Drive call boundaries from audio, use the console only for the label

The control channel was wrong in both directions. Grants fire 0.84-1.62s
before anyone speaks, and srcaddr can drop to 0 while someone is still
talking - one recording came back "-1.61s lead, -0.00s tail", the trim
finding nothing to remove because the window had closed on live speech.
Confirmed by ear: the cut lands at a word boundary on an unfinished word.

Audio is ground truth for WHEN. The console remains the only source of
WHO, so it still supplies talkgroup, alias and rid.

  START  voice onset in the captured audio, with a 0.25s pre-roll that
         now covers only chunk quantisation and threshold ramp-up rather
         than a variable control-channel offset.
  STOP   call_silence_timeout seconds of silence heard in the audio.
  LABEL  resolved AT CLOSE from a bounded rolling history of console
         observations overlapping the window, +4s/-2s, because there is
         no guaranteed ordering between a grant and its audio.
  SPLIT  a console talkgroup change still forces a cut, since two calls
         with no silence between them would otherwise merge into one.

Capture now emits raw PCM instead of MP3. Silence detection becomes
integer arithmetic per chunk with no decode, trimming becomes a byte
offset slice rather than a second ffmpeg pass, and MP3 encoding happens
exactly once at save - uploads are no longer double-encoded.

Audio with no talkgroup anywhere in its window is discarded rather than
uploaded: an untagged call silently poisons incident correlation, which
is worse than losing the audio. Logged at ERROR and counted on
/api/status.

When capture produces no audio at all the old console state machine
still runs, so a node with a broken audio path keeps reporting radio
activity. That is now the only consumer of call_idle_timeout.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

.env.example

@@ -29,29 +29,49 @@ PULSE_SOURCE=drb_sink.monitor
 # Seconds to wait for the shared PulseAudio socket before giving up and retrying.
 PULSE_WAIT_TIMEOUT=30
 
-# Call segmentation: seconds of radio silence before the current recording is
+# --- Call segmentation -------------------------------------------------------
-# closed. Grants on the same talkgroup within this window stay in ONE recording.
+# Recording boundaries come from the AUDIO, not the control channel: a recording
-# Tune ONLY from the "measured control-channel idle" line the edge node logs on
+# starts at voice onset and ends after this many seconds of silence actually
-# every idle-timeout close — silence measured in the audio is a different clock
+# heard in the stream. Transmissions on the same talkgroup separated by less
-# (it also contains the ~1.9s P25 grant-to-speech delay).
+# than this stay in ONE recording, so back-and-forth traffic is one file.
+# Tune from the "measured trailing silence" line logged on every close.
+CALL_SILENCE_TIMEOUT=3.0
+
+# dBFS (RMS over one ~46ms chunk) below which audio counts as silence and the
+# recording is allowed to close.
+#
+# This does NOT need calibrating against your radio's noise floor. Between
+# transmissions the capture is the monitor of a PulseAudio *null sink*, which
+# emits DIGITAL silence: measured on a live node it sits at about -91 dBFS —
+# one least-significant bit of a 16-bit sample — while speech averages about
+# -18 dBFS. Anything from roughly -70 to -40 behaves identically. Only change
+# this if you have replaced the audio path with something that has a real
+# analog noise floor.
+CALL_SILENCE_THRESHOLD_DB=-50
+
+# DEPRECATED as a primary control. Used ONLY when PulseAudio capture is not
+# producing audio, where the old control-channel state machine takes over so
+# the node still reports radio activity (with no recordings) while its audio
+# path is broken.
 CALL_IDLE_TIMEOUT=3
 
-# Seconds of audio kept after the last transmission ends. This is the only
+# Seconds of audio kept past a CONTROL-CHANNEL-derived boundary — a talkgroup
-# headroom protecting the final word of a transmission — usually the disposition
+# change, or a close in the fallback mode above. Buffered audio lags the
-# or the address. Raised 1.0 -> 3.0 after field measurement showed the
+# control channel by ~1.5s (grant-to-speech offset measured 0.84-1.62s), so
-# grant-to-speech offset is ~0.84-1.62s (typically ~1.5s): at 1.0s pad, short
+# cutting at the exact control-channel timestamp clipped the last words of the
-# calls had their recording window close before the voice even started,
+# outgoing call. Does NOT apply to the normal end of a call any more; that
-# clipping speech mid-word. Safe to be generous — trim_silence already strips
+# boundary comes from the audio and needs no pad. Safe to be generous — the
-# the extra back off long calls before upload, so only short transmissions
+# extra is trimmed off again before upload.
-# actually benefit from the larger window.
 CALL_TAIL_PAD_SECONDS=3.0
 
-# Strip leading/trailing dead air before upload. ~63% of an untrimmed recording
+# Strip leading/trailing dead air before upload. Recordings deliberately
-# is silence, which costs Whisper spend and makes it hallucinate text that was
+# over-capture at both ends, and silence costs Whisper spend and makes it
-# never spoken. Only the head and tail are touched, with a guard margin so no
+# hallucinate text that was never spoken. Trimming is a sample-offset slice of
-# syllable is clipped. Set to false to upload raw audio.
+# the buffered PCM (no re-encode) and only ever touches the head and tail, with
+# a guard margin so no syllable is clipped. Set to false to upload raw audio.
 TRIM_SILENCE=true
-# dBFS below which audio counts as silence for detection.
+# dBFS (RMS) below which audio counts as silence when trimming the ends. Kept
+# stricter than CALL_SILENCE_THRESHOLD_DB on purpose.
 TRIM_SILENCE_THRESHOLD_DB=-40
 # Seconds of audio kept either side of detected speech.
 TRIM_SILENCE_GUARD_SECONDS=0.25

drb-edge-node/Dockerfile

@@ -15,5 +15,8 @@ RUN pip install uv && uv pip install --system --no-cache-dir -r requirements.txt
 
 COPY app/ ./app/
 COPY tests/ ./tests/
+# Without this the container runs pytest with asyncio_mode defaulting to strict,
+# so unmarked async tests error out even though they pass locally.
+COPY pytest.ini .
 
 CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "80", "--reload"]

drb-edge-node/app/config.py

@@ -34,49 +34,80 @@ class Settings(BaseSettings):
     # Bounded wait for the shared PulseAudio socket before launching FFmpeg.
     pulse_wait_timeout: float = 30.0
 
-    # Call segmentation — seconds with no active transmission before the current
+    # ------------------------------------------------------------------
-    # recording is closed out. Consecutive grants on the SAME talkgroup inside this
+    # Call segmentation
-    # window are kept in one recording so back-and-forth traffic stays together.
     #
-    # Do NOT tune this against measured *audio* silence: audio gaps also contain
+    # Boundaries come from the AUDIO, not the control channel. A recording
-    # the ~1.9 s P25 grant→speech delay, so they are always longer than the
+    # starts at voice onset and ends after call_silence_timeout seconds of
-    # control-channel idle this timer measures. metadata_watcher logs the real
+    # silence actually heard in the stream. See internal/metadata_watcher.py
-    # measured idle on every idle-timeout close — tune from that.
+    # for why the control channel is no longer trusted for either edge.
+    # ------------------------------------------------------------------
+
+    # Seconds of continuous silence IN THE AUDIO before the current recording is
+    # closed. This is the primary segmentation control. Consecutive
+    # transmissions on the SAME talkgroup separated by less than this stay in
+    # one recording, so back-and-forth traffic is one file.
+    #
+    # Defaults to 3.0 to match the behaviour of the control-channel idle timer
+    # it replaces, but it is NOT the same clock: this one measures real silence
+    # in the audio, with no grant->speech delay mixed in. metadata_watcher logs
+    # the measured trailing silence on every close — tune from that number.
+    call_silence_timeout: float = 3.0
+
+    # dBFS (RMS, measured over one ~46ms capture chunk) below which audio counts
+    # as silence for the purpose of ending a recording.
+    #
+    # This does NOT need field calibration against radio noise. Between
+    # transmissions the capture is the monitor of a PulseAudio *null sink*,
+    # which emits digital silence, not an analog noise floor: measured on a live
+    # node the gap sits at about -91 dBFS, i.e. one least-significant bit of a
+    # 16-bit sample. Speech on the same node averages about -18 dBFS. Anything
+    # between roughly -70 and -40 therefore behaves identically; -50 is chosen
+    # to sit far below even quiet speech while staying far above the floor.
+    call_silence_threshold_db: float = -50.0
+
+    # DEPRECATED as a primary control — used ONLY in console fallback mode, i.e.
+    # when PulseAudio capture is not producing audio and there is nothing to
+    # segment on. Then, and only then, the old control-channel state machine
+    # runs and closes a segment this many seconds after the last observed
+    # transmission. Those segments carry no audio; they exist so the node keeps
+    # reporting real radio activity to C2 while its audio path is broken.
+    #
+    # Do NOT tune this against measured *audio* silence — use
+    # call_silence_timeout for that.
     call_idle_timeout: float = 3.0
 
-    # Audio kept after the observed end of the last transmission. The srcaddr
+    # Audio kept past a CONSOLE-DERIVED segment boundary, covering the fact that
-    # 1→0 edge can be up to one poll (0.5 s) late and the encoder adds its own
+    # buffered audio lags control-channel timestamps by ~1.5s (grant->speech
-    # latency, so this is the only headroom protecting the last word of a
+    # offset measured 0.84-1.62s across 7 field calls).
-    # transmission — which is usually the disposition or the address.
     #
-    # Raised 1.0 -> 3.0 after field measurement showed the recording WINDOW
+    # Still needed, with a narrower job than before. It no longer pads the
-    # (anchored to OP25 control-channel timestamps) closing well before the
+    # normal end of a call — that boundary now comes from the audio itself and
-    # actual voice audio arrives: grant->speech offset measured 0.84-1.62s
+    # needs no pad at all. It applies to the three boundaries that are still
-    # across 7 calls (~1.5s typical). At the old 1.0s pad, a short
+    # control-channel timestamps:
-    # transmission (e.g. a 0.97s control-channel call) had its window close
-    # at T+1.97 while voice didn't start until ~T+1.5 — leaving ~0.4s of
-    # captured speech, clipped mid-word. Confirmed by a 0.57s output file
-    # whose final 0.10s measured -12.2dB, louder than its own -18.2dB
-    # average (i.e. clipped speech, not trailing silence), and by two short
-    # calls that produced no "Trimmed" log line at all because there was no
-    # trailing silence left to trim.
     #
-    # Safe to be generous here: trim_silence already strips trailing silence
+    #   tgid_change            close the outgoing call at the new grant + pad
-    # back to trim_silence_guard_seconds before upload, so a larger pad costs
+    #   tgid_change_unlogged   close at the observing poll + pad
-    # long calls nothing (the extra is trimmed away) while giving short
+    #   idle_timeout           console fallback mode only
-    # transmissions enough window to actually capture the voice. Over-capture
+    #
-    # is free; under-capture loses words permanently. Do not tune this back
+    # Safe to be generous: trim_silence strips trailing silence back to
-    # down without new field data showing the grant->speech offset has
+    # trim_silence_guard_seconds before upload, so a larger pad costs long calls
-    # shrunk — see DEFERRED.md for the call_idle_timeout coupling this value
+    # nothing. Over-capture is free; under-capture loses words permanently. If
-    # now sits at.
+    # the outgoing and incoming recordings overlap in the underlying audio
+    # because of this pad, that is correct — the audio contains both.
     call_tail_pad_seconds: float = 3.0
 
-    # Strip leading/trailing dead air before upload. ~63% of a typical recording
+    # Strip leading/trailing dead air before upload. A recording deliberately
-    # is silence (the grant→speech delay plus the tail pad), which inflates
+    # over-captures at both ends (pre-roll at the head, the whole measured
-    # Whisper cost and is a well-documented trigger for hallucinated transcript
+    # silence run at the tail), which inflates Whisper cost and is a
-    # text. Trimming is conservative — see internal/audio_trim.py.
+    # well-documented trigger for hallucinated transcript text. Trimming is a
+    # sample-offset slice of the buffered PCM — no re-encode — and only ever
+    # touches the head and tail. See internal/audio_trim.py.
     trim_silence: bool = True
-    # Anything quieter than this counts as silence for detection purposes.
+    # dBFS (RMS) below which audio counts as silence when trimming the ends.
+    # Kept above call_silence_threshold_db on purpose: the closer must not miss
+    # speech (permissive), the trimmer must not leave dead air (stricter), and
+    # trim_silence_guard_seconds protects the syllable either way.
     trim_silence_threshold_db: float = -40.0
     # Guard margin kept around detected speech so no syllable is clipped.
     trim_silence_guard_seconds: float = 0.25

drb-edge-node/app/internal/audio_trim.py

@@ -1,270 +1,212 @@
 """
-Conservative leading/trailing silence removal for finished call recordings.
+Leading/trailing silence removal, as a slice of raw PCM.
 
 WHY: P25 grants the channel, radios tune, and only then does a human start
-talking. Measured on six real recordings from a live node, that costs 1.71–2.45 s
+talking; the recorder also deliberately over-captures at the tail (it closes a
-of dead air at the head of every single file, and the tail pad adds its own
+call only after N seconds of silence have actually been HEARD). Both ends
-~0.3–1.0 s. Roughly 63% of every uploaded MP3 was silence. That is not just
+therefore carry dead air. That is not just wasted Whisper spend: silence is a
-wasted Whisper spend: silence is a well-documented trigger for Whisper
+well-documented trigger for Whisper hallucinating text that was never spoken,
-hallucinating text that was never spoken, and a hallucinated sentence poisons
+and a hallucinated sentence poisons entity extraction and then incident
-entity extraction and then incident correlation downstream.
+correlation downstream.
 
 WHY IT IS SAFE: only the head and tail are touched, never the middle, and a
 guard margin is kept around the detected speech so no syllable can be clipped.
-If detection says the whole file is silent we do NOT emit a zero-length file —
+If detection says the whole buffer is silent we do NOT emit a zero-length
-the caller is told and decides (see call_recorder: it skips the upload and logs).
+recording — the caller is told and decides (see call_recorder: it skips the
+upload and logs).
 
 TIMING: trimming changes the audio's duration relative to the call's wall-clock
 start/end, so every trim reports exactly how much was removed from each end.
 Callers must carry those offsets forward — `started_at`/`ended_at` keep meaning
 the CALL's bounds, and the trimmed audio's own bounds are reported separately.
 
-Implementation is two FFmpeg passes (detect, then cut). FFmpeg is already a hard
+HISTORY — THIS USED TO BE TWO FFMPEG PASSES. Detection was `silencedetect`
-dependency of this container and is already running the capture, so this adds no
+parsed out of FFmpeg's stderr, and the cut was a second FFmpeg re-encode. Both
-new moving parts. `silenceremove` in one pass was rejected on purpose: it gives
+are gone: the recorder now buffers PCM, so detection is arithmetic over the
-no way to learn how much it removed, which would make the timing metadata above
+samples and the cut is a byte-offset slice. Consequences worth keeping in mind:
-impossible to produce.
+
+  * The recording is encoded to MP3 exactly ONCE, after this runs, instead of
+    being captured as MP3 and then re-encoded. One less generation of lossy
+    encoding on every upload, and one less subprocess per call.
+  * The threshold is now RMS over a short window (see pcm.rms_dbfs), where
+    FFmpeg's silencedetect compared |sample| per sample. Same units (dBFS),
+    slightly different meaning — do not port an old threshold across without
+    re-reading the field logs.
+  * There is no "is it worth re-encoding" minimum any more. A slice is free, so
+    even a 0.05 s trim is applied.
 """
-import asyncio
-import re
 from dataclasses import dataclass
-from pathlib import Path
+from typing import Optional, Tuple
-from typing import List, Optional, Tuple
 
 from app.config import settings
+from app.internal import pcm
 from app.internal.logger import logger
 
-# Minimum run of quiet before FFmpeg calls it a silence region. Below ~0.3 s this
+# Window the head/tail scan works in. 20 ms is short enough that the guard
-# starts firing on the natural pauses between words, which is not what we want —
+# margin below dwarfs the quantisation error, and long enough that RMS means
-# we only care about the big block of dead air at each end.
+# something.
-MIN_SILENCE_SECONDS = 0.3
+ANALYSIS_WINDOW_SECONDS = 0.02
 
-# A silence region only counts as "leading" if it begins essentially at the file
+# How far in from each end the scan is willing to look before giving up.
-# head. One chunk of slop.
+#
-HEAD_EPSILON_SECONDS = 0.15
+# Bounds the only unbounded cost in this module: the per-sample RMS loop. A
-
+# normal recording resolves within a window or two at the head (the recorder
-# ...and only as "trailing" if it reaches the end of the audio. Do NOT assume a
+# starts on voice onset) and within the silence run at the tail, so this cap is
-# missing `silence_end` marks that case: FFmpeg 6.x flushes a closing
+# never reached in practice. If it IS reached, we leave the audio untrimmed and
-# `silence_end` at EOF, so an all-silence file looks exactly like a file with one
+# say so — shipping an untrimmed recording is always better than shipping none.
-# closed silence region. Verified against ffmpeg 6.1.1 — the reported end lands
+MAX_SCAN_SECONDS = 30.0
-# ~0.03 s short of the (offset-corrected) duration, hence this epsilon. The
-# residual risk is clipping <=0.1 s of audio that follows a >=0.3 s gap right at
-# EOF, which the guard margin below more than covers.
-TAIL_EPSILON_SECONDS = 0.10
-
-# Don't bother re-encoding to reclaim less than this — a re-encode costs a CPU
-# spike on a Pi and a generation of MP3 quality, which is a bad trade for
-# a fraction of a second.
-MIN_TRIM_SECONDS = 0.20
-
-# Bounded so a wedged FFmpeg can never stall the upload path.
-FFMPEG_TIMEOUT_SECONDS = 30.0
-
-_SILENCE_START_RE = re.compile(r"silence_start:\s*(-?[0-9.]+)")
-_SILENCE_END_RE = re.compile(r"silence_end:\s*(-?[0-9.]+)")
-_DURATION_RE = re.compile(r"Duration:\s*(\d+):(\d\d):(\d\d\.\d+)")
-_START_RE = re.compile(r"Duration:[^\n]*?start:\s*(-?[0-9.]+)")
 
 
 @dataclass(frozen=True)
 class TrimResult:
     """Outcome of a trim attempt. `lead`/`tail` are seconds actually removed."""
 
-    path: Optional[Path]
     lead: float = 0.0
     tail: float = 0.0
     duration_before: float = 0.0
     duration_after: float = 0.0
     all_silence: bool = False
     applied: bool = False
+    # True when the scan hit MAX_SCAN_SECONDS without finding speech, so
+    # `all_silence` could not be determined and nothing was trimmed.
+    scan_truncated: bool = False
 
     @property
     def trimmed_seconds(self) -> float:
         return self.lead + self.tail
 
 
-async def _run(cmd: List[str]) -> Tuple[int, str]:
+def _window_bytes() -> int:
-    """Run FFmpeg and return (returncode, stderr). FFmpeg reports on stderr."""
+    return max(pcm.FRAME_BYTES, pcm.byte_offset(ANALYSIS_WINDOW_SECONDS))
-    proc = await asyncio.create_subprocess_exec(
-        *cmd,
-        stdout=asyncio.subprocess.DEVNULL,
-        stderr=asyncio.subprocess.PIPE,
-    )
-    try:
-        _, stderr = await asyncio.wait_for(proc.communicate(), timeout=FFMPEG_TIMEOUT_SECONDS)
-    except asyncio.TimeoutError:
-        try:
-            proc.kill()
-        except Exception:
-            pass
-        raise
-    return proc.returncode or 0, stderr.decode(errors="replace")
 
 
-def _parse_duration(stderr: str) -> Optional[float]:
+def first_signal_offset(
+    audio: bytes,
+    threshold_db: float,
+    limit_seconds: float = MAX_SCAN_SECONDS,
+) -> Optional[int]:
     """
-    Length of the decodable audio, in silencedetect's coordinates.
+    Byte offset of the first window carrying signal, scanning forward.
 
-    MP3 carries encoder delay/padding, which FFmpeg reports as a container
+    None means "no signal found" — either the buffer really is all silence or
-    `start:` offset — the container duration is that much longer than the audio
+    the scan hit `limit_seconds` first; the caller distinguishes the two by
-    silencedetect actually timestamps. Subtracting it is what lets
+    comparing the scanned span against the buffer length.
-    TAIL_EPSILON_SECONDS stay tight enough to be safe.
     """
-    match = _DURATION_RE.search(stderr)
+    window = _window_bytes()
-    if not match:
+    limit = min(len(audio), pcm.byte_offset(limit_seconds) or len(audio))
+    offset = 0
+    while offset < limit:
+        chunk = audio[offset:offset + window]
+        if not pcm.is_silent(chunk, threshold_db):
+            return offset
+        offset += window
     return None
-    hours, minutes, seconds = match.groups()
-    duration = int(hours) * 3600 + int(minutes) * 60 + float(seconds)
-
-    start = _START_RE.search(stderr)
-    if start:
-        offset = float(start.group(1))
-        if 0.0 <= offset < duration:
-            duration -= offset
-    return duration
 
 
-def _parse_silences(stderr: str) -> List[Tuple[float, Optional[float]]]:
+def last_signal_offset(
+    audio: bytes,
+    threshold_db: float,
+    limit_seconds: float = MAX_SCAN_SECONDS,
+) -> Optional[int]:
     """
-    Extract silence regions as (start, end) with end=None when it runs to EOF.
+    Byte offset of the END of the last window carrying signal, scanning back.
 
-    FFmpeg emits `silence_start:` and a later `silence_end:` per region, and
+    Returns the offset one past the last signal-bearing window, so it can be
-    simply never emits the closing line for a region that reaches EOF.
+    used directly as a slice bound.
     """
-    regions: List[Tuple[float, Optional[float]]] = []
+    window = _window_bytes()
-    pending: Optional[float] = None
+    total = pcm.align(len(audio))
-    for line in stderr.splitlines():
+    floor = max(0, total - (pcm.byte_offset(limit_seconds) or total))
-        if "silencedetect" not in line:
+    offset = total
-            continue
+    while offset > floor:
-        start = _SILENCE_START_RE.search(line)
+        start = max(floor, offset - window)
-        if start:
+        if not pcm.is_silent(audio[start:offset], threshold_db):
-            pending = float(start.group(1))
+            return offset
-            continue
+        offset = start
-        end = _SILENCE_END_RE.search(line)
+    return None
-        if end and pending is not None:
-            regions.append((pending, float(end.group(1))))
-            pending = None
-    if pending is not None:
-        regions.append((pending, None))
-    return regions
 
 
-def speech_bounds(
+def keep_window(
-    regions: List[Tuple[float, Optional[float]]],
+    first_signal: Optional[int],
-    duration: float,
+    last_signal: Optional[int],
-    guard: float,
+    total_bytes: int,
-) -> Tuple[float, float, bool]:
+    guard_bytes: int,
+) -> Tuple[int, int]:
     """
-    Turn detected silence regions into the [start, end] window to keep.
+    Turn detected signal bounds into the byte range to keep.
 
-    Pure and side-effect free so the decision logic is unit-testable without
+    Pure and side-effect free so the decision that can destroy a transmission
-    FFmpeg. Returns (start, end, all_silence).
+    stays unit-testable without any audio. Offsets are sample-aligned and
+    clamped to the buffer.
     """
-    speech_start = 0.0
+    total = pcm.align(total_bytes)
-    speech_end = duration
+    start = 0 if first_signal is None else max(0, first_signal - guard_bytes)
-
+    end = total if last_signal is None else min(total, last_signal + guard_bytes)
-    if regions:
+    start = pcm.align(start)
-        head_start, head_end = regions[0]
+    end = pcm.align(end)
-        if head_start <= HEAD_EPSILON_SECONDS and head_end is not None:
+    if end <= start:
-            speech_start = head_end
+        return 0, total
-
+    return start, end
-        tail_start, tail_end = regions[-1]
-        # `None` = pre-6.x FFmpeg, which simply stopped reporting at EOF.
-        reaches_eof = tail_end is None or (duration - tail_end) <= TAIL_EPSILON_SECONDS
-        if reaches_eof:
-            speech_end = min(speech_end, tail_start)
-
-    if speech_end <= speech_start:
-        # Detection says there is no speech anywhere in the file.
-        return 0.0, duration, True
-
-    # Guard margin: never trim right up against the first/last syllable.
-    keep_start = max(0.0, speech_start - guard)
-    keep_end = min(duration, speech_end + guard)
-    return keep_start, keep_end, False
 
 
-async def trim_silence(
+def trim_pcm(
-    path: Path,
+    audio: bytes,
-    sample_rate: str,
-    bitrate: str,
     threshold_db: Optional[float] = None,
     guard: Optional[float] = None,
-) -> TrimResult:
+) -> Tuple[bytes, TrimResult]:
     """
-    Trim leading/trailing silence in place, preserving the original on failure.
+    Return (kept_audio, result). Never raises and never returns empty audio.
 
-    Never raises: any problem degrades to "leave the file exactly as it was",
+    An all-silence buffer is returned UNCHANGED with `all_silence=True`: the
-    because shipping an untrimmed recording is far better than shipping none.
+    caller decides what to do with a recording that contains no speech at all —
+    that is itself a signal (squelch misconfigured, wrong sink, dead audio
+    path), not something to silently truncate to nothing.
     """
     threshold = settings.trim_silence_threshold_db if threshold_db is None else threshold_db
     margin = settings.trim_silence_guard_seconds if guard is None else guard
 
-    try:
+    total = pcm.align(len(audio))
-        _, stderr = await _run([
+    duration = pcm.seconds(total)
-            "ffmpeg", "-hide_banner", "-nostdin", "-nostats",
+    if total <= 0:
-            "-i", str(path),
+        return audio, TrimResult()
-            "-af", f"silencedetect=noise={threshold}dB:d={MIN_SILENCE_SECONDS}",
-            "-f", "null", "-",
-        ])
-    except Exception as e:
-        logger.warning(f"Silence detection failed for {path.name} ({e}) — uploading untrimmed.")
-        return TrimResult(path=path)
 
-    duration = _parse_duration(stderr)
+    first = first_signal_offset(audio, threshold)
-    if duration is None or duration <= 0:
+    if first is None:
-        logger.warning(f"Could not determine duration of {path.name} — uploading untrimmed.")
+        scanned = min(total, pcm.byte_offset(MAX_SCAN_SECONDS) or total)
-        return TrimResult(path=path)
+        if scanned < total:
-
+            # Could not prove it is all silence; refuse to guess.
-    regions = _parse_silences(stderr)
-    keep_start, keep_end, all_silence = speech_bounds(regions, duration, margin)
-
-    if all_silence:
-        # Deliberately NOT trimmed to nothing. The caller decides what to do with
-        # a recording that contains no speech at all — that is itself a signal
-        # (squelch misconfigured, wrong sink, dead audio path).
             logger.warning(
-            f"{path.name} is entirely silence ({duration:.2f}s, threshold {threshold}dB) — "
+                f"Silence scan gave up after {MAX_SCAN_SECONDS:.0f}s without finding speech in a "
+                f"{duration:.1f}s recording — leaving it untrimmed."
+            )
+            return audio, TrimResult(
+                duration_before=duration, duration_after=duration, scan_truncated=True
+            )
+        logger.warning(
+            f"Recording is entirely silence ({duration:.2f}s, threshold {threshold:.1f}dBFS RMS) — "
             "no speech detected."
         )
-        return TrimResult(path=path, duration_before=duration, duration_after=duration, all_silence=True)
+        return audio, TrimResult(duration_before=duration, duration_after=duration, all_silence=True)
 
-    lead = keep_start
+    last = last_signal_offset(audio, threshold)
-    tail = duration - keep_end
+    guard_bytes = pcm.byte_offset(margin)
-    if (lead + tail) < MIN_TRIM_SECONDS:
+    keep_start, keep_end = keep_window(first, last, total, guard_bytes)
-        return TrimResult(path=path, duration_before=duration, duration_after=duration)
 
-    trimmed = path.with_name(f"{path.stem}_trimmed{path.suffix}")
+    lead = pcm.seconds(keep_start)
-    try:
+    tail = pcm.seconds(total - keep_end)
-        code, err = await _run([
+    if keep_start <= 0 and keep_end >= total:
-            "ffmpeg", "-hide_banner", "-nostdin", "-nostats",
+        return audio[:total], TrimResult(duration_before=duration, duration_after=duration)
-            "-loglevel", "warning", "-y",
-            "-ss", f"{keep_start:.3f}",
-            "-i", str(path),
-            "-t", f"{keep_end - keep_start:.3f}",
-            "-ac", "1", "-ar", sample_rate, "-b:a", bitrate,
-            "-f", "mp3", str(trimmed),
-        ])
-    except Exception as e:
-        logger.warning(f"Silence trim failed for {path.name} ({e}) — uploading untrimmed.")
-        trimmed.unlink(missing_ok=True)
-        return TrimResult(path=path, duration_before=duration, duration_after=duration)
 
-    if code != 0 or not trimmed.exists() or trimmed.stat().st_size == 0:
+    kept = audio[keep_start:keep_end]
-        logger.warning(f"Silence trim produced no output for {path.name} ({err.strip()}) — uploading untrimmed.")
+    after = pcm.seconds(len(kept))
-        trimmed.unlink(missing_ok=True)
-        return TrimResult(path=path, duration_before=duration, duration_after=duration)
-
-    trimmed.replace(path)
     logger.info(
-        f"Trimmed {path.name}: -{lead:.2f}s lead, -{tail:.2f}s tail "
+        f"Trimmed recording: -{lead:.2f}s lead, -{tail:.2f}s tail "
-        f"({duration:.2f}s → {keep_end - keep_start:.2f}s)"
+        f"({duration:.2f}s -> {after:.2f}s, threshold {threshold:.1f}dBFS RMS)"
     )
-    return TrimResult(
+    return kept, TrimResult(
-        path=path,
         lead=lead,
         tail=tail,
         duration_before=duration,
-        duration_after=keep_end - keep_start,
+        duration_after=after,
         applied=True,
     )

drb-edge-node/app/internal/call_recorder.py

@@ -1,16 +1,26 @@
 """
 Continuous PulseAudio capture: a ring buffer for PRE-ROLL, a per-call
-accumulator for the call itself.
+accumulator for the call itself, and the voice-activity signal that decides
+where calls begin and end.
 
 A persistent capture process runs for the lifetime of the node. Spawning FFmpeg
 per call used to lose the first 1-2 s to process startup, which meant short
 transmissions produced empty files, so capture never stops.
 
+RAW PCM, NOT MP3 — this is the change everything else hangs off. FFmpeg is
+asked for s16le/22050/mono on stdout instead of an MP3 stream, so:
+
+  * silence detection is integer arithmetic over each chunk as it arrives, with
+    no decode, which is what makes AUDIO-DRIVEN call boundaries possible;
+  * trimming is a byte-offset slice, not a second FFmpeg pass;
+  * MP3 encoding happens exactly ONCE, at save time, so uploads are no longer
+    double-encoded.
+
 TWO BUFFERS, TWO JOBS — this split is load-bearing:
 
   RING BUFFER   holds the last RING_BUFFER_SECONDS of audio at all times. Its
-                only job is PRE-ROLL: however late we notice a grant, we can
+                only job is PRE-ROLL: however late the segmenter notices voice
-                still seek back to OP25's exact timestamp. It is sized for
+                onset, we can still seek back before it. It is sized for
                 detection latency, nothing else.
 
   ACCUMULATOR   opened by start_recording(), fed by every subsequent chunk, and
@@ -20,6 +30,14 @@ TWO BUFFERS, TWO JOBS — this split is load-bearing:
                 which silently clamped the front of any call longer than
                 RING_BUFFER_SECONDS.
 
+VOICE ACTIVITY is tracked CONTINUOUSLY, not only while recording, because the
+segmenter starts a call from audio onset. `_last_voice_epoch` and
+`_voice_onset_epoch` are the whole interface: metadata_watcher polls them via
+audio_activity() and owns every decision about segment boundaries. This module
+deliberately does not open or close calls by itself — attribution (which
+talkgroup this audio belongs to) lives in the watcher, and audio alone cannot
+answer it.
+
 Why not Icecast: it lags ~1 s at connect and drifts progressively to 100 s+, so
 slice timestamps and audio content diverge without bound. Icecast stays in the
 stack for frontend/mobile live listening; it is not an accuracy path.
@@ -31,6 +49,12 @@ the same clock and the pre-roll arithmetic below is a direct subtraction with no
 offset mapping. (A wall-clock STEP — e.g. a large NTP correction — would corrupt
 at most the calls in flight at that instant; the buffer self-heals within
 RING_BUFFER_SECONDS.)
+
+NOTE on what a chunk timestamp means: it is the ARRIVAL time of that audio at
+this process, which lags the moment the words were spoken by the PulseAudio →
+FFmpeg → pipe latency. Every timestamp this module produces is in that same
+arrival clock, so differences between them are exact; only comparisons against
+OP25's control-channel timestamps carry the lag, and those are padded for it.
 """
 import asyncio
 import time
@@ -42,57 +66,72 @@ from typing import List, Optional, Tuple
 
 import httpx
 from app.config import settings
-from app.internal import audio_trim, credentials, pulse
+from app.internal import audio_trim, credentials, pcm, pulse
 from app.internal.logger import logger
 
 # Safety cap on a single recording; mirrors MAX_SEGMENT_SECONDS in metadata_watcher.
 MAX_RECORDING_SECONDS = 600
 
-# Audio included ahead of OP25's call_log timestamp. The grant is logged when the
+# Audio included ahead of the detected voice onset.
-# channel is granted, so the first syllable can land marginally before it.
 #
-# Kept small on purpose: measurement on a live node shows 1.71–2.45 s of real
+# Under audio-driven segmentation this is no longer covering a variable
-# grant→speech delay on every call, so there is no clipping risk at the head and
+# control-channel offset — it covers exactly two things: the analysis chunk
-# a larger pre-roll would only add dead air.
+# quantisation (~46 ms) and the possibility that the first syllable ramps up
+# through the silence threshold rather than crossing it instantly. 0.25 s is
+# generous for both, and anything it drags in that is genuinely silence gets
+# trimmed off again before upload.
 PRE_ROLL_SECONDS = 0.25
 
 # Rolling history kept for PRE-ROLL ONLY. Budget for the worst realistic
-# detection latency: 0.5 s poll interval + ~0.2 s http_server blocking floor +
+# detection latency: the segmenter polls every 0.5 s and can stall for up to a
-# up to 3 s httpx timeout on a stalled poll + callback work ≈ 4 s from grant to
+# 3 s httpx timeout on a bad OP25 poll, so ~4 s from onset to start_recording().
-# start_recording(). 30 s is ~7x that margin, and at 16 kbps costs only ~60 KB of
+# 30 s is ~7x that margin. At 44.1 KB/s of PCM it costs ~1.3 MB of RAM. This
-# RAM. This value does NOT bound call length — the accumulator does.
+# value does NOT bound call length — the accumulator does.
 RING_BUFFER_SECONDS = 30
 
-# ~128 ms of audio per chunk at 16 kbps. Chunk size IS the timestamp resolution of
+# ~46 ms of audio per chunk. Chunk size is BOTH the timestamp resolution of the
-# the ring buffer, so it must stay well under PRE_ROLL_SECONDS — the old 4096-byte
+# ring buffer AND the window silence detection runs over, so it has to stay well
-# reads were ~2 s per chunk, which made sub-second slicing meaningless.
+# under PRE_ROLL_SECONDS and well under the shortest utterance we care about.
-READ_CHUNK_BYTES = 256
+READ_CHUNK_BYTES = 2048
 
-# Encoder settings, matched on purpose to what Liquidsoap already pushes to
+# Encoder settings for the single encode at save time, matched on purpose to
-# Icecast — %mp3(bitrate=16, samplerate=22050, stereo=false) — so the C2 /upload
+# what Liquidsoap already pushes to Icecast — %mp3(bitrate=16, samplerate=22050,
-# endpoint keeps receiving exactly the kind of MP3 it has always received
+# stereo=false) — so the C2 /upload endpoint keeps receiving exactly the kind of
-# (multipart "audio/mpeg", stored to GCS as .mp3, then fed to Whisper).
+# MP3 it has always received (multipart "audio/mpeg", stored to GCS as .mp3,
-# Change both of these together if you ever want higher-fidelity uploads.
+# then fed to Whisper). MP3_SAMPLE_RATE MUST equal pcm.SAMPLE_RATE: the encode
+# is a straight pass with no resampling.
 MP3_BITRATE = "16k"
-MP3_SAMPLE_RATE = "22050"
+MP3_SAMPLE_RATE = str(pcm.SAMPLE_RATE)
 
-# Hard memory ceiling for one call's accumulator. 16 kbps is 2 KB/s, so 600 s of
+# Bounded so a wedged encoder can never stall the upload path.
-# call is ~1.2 MB; 4x that is the ceiling, which both leaves room for encoder
+ENCODE_TIMEOUT_SECONDS = 60.0
-# overshoot and guarantees a runaway call can never eat a Pi's RAM.
+
-_MP3_BYTES_PER_SECOND = 16_000 // 8
+# Hard memory ceiling for one call's accumulator.
-MAX_RECORDING_BYTES = MAX_RECORDING_SECONDS * _MP3_BYTES_PER_SECOND * 4
+#
+# PCM costs 44.1 KB/s where the old MP3 buffer cost 2 KB/s, so this had to be
+# re-derived rather than carried over. 600 s (the time cap) of PCM is 26.5 MB;
+# 32 MiB is ~761 s, which guarantees the TIME cap always bites first and a legal
+# call is never truncated by the byte cap. Peak resident audio is therefore
+# ~33.5 MB for the accumulator plus ~1.3 MB for the ring buffer.
+#
+# Rejected alternatives, for the record: spilling to disk (SD-card wear on a Pi,
+# and I/O in the close path); encoding incrementally into MP3 as chunks arrive
+# (puts a subprocess back in the hot path and makes the sample-accurate post-hoc
+# trim impossible); a lower capture sample rate (changes what Whisper receives).
+MAX_RECORDING_BYTES = 32 * 1024 * 1024
 
 # How long stop_recording() will wait for captured audio to actually reach the
-# call's end timestamp. PulseAudio → FFmpeg → MP3 encoder → muxer → our pipe read
+# call's end timestamp. PulseAudio → FFmpeg → our pipe read is a pipeline with
-# is a pipeline with latency, so at the instant a call ends the newest buffered
+# latency, so at the instant a CONTROL-CHANNEL derived end is computed the
-# chunk is typically a few hundred ms OLDER than the end epoch. Slicing
+# newest buffered chunk is typically a few hundred ms OLDER than that epoch.
-# immediately therefore cuts the tail short — which costs the last word of the
+# Slicing immediately therefore cuts the tail short — which costs the last word
-# transmission, usually the disposition or the address. Bounded so a dead capture
+# of the transmission, usually the disposition or the address.
-# can never hang the upload path.
 #
-# Must exceed settings.call_tail_pad_seconds (default 3.0), otherwise a
+# An audio-driven close never needs this (its end epoch is derived from audio
-# tgid_change close — which pads past a timestamp that is still ~now — gives up
+# that is already buffered, by construction), but a tgid_change close still
-# before the padded audio has been captured and warns on every talkgroup switch.
+# pads past a control-channel timestamp that is ~now, so the wait must exceed
+# settings.call_tail_pad_seconds (default 3.0) or it would warn on every
+# talkgroup switch. Bounded so a dead capture can never hang the upload path.
 TAIL_WAIT_TIMEOUT_SECONDS = 4.0
 TAIL_WAIT_POLL_SECONDS = 0.05
 
@@ -114,12 +153,38 @@ _SOURCE_MISSING_MARKERS = ("no such process", "no such device")
 _NO_DAEMON_MARKERS = ("connection refused", "connection failure")
 
 
+@dataclass(frozen=True)
+class AudioActivity:
+    """
+    What the capture stream is doing right now, as raw facts.
+
+    Deliberately carries no decision — no "should a call be open" boolean —
+    because every threshold comparison belongs to metadata_watcher, which owns
+    the segment state machine and (in tests) an injectable clock. This is a
+    snapshot of observations, nothing more.
+
+    `voice_onset_epoch` is the arrival timestamp of the first chunk of the most
+    recent run of voice. It is NOT cleared when that run ends, so a caller must
+    check `last_voice_epoch` against its own clock before treating the run as
+    live. That is intentional: the segmenter needs the onset of the run it just
+    finished recording in order to avoid re-opening on the same run.
+    """
+
+    capturing: bool
+    recording: bool
+    last_voice_epoch: Optional[float] = None
+    voice_onset_epoch: Optional[float] = None
+    # Convenience for /api/status only; the segmenter recomputes this against
+    # its own clock.
+    silence_seconds: float = 0.0
+
+
 @dataclass
 class _ActiveRecording:
     """Audio accumulating for the call currently being recorded."""
 
     call_id: str
-    call_start: float           # OP25 grant epoch
+    call_start: float           # detected voice onset (or a caller-supplied epoch)
     slice_start: float          # call_start - PRE_ROLL_SECONDS
     chunks: List[Tuple[float, bytes]] = field(default_factory=list)
     total_bytes: int = 0
@@ -140,7 +205,7 @@ class Recording:
 
         wall_clock_of(audio_offset_t) == audio_start_epoch + t
 
-    Bounds are accurate to ±one capture chunk (~128 ms).
+    Bounds are accurate to ±one capture chunk (~46 ms).
     """
 
     call_id: str
@@ -153,13 +218,66 @@ class Recording:
     all_silence: bool = False
 
 
+async def encode_mp3(audio: bytes, path: Path) -> bool:
+    """
+    The one and only encode in the pipeline: raw PCM in, MP3 file out.
+
+    Module-level rather than a method so tests can substitute it without
+    needing FFmpeg, and so the "exactly one encode per call" property is
+    trivially observable.
+    """
+    if not audio:
+        return False
+    cmd = [
+        "ffmpeg",
+        "-hide_banner", "-nostdin", "-nostats",
+        "-loglevel", "warning", "-y",
+        "-f", "s16le",
+        "-ar", MP3_SAMPLE_RATE,
+        "-ac", str(pcm.CHANNELS),
+        "-i", "pipe:0",
+        "-ar", MP3_SAMPLE_RATE,
+        "-ac", str(pcm.CHANNELS),
+        "-b:a", MP3_BITRATE,
+        "-f", "mp3", str(path),
+    ]
+    try:
+        proc = await asyncio.create_subprocess_exec(
+            *cmd,
+            stdin=asyncio.subprocess.PIPE,
+            stdout=asyncio.subprocess.DEVNULL,
+            stderr=asyncio.subprocess.PIPE,
+        )
+    except Exception as e:
+        logger.error(f"Could not launch the MP3 encoder ({e}) — recording not saved.")
+        return False
+
+    try:
+        _, stderr = await asyncio.wait_for(proc.communicate(audio), timeout=ENCODE_TIMEOUT_SECONDS)
+    except asyncio.TimeoutError:
+        try:
+            proc.kill()
+        except Exception:
+            pass
+        logger.error(f"MP3 encode timed out after {ENCODE_TIMEOUT_SECONDS:.0f}s — recording not saved.")
+        return False
+    except Exception as e:
+        logger.error(f"MP3 encode failed ({e}) — recording not saved.")
+        return False
+
+    if proc.returncode != 0:
+        logger.error(f"MP3 encode exited {proc.returncode}: {stderr.decode(errors='replace').strip()}")
+        return False
+    return True
+
+
 class CallRecorder:
     """Continuous PulseAudio capture: ring buffer for pre-roll, accumulator per call."""
 
     def __init__(self):
         self._recordings_dir = Path(settings.recordings_path)
 
-        # Ring buffer: deque of (wall_clock_epoch_at_arrival, mp3_bytes).
+        # Ring buffer: deque of (wall_clock_epoch_at_arrival, pcm_bytes).
         # Pre-roll only — see the module docstring.
         self._buffer: deque[Tuple[float, bytes]] = deque()
         self._buffer_bytes: int = 0
@@ -168,6 +286,10 @@ class CallRecorder:
         self._proc: Optional[asyncio.subprocess.Process] = None
         self._capturing: bool = False
 
+        # Voice activity, tracked continuously — not only while recording.
+        self._last_voice_epoch: Optional[float] = None
+        self._voice_onset_epoch: Optional[float] = None
+
         # Active recording state (None when idle)
         self._active: Optional[_ActiveRecording] = None
 
@@ -204,14 +326,12 @@ class CallRecorder:
             "-hide_banner", "-nostdin", "-nostats",
             "-loglevel", "warning",
             "-f", "pulse", "-i", settings.pulse_source,
-            "-ac", "1",
+            "-ac", str(pcm.CHANNELS),
             "-ar", MP3_SAMPLE_RATE,
-            "-b:a", MP3_BITRATE,
+            # Raw PCM on stdout. No muxer, so no -flush_packets games: s16le is
-            # Without this, the MP3 muxer fills its 32 KB AVIO buffer before
+            # a bare byte stream and every byte FFmpeg produces is immediately
-            # writing anything — 16 s of audio per burst at 16 kbps, which would
+            # readable, which is what keeps arrival timestamps honest.
-            # destroy the arrival timestamps the slicing depends on.
+            "-f", "s16le", "-",
-            "-flush_packets", "1",
-            "-f", "mp3", "-",
         ]
 
     async def _capture_loop(self) -> None:
@@ -240,7 +360,7 @@ class CallRecorder:
 
     async def _run_capture(self) -> None:
         cmd = self._ffmpeg_command()
-        logger.info(f"Starting capture: ffmpeg -f pulse -i {settings.pulse_source}")
+        logger.info(f"Starting capture: ffmpeg -f pulse -i {settings.pulse_source} (s16le/{MP3_SAMPLE_RATE}/mono)")
         self._last_stderr_lines.clear()
         proc = await asyncio.create_subprocess_exec(
             *cmd,
@@ -252,8 +372,14 @@ class CallRecorder:
         try:
             assert proc.stdout is not None
             while True:
-                chunk = await proc.stdout.read(READ_CHUNK_BYTES)
+                # readexactly, not read: a fixed chunk keeps every analysis
-                if not chunk:
+                # window the same length AND guarantees sample alignment, so a
+                # short read can never split a 16-bit sample across chunks.
+                try:
+                    chunk = await proc.stdout.readexactly(READ_CHUNK_BYTES)
+                except asyncio.IncompleteReadError as partial:
+                    if partial.partial:
+                        self._ingest(partial.partial)
                     break  # EOF — FFmpeg died or the source went away
                 if not self._capturing:
                     self._capturing = True
@@ -339,9 +465,44 @@ class CallRecorder:
         except Exception:
             pass
 
+    # ------------------------------------------------------------------
+    # Voice activity
+    # ------------------------------------------------------------------
+
+    def _note_voice(self, chunk: bytes, now: float) -> None:
+        """
+        Update the continuous voice-activity marks from one arriving chunk.
+
+        A chunk carrying signal starts a NEW run whenever the gap since the last
+        one is at least the configured silence timeout — i.e. runs are separated
+        by exactly the same threshold that ends a recording, so the segmenter's
+        "did a new run begin" and "did the recording end" questions can never
+        disagree with each other.
+        """
+        if pcm.is_silent(chunk, settings.call_silence_threshold_db):
+            return
+        gap = settings.call_silence_timeout
+        if self._last_voice_epoch is None or (now - self._last_voice_epoch) >= gap:
+            self._voice_onset_epoch = now
+        self._last_voice_epoch = now
+
+    def audio_activity(self) -> AudioActivity:
+        """Snapshot of the capture stream for metadata_watcher (and /api/status)."""
+        last = self._last_voice_epoch
+        silence = (time.time() - last) if last is not None else 0.0
+        return AudioActivity(
+            capturing=self._capturing,
+            recording=self._active is not None,
+            last_voice_epoch=last,
+            voice_onset_epoch=self._voice_onset_epoch,
+            silence_seconds=max(0.0, silence),
+        )
+
     def _ingest(self, chunk: bytes) -> None:
         """Append a chunk to the ring buffer and, if recording, the accumulator."""
         now = time.time()
+        self._note_voice(chunk, now)
+
         self._buffer.append((now, chunk))
         self._buffer_bytes += len(chunk)
 
@@ -375,9 +536,10 @@ class CallRecorder:
 
     async def start_recording(self, call_id: str, start_epoch: Optional[float] = None) -> bool:
         """
-        Open a recording. `start_epoch` is OP25's call_log timestamp (host wall
+        Open a recording. `start_epoch` is the detected voice onset (host wall
-        clock); the slice begins PRE_ROLL_SECONDS before it. Omit it only when no
+        clock, same clock as the chunk stamps); the slice begins
-        OP25 timestamp is available — then we fall back to "now", losing precision.
+        PRE_ROLL_SECONDS before it. Omit it only when no onset is available —
+        then we fall back to "now", losing the pre-roll's precision.
         """
         if self._active is not None:
             logger.warning(f"Recording already active ({self._active.call_id}) — ignoring start for {call_id}.")
@@ -396,8 +558,8 @@ class CallRecorder:
         oldest = self._buffer[0][0] if self._buffer else None
         if oldest is not None and slice_start < oldest:
             # Pre-roll predates the buffer: node just started, capture restarted,
-            # or OP25's timestamp is far in the past. Clamp and say so LOUDLY —
+            # or the onset is far in the past. Clamp and say so LOUDLY — this is
-            # this is silent audio loss otherwise.
+            # silent audio loss otherwise.
             clamped = oldest - slice_start
             logger.warning(
                 f"BUFFER CLAMP: pre-roll for call {call_id} predates buffered audio by "
@@ -417,13 +579,28 @@ class CallRecorder:
         logger.info(f"Recording started: {call_id} (slice from {slice_start:.3f})")
         return True
 
+    async def discard_recording(self) -> None:
+        """
+        Drop the open recording without writing anything.
+
+        Used when the segmenter decides the audio must not be kept — today only
+        the unattributed/orphan-audio path, where uploading would inject a call
+        with no talkgroup into correlation.
+        """
+        active, self._active = self._active, None
+        if active is not None:
+            logger.info(f"Discarded buffered audio for {active.call_id} ({active.total_bytes} bytes).")
+
     async def stop_recording(self, end_epoch: Optional[float] = None) -> Optional[Recording]:
         """
-        Close the recording and write the file. `end_epoch` is host wall clock.
+        Close the recording, trim it, encode it once and write the file.
+        `end_epoch` is host wall clock.
 
         Waits (bounded) for captured audio to actually cover `end_epoch` before
-        slicing — see TAIL_WAIT_TIMEOUT_SECONDS. Returns None only when there was
+        slicing — see TAIL_WAIT_TIMEOUT_SECONDS. An audio-driven close never
-        no recording open or no audio at all.
+        needs that wait, because its end epoch is derived from audio that is
+        already buffered; a control-channel-derived close (tgid_change) does.
+        Returns None only when there was no recording open or no audio at all.
         """
         active = self._active
         if active is None:
@@ -439,19 +616,21 @@ class CallRecorder:
         await self._await_tail(end, call_id)
         self._active = None
 
-        chunks: List[bytes] = []
+        parts: List[bytes] = []
+        total = 0
         last_ts = slice_start
         for ts, chunk in active.chunks:
             if ts < slice_start:
                 continue
-            chunks.append(chunk)
+            parts.append(chunk)
+            total += len(chunk)
             last_ts = ts
             if ts >= end:
                 # Include the chunk straddling `end` so the tail is never clipped,
                 # then stop.
                 break
 
-        if not chunks:
+        if not parts:
             logger.warning(
                 f"No buffered audio for call {call_id} "
                 f"(window {slice_start:.3f}–{end:.3f}) — PulseAudio capture may be down."
@@ -464,29 +643,55 @@ class CallRecorder:
                 "audio — the tail is short. Capture may be stalled or restarting."
             )
 
-        self._recordings_dir.mkdir(parents=True, exist_ok=True)
+        raw = b"".join(parts)
-        ts_str = datetime.now(timezone.utc).strftime("%Y%m%d_%H%M%S")
-        output_path = self._recordings_dir / f"{ts_str}_{call_id}.mp3"
-
-        output_path.write_bytes(b"".join(chunks))
-
-        size = output_path.stat().st_size
-        if size <= 0:
-            output_path.unlink(missing_ok=True)
-            logger.warning(f"Recording for call {call_id} produced an empty file.")
-            return None
-
-        audio_end = min(end, last_ts)
-        logger.info(f"Recording saved: {output_path.name} ({size} bytes, {audio_end - slice_start:.2f}s window)")
-
         recording = Recording(
             call_id=call_id,
-            path=output_path,
+            path=None,
             audio_start_epoch=slice_start,
-            audio_end_epoch=audio_end,
+            audio_end_epoch=slice_start + pcm.seconds(len(raw)),
             clamped_seconds=active.clamped_seconds,
         )
-        return await self._apply_trim(recording)
+        return await self._finish(recording, raw)
+
+    async def _finish(self, recording: Recording, raw: bytes) -> Optional[Recording]:
+        """Trim, encode exactly once, and write the MP3."""
+        audio = raw
+        if settings.trim_silence:
+            audio, result = audio_trim.trim_pcm(audio)
+            if result.all_silence:
+                logger.warning(
+                    f"Call {recording.call_id} contains no speech at all — skipping upload. "
+                    "Check squelch, the Liquidsoap output and the drb_sink monitor."
+                )
+                recording.all_silence = True
+                return recording
+            if result.applied:
+                recording.lead_trimmed = result.lead
+                recording.tail_trimmed = result.tail
+                # Wall clock of the trimmed audio's first and last sample.
+                recording.audio_start_epoch += result.lead
+                recording.audio_end_epoch -= result.tail
+
+        self._recordings_dir.mkdir(parents=True, exist_ok=True)
+        ts_str = datetime.now(timezone.utc).strftime("%Y%m%d_%H%M%S")
+        output_path = self._recordings_dir / f"{ts_str}_{recording.call_id}.mp3"
+
+        if not await encode_mp3(audio, output_path):
+            output_path.unlink(missing_ok=True)
+            return None
+
+        size = output_path.stat().st_size if output_path.exists() else 0
+        if size <= 0:
+            output_path.unlink(missing_ok=True)
+            logger.warning(f"Recording for call {recording.call_id} produced an empty file.")
+            return None
+
+        recording.path = output_path
+        logger.info(
+            f"Recording saved: {output_path.name} ({size} bytes, "
+            f"{pcm.seconds(len(audio)):.2f}s audio)"
+        )
+        return recording
 
     async def _await_tail(self, end: float, call_id: str) -> float:
         """
@@ -519,41 +724,6 @@ class CallRecorder:
         )
         return waited
 
-    async def _apply_trim(self, recording: Recording) -> Recording:
-        """
-        Strip leading/trailing dead air and keep the timing metadata honest.
-
-        An all-silence recording is NOT uploaded: it carries no information and
-        silence is exactly what makes Whisper hallucinate. It is logged instead,
-        because it also means something is wrong with the audio path.
-        """
-        if not settings.trim_silence or recording.path is None:
-            return recording
-
-        result = await audio_trim.trim_silence(
-            recording.path,
-            sample_rate=MP3_SAMPLE_RATE,
-            bitrate=MP3_BITRATE,
-        )
-
-        if result.all_silence:
-            logger.warning(
-                f"Call {recording.call_id} contains no speech at all — skipping upload. "
-                "Check squelch, the Liquidsoap output and the drb_sink monitor."
-            )
-            recording.path.unlink(missing_ok=True)
-            recording.path = None
-            recording.all_silence = True
-            return recording
-
-        if result.applied:
-            recording.lead_trimmed = result.lead
-            recording.tail_trimmed = result.tail
-            # Wall clock of the trimmed audio's first and last sample.
-            recording.audio_start_epoch += result.lead
-            recording.audio_end_epoch -= result.tail
-        return recording
-
     # ------------------------------------------------------------------
     # Upload (unchanged interface)
     # ------------------------------------------------------------------

drb-edge-node/app/internal/metadata_watcher.py

@@ -1,81 +1,160 @@
 """
-Event-driven call state machine.
+Call segmentation: AUDIO decides the boundaries, the CONSOLE decides the label.
 
-Replaces the old hang-counter inference (which derived call start from "a tgid
+    START  first chunk of audio above the silence threshold (voice onset).
-appeared in channel_update" and call end from N polls of silence) with the two
+    STOP   settings.call_silence_timeout seconds of continuous silence HEARD in
-authoritative signals OP25 actually exposes:
+           that audio.
+    LABEL  talkgroup / alias / rid, resolved from OP25 console observations that
+           fall inside the recording's window, resolved AT CLOSE TIME.
+    SPLIT  a console talkgroup change still forces a cut, even mid-audio.
 
-  START — a `call_log` entry. OP25 appends one at channel-grant time stamped with
+WHY THE CONTROL CHANNEL NO LONGER DECIDES BOUNDARIES. The previous design
-          its own time.time(). This is an exact start timestamp, not the moment
+started a segment on an OP25 `call_log` grant and ended it by inferring from the
-          our poll happened to notice, so recordings can be sliced back to it.
+control channel: the `srcaddr != 0 -> 0` edge started an idle timer and the
+segment closed call_idle_timeout seconds later. Both halves were measured wrong
+in the field:
 
-  END   — the `srcaddr` != 0 → `srcaddr` == 0 transition in `channel_update`.
+  * The grant fires 0.84-1.62 s (variable) before anyone speaks, so a
-          OP25 never reports call termination externally: internally it ends a
+    grant-anchored window is always guessing at the offset.
-          call on the P25 Terminator Data Unit (duid15) or 3 voice-framing
+  * `srcaddr` can drop to 0 WHILE SOMEONE IS STILL TALKING. Measured across six
-          timeouts, but neither becomes a log entry. What *is* observable is that
+    recordings, five had healthy trailing silence trimmed (-0.53 s to -2.48 s)
-          `srcaddr`/`svcopts` reset to 0/false the instant the call ends, while
+    but one reported "-1.61s lead, -0.00s tail" — the trim found nothing to
-          `tgid`/`hold_tgid` keep showing the just-ended talkgroup for
+    remove because the capture window had closed on top of live speech. The
-          TGID_HOLD_TIME (2 s). So the srcaddr edge is a real state change, not a
+    recording ends on an unfinished word. Working backwards from its lead trim,
-          timeout heuristic.
+    the audio pipeline lag was at most 1.36 s, so the window should have held
+    ~1.6 s more; the only consistent explanation is a false early `srcaddr -> 0`.
+
+Audio is the ground truth for WHEN. It cannot answer WHO, so the console is
+still the only source of talkgroup, alias and radio id.
+
+WHY ATTRIBUTION HAPPENS AT CLOSE, NOT AT OPEN. There is no guaranteed ordering
+between a grant and the audio it belongs to: the console is polled every 500 ms
+and the audio pipeline lag is variable, so the grant can land after voice onset
+just as easily as before it. A segment may therefore open unattributed and
+acquire its talkgroup part-way through, which is expected and fine. At close we
+have seen the whole window and ask the rolling console history "what was active
+during this audio, give or take a few seconds" — see _attribute and the
+ATTRIBUTION_* constants.
+
+ORPHAN AUDIO. If nothing in the console history overlaps the window, the audio
+is unattributed: Liquidsoap fallback, a test tone, stray noise, or a dropped
+`call_log`. Policy is DISCARD AND SHOUT — the recording is not uploaded and no
+call_start/call_end is published, because a call with no talkgroup silently
+poisons incident correlation downstream, and that is worse than losing the
+audio. It is logged at ERROR with the window and everything nearby that was
+considered, and counted on /api/status so it cannot pass unnoticed.
+
+FALLBACK MODE. When PulseAudio capture is NOT producing audio there is nothing
+to segment on, so the old console state machine still runs (grant opens,
+srcaddr edge + call_idle_timeout closes). It produces no audio — capture is
+down — but it keeps the node reporting real radio activity to C2 while the
+audio path is broken. This is the only remaining consumer of
+settings.call_idle_timeout.
 
 SEGMENTS: one emitted call (= one recording, one Firestore doc) spans a whole
 conversation, not a single transmission. It stays open across repeated grants on
-the same talkgroup and closes when the talkgroup changes or the radio goes quiet
+the same talkgroup and closes when the talkgroup changes or the AUDIO goes quiet
-for settings.call_idle_timeout seconds.
+for settings.call_silence_timeout seconds.
 
 CLOCKS: `call_log["time"]` is time.time() inside the op25 container. All three
 client containers run network_mode: host and share the host kernel clock, so that
 value is directly comparable to time.time() here — no offset mapping needed. The
-call recorder's ring buffer is stamped with the same clock for the same reason.
+call recorder's chunk timestamps are the same clock, with the caveat that they
+are ARRIVAL times and therefore lag the moment of speech by the pipeline
+latency. Comparisons between two audio timestamps are exact; comparisons between
+audio and console timestamps carry that lag, which is what _tail_pad() covers.
 """
 import asyncio
 import time
 import uuid
+from collections import deque
+from dataclasses import dataclass, field
 from datetime import datetime, timezone
 from typing import Optional, Callable, Awaitable, Any, List, Dict
 
 from app.config import settings
+from app.internal.call_recorder import AudioActivity
 from app.internal.op25_client import op25_client
 from app.internal.logger import logger
 
 CallbackFn = Callable[[dict], Awaitable[None]]
+ActivityFn = Callable[[], AudioActivity]
 
-# 500 ms. Do NOT lower: start precision already comes from OP25's own timestamp,
+# 500 ms. Do NOT lower: audio boundaries come from the recorder's own chunk
-# and http_server.py's request handler has a ~200 ms blocking floor anyway.
+# timestamps (~46 ms resolution), not from when this loop happens to notice
+# them, and http_server.py's request handler has a ~200 ms blocking floor anyway.
 POLL_INTERVAL = 0.5
 
-# Seconds of unreachable OP25 before an open segment is force-closed.
+# Seconds of unreachable OP25 before an open segment is force-closed. Applies in
+# both modes: without the console there is no attribution, and unattributed
+# audio is discarded anyway.
 OP25_OFFLINE_GRACE = 3.0
 
 # Hard ceiling on a single segment; mirrors MAX_RECORDING_SECONDS in call_recorder
-# so a talkgroup that never goes quiet cannot produce an unbounded recording.
+# so a talkgroup that never goes quiet cannot produce an unbounded recording. In
+# audio mode a new segment is opened immediately afterwards if voice is still
+# present, so a genuinely long transmission is split rather than truncated.
 MAX_SEGMENT_SECONDS = 600
 
+# How far either side of the AUDIO window console observations are still
+# accepted as attribution evidence. "Plus or minus some seconds", made explicit:
+#
+#   LOOKBACK  the grant normally PRECEDES the audio — 0.84-1.62 s of
+#             grant-to-speech delay, plus up to ~1.4 s of audio pipeline lag,
+#             plus one 0.5 s poll of detection slack. 4.0 s covers the worst
+#             case measured with margin.
+#   LOOKAHEAD the grant can also FOLLOW voice onset, because the console is only
+#             polled every 500 ms and OP25 logs the grant on its own schedule.
+#             2.0 s is four poll intervals.
+#
+# Both are deliberately asymmetric: the "grant first" direction is the common
+# one and has the larger physical spread.
+ATTRIBUTION_LOOKBACK_SECONDS = 4.0
+ATTRIBUTION_LOOKAHEAD_SECONDS = 2.0
+
+# Rolling console history. Bounded twice — by age and by entry count — so a busy
+# system cannot grow it without limit. At ~2 observations per poll this is a few
+# minutes of history for a few tens of KB.
+CONSOLE_HISTORY_SECONDS = 180.0
+CONSOLE_HISTORY_MAX = 1200
+
+# How far back to look for a duplicate before appending a grant. OP25's call_log
+# deque drains on read so repeats should not happen, but a re-delivered entry
+# would otherwise inflate the transmission count and the attribution score.
+_GRANT_DEDUPE_DEPTH = 24
+
+# Close reasons where the console explicitly told us the talkgroup changed, so
+# the segment's label is already known first-hand and close-time attribution
+# would only be able to make it worse (the window extends past the split).
+_SPLIT_REASONS = ("tgid_change", "tgid_change_unlogged")
+
 
 def _tail_pad() -> float:
     """
-    Audio kept after the observed end of the last transmission, so the srcaddr
+    Audio kept past a CONSOLE-DERIVED segment boundary, to cover the fact that
-    edge (up to one poll late) plus encoder latency never clips the tail.
+    buffered audio lags control-channel timestamps.
 
-    Read live from settings (env CALL_TAIL_PAD_SECONDS) rather than frozen into a
+    Under audio-driven segmentation this no longer applies to the normal end of
-    module constant, so it is tunable per node. See the setting in config.py for
+    a call — that boundary now comes from the audio itself and needs no pad. It
-    why the default moved 1.0 → 3.0 (short calls' recording window was closing
+    still applies wherever a boundary is a control-channel timestamp:
-    before the ~1.5s grant→speech offset let voice audio even start).
 
-    All three close paths use this pad: idle_timeout, tgid_change, and
+        tgid_change            close at the new grant's timestamp + pad
-    tgid_change_unlogged. An earlier version of this docstring claimed the
+        tgid_change_unlogged   close at the observing poll's timestamp + pad
-    latter two close at "an exact, already-known boundary" (the new grant's
+        idle_timeout           console fallback mode only
-    timestamp, or the same poll tick) and so intentionally added no pad — THAT
+
+    Read live from settings (env CALL_TAIL_PAD_SECONDS) rather than frozen into
+    a module constant, so it is tunable per node.
+
+    An earlier version of this docstring claimed the tgid_change paths close at
+    "an exact, already-known boundary" and so intentionally added no pad — THAT
     REASONING WAS WRONG and produced real truncated recordings. The boundary is
     exact only in CONTROL-CHANNEL time; the buffered AUDIO lags control-channel
     timestamps by ~1.5 s (measured: 0.84-1.62 s of lead trimmed across 7 field
     calls), so slicing the outgoing call at the new grant's exact timestamp cut
-    roughly the last 1.5s of its real speech — calls ending mid-word with ~0s
+    roughly the last 1.5 s of its real speech. Do not reintroduce a zero-pad
-    trailing silence. Do not reintroduce a zero-pad close for tgid_change or
+    close for tgid_change or tgid_change_unlogged; if the outgoing and incoming
-    tgid_change_unlogged; if the outgoing and incoming recordings end up
+    recordings end up overlapping in the underlying audio because of this pad,
-    overlapping in the underlying audio because of this pad, that is correct —
+    that is correct — the audio genuinely contains both.
-    the audio genuinely contains both. See _handle_call_log and
-    _handle_channels for how each path sources the timestamp this gets added to.
     """
     return settings.call_tail_pad_seconds
 
@@ -104,6 +183,35 @@ def _iso(epoch: Optional[float]) -> Optional[str]:
     return datetime.fromtimestamp(epoch, timezone.utc).isoformat()
 
 
+@dataclass(frozen=True)
+class ConsoleEvent:
+    """One thing the OP25 console said, kept so a closing segment can ask about it."""
+
+    epoch: float
+    tgid: int
+    name: str = ""
+    freq: Any = None
+    rid: Optional[int] = None
+    # True for a `call_log` grant, False for an active `channel_update` row.
+    is_grant: bool = False
+
+
+@dataclass
+class Attribution:
+    """Who a stretch of audio belonged to, and how confident we are."""
+
+    tgid: int
+    name: str = ""
+    freq: Any = None
+    rid: Optional[int] = None
+    grants: int = 0
+    # Observations that fall strictly inside the audio window (vs only inside
+    # the tolerance band around it).
+    overlap: int = 0
+    nearby: int = 0
+    competing: List[int] = field(default_factory=list)
+
+
 class MetadataWatcher:
     def __init__(self):
         self._running = False
@@ -114,21 +222,38 @@ class MetadataWatcher:
         self._current_tgid_name: Optional[str] = None
         self._current_freq: Any = None
         self._current_srcaddr: Optional[int] = None
-        self._started_at: Optional[float] = None   # OP25 epoch of the first grant
+        self._started_at: Optional[float] = None   # audio onset, or grant epoch in fallback mode
         self._transmissions: int = 0
+        # True when the open segment is governed by audio, False for the
+        # console fallback. Fixed at open so capture flapping cannot switch the
+        # rules underneath a live segment.
+        self._audio_driven: bool = False
 
-        # Transmission tracking within the open segment
+        # Transmission tracking within the open segment (console fallback mode)
         self._tx_active: bool = False              # last poll saw srcaddr != 0
         self._last_activity: float = 0.0           # epoch of last evidence of traffic
         self._last_tx_end: Optional[float] = None  # epoch of the srcaddr 1→0 edge
         self._last_ok_poll: float = 0.0
 
+        # Rolling console history for close-time attribution.
+        self._console: deque[ConsoleEvent] = deque(maxlen=CONSOLE_HISTORY_MAX)
+
+        # Onset of the voice run the last audio-driven segment covered, so the
+        # same run cannot immediately re-open a second segment.
+        self._consumed_onset: Optional[float] = None
+
+        # Field-visible counter of discarded orphan audio.
+        self._unattributed_segments: int = 0
+
         # Injectable for tests; production is always the host wall clock.
         self._clock: Callable[[], float] = time.time
 
         # Set these before calling start()
         self.on_call_start: Optional[CallbackFn] = None
         self.on_call_end: Optional[CallbackFn] = None
+        # Supplies the audio-activity snapshot. None (or a snapshot reporting
+        # capturing=False) puts the watcher in console fallback mode.
+        self.audio_activity: Optional[ActivityFn] = None
 
     # ------------------------------------------------------------------
     # Lifecycle
@@ -138,7 +263,7 @@ class MetadataWatcher:
         self._running = True
         self._last_ok_poll = self._clock()
         asyncio.create_task(self._poll_loop())
-        logger.info("Metadata watcher started (call_log driven).")
+        logger.info("Metadata watcher started (audio-driven segmentation, console attribution).")
 
     async def stop(self):
         self._running = False
@@ -168,6 +293,307 @@ class MetadataWatcher:
             return
 
         self._last_ok_poll = now
+        self._record_console(update, now)
+
+        activity = self._snapshot()
+        if activity is None or not activity.capturing:
+            await self._console_tick(update, now)
+            return
+
+        await self._audio_tick(update, activity, now)
+
+    def _snapshot(self) -> Optional[AudioActivity]:
+        if self.audio_activity is None:
+            return None
+        try:
+            return self.audio_activity()
+        except Exception as e:
+            logger.warning(f"Audio activity unavailable ({e}) — falling back to console segmentation.")
+            return None
+
+    # ------------------------------------------------------------------
+    # Console history (feeds close-time attribution)
+    # ------------------------------------------------------------------
+
+    def _record_console(self, update: Any, now: float) -> None:
+        for entry in update.call_log:
+            tgid = _as_int(entry.get("tgid"))
+            if tgid is None:
+                continue
+            epoch = _as_float(entry.get("time"))
+            event = ConsoleEvent(
+                epoch=now if epoch is None else epoch,
+                tgid=tgid,
+                name=entry.get("tgtag") or "",
+                freq=entry.get("freq"),
+                rid=_as_int(entry.get("rid")),
+                is_grant=True,
+            )
+            if not self._is_duplicate_grant(event):
+                self._console.append(event)
+
+        for channel in update.channels:
+            tgid = _as_int(channel.get("tgid"))
+            srcaddr = _as_int(channel.get("srcaddr"))
+            if tgid is None or srcaddr is None:
+                continue  # idle channel says nothing about who is talking
+            self._console.append(ConsoleEvent(
+                epoch=now,
+                tgid=tgid,
+                name=channel.get("tag") or "",
+                freq=channel.get("freq"),
+                rid=srcaddr,
+                is_grant=False,
+            ))
+
+        cutoff = now - CONSOLE_HISTORY_SECONDS
+        while self._console and self._console[0].epoch < cutoff:
+            self._console.popleft()
+
+    def _is_duplicate_grant(self, event: ConsoleEvent) -> bool:
+        for index in range(len(self._console) - 1, -1, -1):
+            if len(self._console) - index > _GRANT_DEDUPE_DEPTH:
+                return False
+            known = self._console[index]
+            if known.is_grant and known.tgid == event.tgid and known.epoch == event.epoch:
+                return True
+        return False
+
+    def _attribute(self, start: float, end: float) -> Optional[Attribution]:
+        """
+        Resolve which talkgroup a stretch of audio belongs to.
+
+        Scores every talkgroup seen in [start - LOOKBACK, end + LOOKAHEAD] by
+        how well its console activity overlaps the audio itself, preferring
+        real overlap over merely being nearby, and grants over channel rows.
+        Returns None only when NOTHING was observed in that band at all — the
+        orphan-audio case.
+        """
+        low = start - ATTRIBUTION_LOOKBACK_SECONDS
+        high = end + ATTRIBUTION_LOOKAHEAD_SECONDS
+        candidates: Dict[int, Attribution] = {}
+        firsts: Dict[int, float] = {}
+
+        for event in self._console:
+            if event.epoch < low or event.epoch > high:
+                continue
+            found = candidates.get(event.tgid)
+            if found is None:
+                found = Attribution(tgid=event.tgid)
+                candidates[event.tgid] = found
+                firsts[event.tgid] = event.epoch
+            found.nearby += 1
+            if start <= event.epoch <= end:
+                found.overlap += 1
+            if event.is_grant:
+                found.grants += 1
+            if event.name and not found.name:
+                found.name = event.name
+            if event.freq and found.freq is None:
+                found.freq = event.freq
+            if event.rid is not None:
+                found.rid = event.rid  # most recent wins
+
+        if not candidates:
+            return None
+
+        best = max(
+            candidates.values(),
+            key=lambda a: (a.overlap, a.grants, a.nearby, -firsts[a.tgid]),
+        )
+        best.competing = sorted(
+            tgid for tgid, a in candidates.items() if tgid != best.tgid and a.overlap > 0
+        )
+        return best
+
+    # ------------------------------------------------------------------
+    # Audio-driven segmentation
+    # ------------------------------------------------------------------
+
+    async def _audio_tick(self, update: Any, activity: AudioActivity, now: float) -> None:
+        if self._active_call_id is not None and not self._audio_driven:
+            # A segment that opened while capture was down finishes under the
+            # rules it started with rather than switching mid-flight.
+            await self._console_tick(update, now)
+            return
+
+        # 1. Console first: a talkgroup change must still force a split even
+        #    when the audio never went quiet, and a grant may be the thing that
+        #    finally attributes an already-open segment.
+        for entry in sorted(update.call_log, key=lambda e: _as_float(e.get("time")) or 0.0):
+            await self._handle_grant(entry, now)
+        await self._scan_channels(update.channels, now)
+
+        # 2. Then the audio decides the boundaries.
+        last_voice = activity.last_voice_epoch
+        voice_active = last_voice is not None and (now - last_voice) < settings.call_silence_timeout
+
+        if self._active_call_id is None:
+            onset = activity.voice_onset_epoch
+            if voice_active and onset is not None and (
+                self._consumed_onset is None or onset > self._consumed_onset
+            ):
+                await self._open_from_audio(onset, now)
+            return
+
+        if not voice_active:
+            silence = (now - last_voice) if last_voice is not None else settings.call_silence_timeout
+            # The measured trailing silence, in the AUDIO's own clock. This is
+            # the number to tune settings.call_silence_timeout from — unlike the
+            # old control-channel idle it contains no grant-to-speech delay, so
+            # it means exactly what it says.
+            logger.info(
+                f"Audio silence close for tgid {self._current_tgid}: measured trailing silence "
+                f"{silence:.2f}s (threshold {settings.call_silence_timeout:.2f}s at "
+                f"{settings.call_silence_threshold_db:.1f}dBFS)."
+            )
+            self._consumed_onset = activity.voice_onset_epoch
+            end = (last_voice + settings.call_silence_timeout) if last_voice is not None else now
+            await self._close_segment(min(end, now), reason="audio_silence")
+            return
+
+        if self._started_at is not None and (now - self._started_at) >= MAX_SEGMENT_SECONDS:
+            logger.warning(
+                f"Segment for tgid {self._current_tgid} hit the {MAX_SEGMENT_SECONDS}s cap while audio "
+                "was still live — closing and immediately reopening so nothing is dropped. If this "
+                "repeats, the silence threshold may be low enough that noise reads as voice."
+            )
+            await self._close_segment(now, reason="max_length")
+            await self._open_from_audio(now, now)
+
+    async def _open_from_audio(self, onset: float, now: float) -> None:
+        """Open a segment at a detected voice onset, attributing it if we can."""
+        found = self._attribute(onset, now)
+        await self._open_segment(
+            started_at=onset,
+            now=now,
+            tgid=found.tgid if found else None,
+            tgid_name=found.name if found else "",
+            freq=found.freq if found else None,
+            srcaddr=found.rid if found else None,
+            audio_driven=True,
+            transmissions=found.grants if found else 0,
+        )
+
+    async def _handle_grant(self, entry: Dict[str, Any], now: float) -> None:
+        """A `call_log` grant, interpreted in audio mode: label or split, never start."""
+        tgid = _as_int(entry.get("tgid"))
+        if tgid is None:
+            return  # a grant with no talkgroup is nothing we can label with
+
+        started_at = _as_float(entry.get("time"))
+        if started_at is None:
+            logger.warning(f"call_log entry for tgid={tgid} has no usable time — using local clock.")
+            started_at = now
+
+        if self._active_call_id is None:
+            # Audio starts recordings, not grants. The grant is already in the
+            # console history and will attribute the segment when audio arrives.
+            return
+
+        if self._current_tgid is None:
+            self._current_tgid = tgid
+            self._current_tgid_name = entry.get("tgtag") or ""
+            self._current_freq = entry.get("freq")
+            self._current_srcaddr = _as_int(entry.get("rid"))
+            self._transmissions += 1
+            logger.info(
+                f"Late attribution: segment {self._active_call_id} adopted tgid {tgid} from a grant "
+                f"logged {started_at - (self._started_at or started_at):+.2f}s from audio onset."
+            )
+            return
+
+        if tgid == self._current_tgid:
+            # CONTINUE: same talkgroup, keep one recording so the back-and-forth
+            # of a single conversation lands in one file.
+            self._transmissions += 1
+            self._refresh_meta_from_log(entry)
+            return
+
+        # FORCED SPLIT. Two talkgroups can be back to back with no silence
+        # between them; pure audio segmentation would merge them into one file
+        # under one label, which is exactly the kind of wrong that corrupts
+        # incident correlation. The console change is authoritative here.
+        await self._close_segment(started_at + _tail_pad(), reason="tgid_change")
+        await self._open_segment(
+            started_at=started_at,
+            now=now,
+            tgid=tgid,
+            tgid_name=entry.get("tgtag") or "",
+            freq=entry.get("freq"),
+            srcaddr=_as_int(entry.get("rid")),
+            audio_driven=True,
+            transmissions=1,
+        )
+
+    async def _scan_channels(self, channels: List[Dict[str, Any]], now: float) -> None:
+        """Channel rows in audio mode: refresh metadata, catch an unlogged split."""
+        if self._active_call_id is None:
+            return
+
+        active: List[Dict[str, Any]] = []
+        ours = False
+        for channel in channels:
+            tgid = _as_int(channel.get("tgid"))
+            srcaddr = _as_int(channel.get("srcaddr"))
+            if tgid is None or srcaddr is None:
+                continue
+            active.append(channel)
+            if tgid == self._current_tgid:
+                ours = True
+                self._current_srcaddr = srcaddr
+                self._last_activity = now
+                self._refresh_meta_from_channel(channel)
+
+        if self._current_tgid is None:
+            # Late attribution from a channel row — this is the path that saves
+            # us when the grant itself was dropped from OP25's capped deque.
+            if len(active) == 1:
+                tgid = _as_int(active[0].get("tgid"))
+                self._current_tgid = tgid
+                self._current_tgid_name = active[0].get("tag") or ""
+                self._current_freq = active[0].get("freq")
+                self._current_srcaddr = _as_int(active[0].get("srcaddr"))
+                logger.info(f"Late attribution: segment {self._active_call_id} adopted tgid {tgid} from channel state.")
+            return
+
+        if ours or not active or len(channels) != 1:
+            # Restricted to single-receiver setups on purpose: with several
+            # receivers, another channel being busy says nothing about ours.
+            return
+
+        foreign = _as_int(active[0].get("tgid"))
+        if foreign is None or foreign == self._current_tgid:
+            return
+
+        logger.warning(
+            f"tgid {foreign} active without a call_log entry — splitting segment for tgid "
+            f"{self._current_tgid} (call_log event likely dropped)."
+        )
+        await self._close_segment(now + _tail_pad(), reason="tgid_change_unlogged")
+        await self._open_segment(
+            started_at=now,
+            now=now,
+            tgid=foreign,
+            tgid_name=active[0].get("tag") or "",
+            freq=active[0].get("freq"),
+            srcaddr=_as_int(active[0].get("srcaddr")),
+            audio_driven=True,
+            transmissions=1,
+        )
+
+    # ------------------------------------------------------------------
+    # Console fallback segmentation (capture down)
+    # ------------------------------------------------------------------
+
+    async def _console_tick(self, update: Any, now: float) -> None:
+        if self._active_call_id is not None and self._audio_driven:
+            logger.warning(
+                f"PulseAudio capture stopped while recording {self._active_call_id} — closing the "
+                "segment at the last captured audio; segmentation falls back to the control channel."
+            )
+            await self._close_segment(now, reason="capture_lost")
+            return
 
         # 1. call_log first — these are the authoritative starts, and processing
         #    them before the channel scan means a same-poll grant+state pair is
@@ -176,27 +602,24 @@ class MetadataWatcher:
         for entry in sorted(update.call_log, key=lambda e: _as_float(e.get("time")) or 0.0):
             await self._handle_call_log(entry, now)
 
-        # 2. channel_update — the only external end signal.
+        # 2. channel_update — the only external end signal available here.
         await self._handle_channels(update.channels, now)
 
     async def _handle_call_log(self, entry: Dict[str, Any], now: float) -> None:
         tgid = _as_int(entry.get("tgid"))
         if tgid is None:
-            return  # a grant with no talkgroup is nothing we can record or label
+            return
 
-        # OP25's own stamp. Fall back to now only if the field is missing/garbage.
         started_at = _as_float(entry.get("time"))
         if started_at is None:
             logger.warning(f"call_log entry for tgid={tgid} has no usable time — using local clock.")
             started_at = now
 
         if self._active_call_id is None:
-            await self._open_segment(entry, tgid, started_at, now)
+            await self._open_from_console(entry, tgid, started_at, now)
             return
 
         if tgid == self._current_tgid:
-            # CONTINUE: same talkgroup, keep one recording so the back-and-forth
-            # of a single conversation lands in one file.
             self._transmissions += 1
             self._tx_active = True
             self._last_tx_end = None
@@ -204,18 +627,8 @@ class MetadataWatcher:
             self._refresh_meta_from_log(entry)
             return
 
-        # SPLIT: different talkgroup. The new grant's OP25 timestamp is the most
-        # precise CONTROL-CHANNEL end for the outgoing segment, but the buffered
-        # audio lags control by ~1.5s, so slicing exactly there cut the outgoing
-        # call's last words. Pad past it and let the recorder's bounded tail wait
-        # block until that audio has actually been captured.
-        #
-        # The incoming call's pre-roll comes from the ring buffer, so the delay
-        # costs it nothing, and the two slices overlapping in the underlying
-        # audio is correct — the stream genuinely contains one call's tail and
-        # then the next call's start.
         await self._close_segment(started_at + _tail_pad(), reason="tgid_change")
-        await self._open_segment(entry, tgid, started_at, now)
+        await self._open_from_console(entry, tgid, started_at, now)
 
     async def _handle_channels(self, channels: List[Dict[str, Any]], now: float) -> None:
         if self._active_call_id is None:
@@ -241,19 +654,13 @@ class MetadataWatcher:
             self._last_tx_end = None
             self._last_activity = now
         elif self._tx_active:
-            # The srcaddr != 0 → 0 edge: OP25 has torn the call down.
+            # The srcaddr != 0 → 0 edge. Note this is NOT trusted as an end of
+            # speech any more (it fires mid-word in the field) — in fallback
+            # mode there is simply nothing better available.
             self._tx_active = False
             self._last_tx_end = now
             self._last_activity = now
 
-        # Safety net for a dropped call_log event (deque is capped at 10): the one
-        # receiver we have is plainly on another talkgroup, so our segment is over
-        # even though we never saw its grant. Close now rather than record
-        # call_idle_timeout seconds of the wrong tgid.
-        #
-        # Restricted to single-receiver setups on purpose: with several receivers,
-        # another channel being busy says nothing about ours, and closing on it
-        # would truncate every call whenever a second receiver is active.
         if not tx_active and foreign_active_tgid is not None and len(channels) == 1:
             logger.warning(
                 f"tgid {foreign_active_tgid} active without a call_log entry — "
@@ -263,14 +670,6 @@ class MetadataWatcher:
             return
 
         if (now - self._last_activity) >= settings.call_idle_timeout:
-            # STOP: quiet for long enough. End the audio at the last transmission
-            # plus a short pad, not at "now" — otherwise every recording carries
-            # call_idle_timeout seconds of silence.
-            #
-            # The measured idle below is the CONTROL-CHANNEL idle (srcaddr 1→0
-            # edge → now). It is NOT comparable to silence measured in the audio,
-            # which additionally contains the ~1.9 s P25 grant→speech delay.
-            # Tune settings.call_idle_timeout from THIS number and nothing else.
             if self._last_tx_end is not None:
                 measured_idle = now - self._last_tx_end
                 end = self._last_tx_end + _tail_pad()
@@ -292,6 +691,18 @@ class MetadataWatcher:
             logger.warning(f"Segment for tgid {self._current_tgid} hit the {MAX_SEGMENT_SECONDS}s cap — closing.")
             await self._close_segment(now, reason="max_length")
 
+    async def _open_from_console(self, entry: Dict[str, Any], tgid: int, started_at: float, now: float) -> None:
+        await self._open_segment(
+            started_at=started_at,
+            now=now,
+            tgid=tgid,
+            tgid_name=entry.get("tgtag") or "",
+            freq=entry.get("freq"),
+            srcaddr=_as_int(entry.get("rid")),
+            audio_driven=False,
+            transmissions=1,
+        )
+
     # ------------------------------------------------------------------
     # Segment open / close
     # ------------------------------------------------------------------
@@ -311,35 +722,48 @@ class MetadataWatcher:
         if not self._current_freq and channel.get("freq"):
             self._current_freq = channel.get("freq")
 
-    async def _open_segment(self, entry: Dict[str, Any], tgid: int, started_at: float, now: float) -> None:
+    async def _open_segment(
+        self,
+        started_at: float,
+        now: float,
+        tgid: Optional[int],
+        tgid_name: str,
+        freq: Any,
+        srcaddr: Optional[int],
+        audio_driven: bool,
+        transmissions: int = 1,
+    ) -> None:
         self._active_call_id = str(uuid.uuid4())
         self._current_tgid = tgid
-        self._current_tgid_name = entry.get("tgtag") or ""
+        self._current_tgid_name = tgid_name
-        self._current_freq = entry.get("freq")
+        self._current_freq = freq
-        self._current_srcaddr = _as_int(entry.get("rid"))
+        self._current_srcaddr = srcaddr
         self._started_at = started_at
-        self._transmissions = 1
+        self._transmissions = max(1, transmissions)
+        self._audio_driven = audio_driven
 
-        # Assume the transmission is still up: we learn otherwise from the next
+        # Console fallback assumes the transmission is still up; it learns
-        # channel scan. A grant whose call already ended before we polled simply
+        # otherwise from the next channel scan.
-        # closes on the very next tick via the idle timeout.
+        self._tx_active = not audio_driven
-        self._tx_active = True
         self._last_tx_end = None
         self._last_activity = now
 
         payload = {
             "call_id": self._active_call_id,
             "tgid": tgid,
-            "tgid_name": self._current_tgid_name,
+            "tgid_name": tgid_name,
-            "freq": self._current_freq,
+            "freq": freq,
-            "srcaddr": self._current_srcaddr,
+            "srcaddr": srcaddr,
             "started_at": _iso(started_at),
             # Raw epoch for the recorder's ring-buffer slice — same clock domain.
             "started_at_epoch": started_at,
+            "attributed": tgid is not None,
+            "driver": "audio" if audio_driven else "console",
         }
+        source = "audio onset" if audio_driven else "op25 grant"
         logger.info(
             f"Call start: tgid={tgid} id={self._active_call_id} "
-            f"(op25 t={started_at:.3f}, detected {now - started_at:+.2f}s later)"
+            f"({source} t={started_at:.3f}, detected {now - started_at:+.2f}s later)"
         )
         if self.on_call_start:
             await self.on_call_start(payload)
@@ -352,6 +776,10 @@ class MetadataWatcher:
         if started_at is not None:
             end_epoch = max(end_epoch, started_at)
 
+        if self._audio_driven and reason not in _SPLIT_REASONS:
+            self._resolve_attribution(started_at if started_at is not None else end_epoch, end_epoch)
+
+        attributed = self._current_tgid is not None
         payload = {
             "call_id": self._active_call_id,
             "tgid": self._current_tgid,
@@ -364,8 +792,25 @@ class MetadataWatcher:
             "ended_at_epoch": end_epoch,
             "transmissions": self._transmissions,
             "end_reason": reason,
+            "attributed": attributed,
+            "driver": "audio" if self._audio_driven else "console",
         }
         duration = (end_epoch - started_at) if started_at is not None else 0.0
+
+        if not attributed:
+            self._unattributed_segments += 1
+            window_start = started_at if started_at is not None else end_epoch
+            logger.error(
+                f"ORPHAN AUDIO: {duration:.2f}s of audio ({self._active_call_id}, reason={reason}, "
+                f"window {window_start:.3f}-{end_epoch:.3f}) had NO OP25 talkgroup anywhere within "
+                f"{ATTRIBUTION_LOOKBACK_SECONDS:.0f}s before or {ATTRIBUTION_LOOKAHEAD_SECONDS:.0f}s "
+                f"after it. It will be DISCARDED, not uploaded — an untagged call would poison "
+                f"incident correlation. Causes: Liquidsoap fallback/test audio on drb_sink, OP25 not "
+                f"decoding the control channel, or a dropped call_log. Console history holds "
+                f"{len(self._console)} recent observations; total orphans this run: "
+                f"{self._unattributed_segments}."
+            )
+        else:
             logger.info(
                 f"Call end: id={self._active_call_id} tgid={self._current_tgid} "
                 f"reason={reason} transmissions={self._transmissions} duration={duration:.2f}s"
@@ -382,10 +827,57 @@ class MetadataWatcher:
         self._transmissions = 0
         self._tx_active = False
         self._last_tx_end = None
+        self._audio_driven = False
 
         if self.on_call_end:
             await self.on_call_end(payload)
 
+    def _resolve_attribution(self, start: float, end: float) -> None:
+        """
+        Last chance to label an audio-driven segment, run at close.
+
+        Only ADOPTS a talkgroup when the segment still has none. A tgid we
+        already hold came from a grant or a channel row — the console stating
+        outright who was transmitting — and an inference over a window is not
+        allowed to overrule a direct statement. This matters because the window
+        deliberately extends past the audio (ATTRIBUTION_LOOKAHEAD_SECONDS, and
+        the tail pad on a split), so a neighbouring call's console activity can
+        legitimately fall inside it.
+
+        A disagreement is still worth knowing about, so it is logged: it means
+        two talkgroups' console activity overlaps one recording, i.e. the split
+        logic should have fired and did not.
+        """
+        found = self._attribute(start, end)
+        if found is None:
+            return
+
+        if self._current_tgid is None:
+            logger.info(
+                f"Attributed {self._active_call_id} at close to tgid {found.tgid} "
+                f"(overlap {found.overlap}, grants {found.grants}, nearby {found.nearby})."
+            )
+            self._current_tgid = found.tgid
+            if found.name:
+                self._current_tgid_name = found.name
+            if found.freq is not None and not self._current_freq:
+                self._current_freq = found.freq
+            if found.rid is not None and self._current_srcaddr is None:
+                self._current_srcaddr = found.rid
+            self._transmissions = max(self._transmissions, found.grants)
+            return
+
+        others = sorted(set(found.competing) | ({found.tgid} if found.tgid != self._current_tgid else set()))
+        others = [tgid for tgid in others if tgid != self._current_tgid]
+        if others:
+            logger.warning(
+                f"Segment {self._active_call_id} (tgid {self._current_tgid}) overlaps console "
+                f"activity for {others} as well — the split logic should have fired and did not. "
+                "Keeping the talkgroup the console stated directly."
+            )
+        if not self._current_tgid_name and found.tgid == self._current_tgid and found.name:
+            self._current_tgid_name = found.name
+
     # ------------------------------------------------------------------
     # Public state (consumed by routers/api.py, main.py and the dashboards)
     # ------------------------------------------------------------------
@@ -406,5 +898,10 @@ class MetadataWatcher:
     def is_active(self) -> bool:
         return self._active_call_id is not None
 
+    @property
+    def unattributed_segments(self) -> int:
+        """Orphan-audio segments discarded since start. Surfaced on /api/status."""
+        return self._unattributed_segments
+
 
 metadata_watcher = MetadataWatcher()

drb-edge-node/app/internal/pcm.py

@@ -0,0 +1,131 @@
+"""
+Raw PCM primitives: the one place that knows the capture format.
+
+The capture pipeline buffers RAW PCM (signed 16-bit little-endian, mono,
+22050 Hz) instead of MP3. Three things fall out of that, and they are the whole
+reason for the change:
+
+  1. Silence detection is integer arithmetic over the bytes as they arrive —
+     no decode, no FFmpeg, no second process. That is what makes an
+     AUDIO-DRIVEN call boundary possible at all.
+  2. Trimming becomes a byte-offset slice instead of a second encode pass.
+  3. MP3 encoding happens exactly ONCE, at save time, so uploads stop being
+     double-encoded.
+
+WHY SILENCE IS UNAMBIGUOUS HERE: between transmissions the captured stream is
+the monitor of a PulseAudio *null sink*, which emits digital silence, not an
+analog noise floor. Measured on a live node, the gap between transmissions sits
+at about -91 dBFS — that is 20*log10(1/32768), i.e. one least-significant bit,
+the quietest thing a 16-bit sample can be without being exactly zero. Speech on
+the same node averages about -18 dBFS. There is therefore ~70 dB of daylight
+between "silence" and "voice", and the threshold does NOT need field
+calibration against radio noise the way an analog squelch tail would.
+
+MEASUREMENT IS RMS, NOT PEAK. Peak would be cheaper but a single decoder click
+would read as voice for a whole window; RMS over a window is the honest
+"is there signal here" answer. The cost is a Python loop over the window's
+samples, which is affordable because of how little audio is ever scanned:
+one ~46 ms chunk per chunk arrival at capture time, and only the head/tail of a
+finished recording at trim time (see audio_trim.MAX_SCAN_SECONDS). A cheap
+all-zero fast path in C skips the loop entirely for exactly-silent windows.
+
+BYTE ORDER: FFmpeg is asked for s16le. `array("h")` is native-endian, so on a
+big-endian host the samples are byte-swapped before use. Every DRB target is
+little-endian today; this is three lines of insurance, not a real scenario.
+"""
+import math
+import sys
+from array import array
+from typing import Union
+
+# Capture format. MP3_SAMPLE_RATE in call_recorder must stay equal to
+# SAMPLE_RATE — the encode at save time is a straight pass with no resample.
+SAMPLE_RATE = 22050
+SAMPLE_WIDTH = 2
+CHANNELS = 1
+FRAME_BYTES = SAMPLE_WIDTH * CHANNELS
+BYTES_PER_SECOND = SAMPLE_RATE * FRAME_BYTES  # 44100 B/s
+
+# 16-bit full scale. A sample of 32768 (or -32768) is 0 dBFS.
+FULL_SCALE = 32768.0
+
+# Reported for a window with no signal at all. Any real threshold is far above
+# this, so it always compares as "silent" without special-casing log10(0).
+SILENT_DBFS = -120.0
+
+_NEEDS_BYTESWAP = sys.byteorder != "little"
+
+Buffer = Union[bytes, bytearray]
+
+
+def align(nbytes: int) -> int:
+    """Round a byte count DOWN to a whole number of samples."""
+    if nbytes <= 0:
+        return 0
+    return nbytes - (nbytes % FRAME_BYTES)
+
+
+def seconds(nbytes: int) -> float:
+    """Duration of `nbytes` of PCM."""
+    return nbytes / BYTES_PER_SECOND
+
+
+def byte_offset(sec: float) -> int:
+    """Sample-aligned byte offset of `sec` seconds into a PCM buffer."""
+    return align(int(sec * BYTES_PER_SECOND))
+
+
+def samples(buf: Buffer) -> array:
+    """View a PCM buffer as signed 16-bit samples, dropping any partial frame."""
+    usable = align(len(buf))
+    data = array("h")
+    if usable:
+        data.frombytes(bytes(buf[:usable]))
+        if _NEEDS_BYTESWAP:
+            data.byteswap()
+    return data
+
+
+def is_all_zero(buf: Buffer) -> bool:
+    """
+    True when every byte is zero — exact digital silence.
+
+    `bytes.count` runs in C, so this is the cheap path that lets a long scan
+    over silence stay fast without touching the per-sample loop below.
+    """
+    return len(buf) > 0 and buf.count(0) == len(buf)
+
+
+def rms(buf: Buffer) -> float:
+    """Root-mean-square amplitude in raw sample units (0 .. 32768)."""
+    data = samples(buf)
+    if not data:
+        return 0.0
+    total = 0
+    for sample in data:
+        total += sample * sample
+    return math.sqrt(total / len(data))
+
+
+def rms_dbfs(buf: Buffer) -> float:
+    """RMS level of a PCM window in dBFS. SILENT_DBFS for an empty/zero window."""
+    if not buf or is_all_zero(buf):
+        return SILENT_DBFS
+    value = rms(buf)
+    if value <= 0.0:
+        return SILENT_DBFS
+    return 20.0 * math.log10(min(value, FULL_SCALE) / FULL_SCALE)
+
+
+def is_silent(buf: Buffer, threshold_db: float) -> bool:
+    """
+    True when a PCM window carries no signal above `threshold_db` (dBFS RMS).
+
+    An empty buffer counts as silence: "no audio arrived" must never read as
+    "someone is talking", or a stalled capture would hold a segment open.
+    """
+    if not buf:
+        return True
+    if is_all_zero(buf):
+        return True
+    return rms_dbfs(buf) < threshold_db

drb-edge-node/app/main.py

@@ -30,21 +30,56 @@ def _iso(epoch: Optional[float]) -> Optional[str]:
     return datetime.fromtimestamp(epoch, timezone.utc).isoformat()
 
 
+# call_ids whose `call_start` has already gone out over MQTT. A segment can open
+# before its talkgroup is known (audio onset can precede the OP25 grant), and
+# C2's _on_call_start writes talkgroup_id straight into a new Firestore `calls`
+# doc — publishing early with tgid=None would create a permanently untagged call.
+# So the start is held back until attribution succeeds, and replayed just before
+# the end event if it resolved late.
+_published_starts: set = set()
+
+
 async def on_call_start(data: dict):
     radio_bot.start_stream()
     await mqtt_manager.publish_status("recording")
-    await mqtt_manager.publish_metadata("call_start", data)
+    # started_at_epoch is the detected voice onset (or, in console fallback mode,
-    # started_at_epoch is OP25's own call_log timestamp — the recorder slices the
+    # OP25's call_log timestamp). The recorder slices the ring buffer back to it
-    # ring buffer back to it (minus pre-roll), so however late we detected the
+    # minus the pre-roll, so however late the poll loop noticed, the audio still
-    # grant, the audio still starts in the right place.
+    # starts in the right place.
     await call_recorder.start_recording(
         data["call_id"],
         start_epoch=data.get("started_at_epoch"),
     )
 
+    if data.get("attributed", True):
+        _published_starts.add(data["call_id"])
+        await mqtt_manager.publish_metadata("call_start", data)
+    else:
+        logger.info(
+            f"Call {data['call_id']} started on audio onset with no talkgroup yet — holding the "
+            "call_start event until the console attributes it."
+        )
+
 
 async def on_call_end(data: dict):
     radio_bot.stop_stream()
+    call_id = data["call_id"]
+    published_start = call_id in _published_starts
+    _published_starts.discard(call_id)
+
+    if not data.get("attributed", True):
+        # ORPHAN AUDIO. metadata_watcher has already logged the details at ERROR.
+        # The audio is dropped rather than uploaded: a call with no talkgroup is
+        # worse than no call at all, because it silently poisons correlation.
+        await call_recorder.discard_recording()
+        if published_start:
+            # Should not happen (attribution only ever improves), but if a start
+            # did go out, the doc must not be left hanging in "active".
+            data["audio_skipped"] = "unattributed"
+            await mqtt_manager.publish_metadata("call_end", data)
+        await mqtt_manager.publish_status("online")
+        return
+
     recording = await call_recorder.stop_recording(end_epoch=data.get("ended_at_epoch"))
 
     if recording is not None and recording.path is not None:
@@ -89,6 +124,18 @@ async def on_call_end(data: dict):
             "— PulseAudio capture may be down (check the op25 container and "
             f"the {settings.pulse_source} source)."
         )
+
+    if not published_start:
+        # Attribution arrived after the segment opened. Replay the start so C2
+        # creates the `calls` doc with the right talkgroup before the end event
+        # updates it.
+        start_payload = {
+            key: data[key]
+            for key in ("call_id", "tgid", "tgid_name", "freq", "srcaddr",
+                        "started_at", "started_at_epoch", "attributed", "driver")
+            if key in data
+        }
+        await mqtt_manager.publish_metadata("call_start", start_payload)
     await mqtt_manager.publish_metadata("call_end", data)
     await mqtt_manager.publish_status("online")
 
@@ -225,6 +272,10 @@ async def lifespan(app: FastAPI):
     # Wire callbacks
     metadata_watcher.on_call_start = on_call_start
     metadata_watcher.on_call_end = on_call_end
+    # Segment boundaries come from the audio itself; this is how the watcher
+    # sees it. Without this the watcher falls back to control-channel
+    # segmentation, which is measurably wrong in both directions.
+    metadata_watcher.audio_activity = call_recorder.audio_activity
     mqtt_manager.on_command = on_command
     mqtt_manager.on_config_push = on_config_push
     mqtt_manager.on_api_key = on_api_key

drb-edge-node/app/routers/api.py

@@ -49,9 +49,16 @@ async def get_status():
         "system_name": system_name,
         "is_recording": call_recorder.is_recording,
         # Health of the PulseAudio capture that feeds every recording — the single
-        # most useful signal when recordings come back empty.
+        # most useful signal when recordings come back empty. Segment boundaries
+        # come from this stream, so audio_silence_seconds is also how far the
+        # node currently is from closing whatever it is recording.
         "audio_capture": call_recorder.is_capturing,
         "buffered_seconds": round(call_recorder.buffered_seconds, 1),
+        "audio_silence_seconds": round(call_recorder.audio_activity().silence_seconds, 1),
+        # Audio that was recorded but had no OP25 talkgroup anywhere near it, so
+        # it was discarded rather than uploaded. Non-zero means either the
+        # console is not decoding or something else is feeding drb_sink.
+        "unattributed_segments": metadata_watcher.unattributed_segments,
         "active_tgid": active_tgid,
         "active_tgid_name": active_tgid_name,
         "active_call_id": metadata_watcher.active_call_id,

drb-edge-node/tests/test_audio_trim.py

@@ -1,160 +1,218 @@
 """
-Unit tests for silence-trim decision logic.
+Unit tests for silence trimming, now a byte-offset slice of raw PCM.
 
-`speech_bounds` is pure on purpose so the "what do we keep" decision — the part
+`keep_window` is pure on purpose so the "what do we keep" decision — the part
-that can destroy a transmission if it is wrong — is testable without FFmpeg.
+that can destroy a transmission if it is wrong — stays testable without any
-The numbers below come from ffmpeg silencedetect run against six real recordings
+audio at all. The rest of the file drives the real detector over synthesised
-off a live P25 node: 1.71–2.45 s of leading silence and 0.00–1.11 s trailing.
+buffers shaped like the six real recordings measured off a live P25 node:
+1.71-2.45 s of leading silence and 0.00-1.11 s trailing.
+
+The old implementation shelled out to FFmpeg twice (silencedetect, then a
+re-encode) and these tests parsed its stderr. Both passes are gone; the recorder
+buffers PCM, so detection is arithmetic and the cut is a slice.
 """
+from array import array
+
 import pytest
 
+from app.config import settings
+from app.internal import audio_trim, pcm
 from app.internal.audio_trim import (
     TrimResult,
-    _parse_duration,
+    first_signal_offset,
-    _parse_silences,
+    keep_window,
-    speech_bounds,
+    last_signal_offset,
+    trim_pcm,
 )
 
 GUARD = 0.25
+SPEECH_LEVEL = 4096   # -18 dBFS, the measured field average
+FLOOR_LEVEL = 1       # -90.3 dBFS, the measured digital-silence floor
 
 
-def test_leading_silence_is_trimmed_with_a_guard_margin():
+def speech(seconds: float) -> bytes:
-    # Real shape of file f4bfaa1f: 1.85s lead, 0.34s trail, 4.54s total.
+    count = int(pcm.SAMPLE_RATE * seconds)
-    regions = [(0.0, 1.85), (4.20, None)]
+    return array("h", [SPEECH_LEVEL, -SPEECH_LEVEL] * (count // 2)).tobytes()
-    start, end, all_silence = speech_bounds(regions, duration=4.54, guard=GUARD)
-
-    assert not all_silence
-    assert start == pytest.approx(1.85 - GUARD)
-    assert end == pytest.approx(4.20 + GUARD)
-    # The guard must never eat into detected speech.
-    assert start < 1.85 and end > 4.20
 
 
-def test_guard_margin_never_runs_past_the_file_bounds():
+def silence(seconds: float, level: int = FLOOR_LEVEL) -> bytes:
-    regions = [(0.0, 0.10), (3.95, None)]
+    count = int(pcm.SAMPLE_RATE * seconds)
-    start, end, _ = speech_bounds(regions, duration=4.0, guard=1.0)
+    return array("h", [level, -level] * (count // 2)).tobytes()
-
-    assert start == 0.0
-    assert end == 4.0
-
-
-def test_trailing_silence_is_trimmed_when_ffmpeg_closes_the_region_at_eof():
-    """
-    FFmpeg 6.x flushes a `silence_end` at EOF, so a trailing region looks closed.
-    Treating "no silence_end" as the only trailing signal silently disabled tail
-    trimming entirely — verified against ffmpeg 6.1.1.
-    """
-    # Real ffmpeg 6.1.1 output for a 2s-silence + 1.5s-tone + 1s-silence file.
-    regions = [(0.0, 2.05361), (3.56367, 4.63102)]
-    start, end, all_silence = speech_bounds(regions, duration=4.65, guard=GUARD)
-
-    assert not all_silence
-    assert start == pytest.approx(2.05361 - GUARD)
-    assert end == pytest.approx(3.56367 + GUARD), "the trailing second must be trimmed"
-
-
-def test_all_silence_survives_ffmpeg_closing_the_region_at_eof():
-    # Real ffmpeg 6.1.1 output for a 4s file of pure silence.
-    _, _, all_silence = speech_bounds([(0.0, 4.0)], duration=4.03, guard=GUARD)
-    assert all_silence
-
-
-def test_trailing_silence_that_does_not_reach_eof_is_left_alone():
-    """
-    A silence region with a closing silence_end is an internal pause between
-    transmissions, not dead air at the tail. Trimming it would cut the middle
-    out of a conversation.
-    """
-    regions = [(0.0, 1.9), (5.0, 7.5)]
-    start, end, _ = speech_bounds(regions, duration=12.0, guard=GUARD)
-
-    assert start == pytest.approx(1.9 - GUARD)
-    assert end == 12.0, "an internal pause must not shorten the file"
-
-
-def test_no_silence_detected_keeps_the_whole_file():
-    start, end, all_silence = speech_bounds([], duration=6.0, guard=GUARD)
-
-    assert (start, end) == (0.0, 6.0)
-    assert not all_silence
-
-
-def test_silence_starting_late_is_not_treated_as_leading():
-    """Only a region at the very head counts as leading silence."""
-    regions = [(1.20, 2.00)]
-    start, end, _ = speech_bounds(regions, duration=5.0, guard=GUARD)
-
-    assert start == 0.0, "speech before 1.20s must not be trimmed away"
-    assert end == 5.0
-
-
-def test_all_silence_is_reported_not_trimmed_to_nothing():
-    # One region covering the whole file and running to EOF.
-    regions = [(0.0, None)]
-    start, end, all_silence = speech_bounds(regions, duration=4.0, guard=GUARD)
-
-    assert all_silence
-    assert (start, end) == (0.0, 4.0), "an all-silence file must not become zero-length"
-
-
-def test_all_silence_when_head_and_tail_regions_overlap():
-    regions = [(0.0, 3.2), (3.0, None)]
-    _, _, all_silence = speech_bounds(regions, duration=4.0, guard=GUARD)
-
-    assert all_silence
 
 
 # ---------------------------------------------------------------------------
-# FFmpeg output parsing
+# keep_window — the pure decision
 # ---------------------------------------------------------------------------
 
-# Verbatim shape of ffmpeg 6.1.1 output.
+def test_guard_margin_is_kept_around_detected_speech():
-FFMPEG_STDERR = """
+    guard = pcm.byte_offset(GUARD)
-Input #0, mp3, from '/recordings/x.mp3':
+    start, end = keep_window(
-  Duration: 00:00:04.70, start: 0.050113, bitrate: 16 kb/s
+        first_signal=pcm.byte_offset(1.85),
-[silencedetect @ 0000029160e63f40] silence_start: 0
+        last_signal=pcm.byte_offset(4.20),
-[silencedetect @ 0000029160e63f40] silence_end: 2.05361 | silence_duration: 2.05361
+        total_bytes=pcm.byte_offset(4.54),
-[silencedetect @ 0000029160e63f40] silence_start: 3.56367
+        guard_bytes=guard,
-[silencedetect @ 0000029160e63f40] silence_end: 4.63102 | silence_duration: 1.06735
+    )
-[out#0/null @ 0x2] video:0kB audio:97kB
+    assert pcm.seconds(start) == pytest.approx(1.85 - GUARD, abs=0.001)
+    assert pcm.seconds(end) == pytest.approx(4.20 + GUARD, abs=0.001)
+
+
+def test_guard_margin_never_runs_past_the_buffer_bounds():
+    total = pcm.byte_offset(4.0)
+    start, end = keep_window(
+        first_signal=pcm.byte_offset(0.10),
+        last_signal=pcm.byte_offset(3.95),
+        total_bytes=total,
+        guard_bytes=pcm.byte_offset(1.0),
+    )
+    assert (start, end) == (0, total)
+
+
+def test_keep_window_offsets_are_sample_aligned():
+    start, end = keep_window(3, 9, 21, 1)
+    assert start % pcm.FRAME_BYTES == 0
+    assert end % pcm.FRAME_BYTES == 0
+
+
+def test_no_signal_found_keeps_everything():
+    total = pcm.byte_offset(6.0)
+    assert keep_window(None, None, total, pcm.byte_offset(GUARD)) == (0, total)
+
+
+def test_an_inverted_window_degrades_to_keeping_everything():
+    """Never return an empty slice, whatever the inputs say."""
+    total = pcm.byte_offset(4.0)
+    assert keep_window(pcm.byte_offset(3.0), pcm.byte_offset(0.5), total, 0) == (0, total)
+
+
+# ---------------------------------------------------------------------------
+# Scanning
+# ---------------------------------------------------------------------------
+
+def test_first_and_last_signal_are_found_in_a_realistic_recording():
+    audio = silence(1.85) + speech(2.35) + silence(0.34)
+    threshold = -40.0
+
+    first = first_signal_offset(audio, threshold)
+    last = last_signal_offset(audio, threshold)
+
+    assert pcm.seconds(first) == pytest.approx(1.85, abs=audio_trim.ANALYSIS_WINDOW_SECONDS)
+    assert pcm.seconds(last) == pytest.approx(4.20, abs=audio_trim.ANALYSIS_WINDOW_SECONDS)
+
+
+def test_internal_pauses_are_not_treated_as_the_tail():
+    """Trimming the middle out of a conversation would be unrecoverable."""
+    audio = silence(1.9) + speech(3.1) + silence(2.5) + speech(4.5)
+    last = last_signal_offset(audio, -40.0)
+    assert pcm.seconds(last) == pytest.approx(12.0, abs=audio_trim.ANALYSIS_WINDOW_SECONDS)
+
+
+def test_all_silence_returns_no_signal_offset():
+    assert first_signal_offset(silence(4.0), -40.0) is None
+    assert last_signal_offset(silence(4.0), -40.0) is None
+
+
+def test_the_scan_is_bounded_so_a_long_buffer_cannot_stall_the_upload():
+    """The per-sample loop is the only unbounded cost; it must have a ceiling."""
+    audio = silence(2.0)
+    assert first_signal_offset(audio, -40.0, limit_seconds=0.5) is None
+    assert first_signal_offset(speech(0.1) + silence(1.9), -40.0, limit_seconds=0.5) == 0
+
+
+# ---------------------------------------------------------------------------
+# trim_pcm — end to end over synthesised audio
+# ---------------------------------------------------------------------------
+
+def test_leading_and_trailing_silence_are_trimmed_to_the_guard_margin():
+    audio = silence(1.85) + speech(2.35) + silence(0.34)
+    kept, result = trim_pcm(audio, threshold_db=-40.0, guard=GUARD)
+
+    assert result.applied and not result.all_silence
+    assert result.lead == pytest.approx(1.85 - GUARD, abs=0.05)
+    assert result.tail == pytest.approx(0.34 - GUARD, abs=0.05)
+    assert pcm.seconds(len(kept)) == pytest.approx(result.duration_after, abs=0.001)
+    assert result.duration_after < result.duration_before
+
+
+def test_the_guard_margin_never_eats_into_speech():
+    audio = silence(2.0) + speech(1.0) + silence(2.0)
+    kept, result = trim_pcm(audio, threshold_db=-40.0, guard=GUARD)
+
+    # Everything removed from the head must be silence, and the first sample of
+    # real speech must survive.
+    assert result.lead < 2.0
+    assert pcm.seconds(len(kept)) > 1.0
+
+
+def test_measured_trailing_silence_is_reported_for_field_tuning():
     """
-
+    The recorder deliberately over-captures the tail (it closes only after the
-FFMPEG_STDERR_OPEN_TAIL = """
+    silence timeout has actually elapsed in the audio), so `tail` is how the
-  Duration: 00:00:04.54, start: 0.000000, bitrate: 16 kb/s
+    real silence run reaches the logs.
-[silencedetect @ 0x1] silence_start: 0
-[silencedetect @ 0x1] silence_end: 1.85042 | silence_duration: 1.85042
-[silencedetect @ 0x1] silence_start: 4.20134
     """
+    audio = silence(0.5) + speech(2.0) + silence(3.0)
+    _, result = trim_pcm(audio, threshold_db=-40.0, guard=GUARD)
+    assert result.tail == pytest.approx(3.0 - GUARD, abs=0.05)
+    assert result.trimmed_seconds == pytest.approx(result.lead + result.tail)
 
 
-def test_duration_is_corrected_for_the_mp3_container_start_offset():
+def test_an_all_silence_buffer_is_reported_not_truncated_to_nothing():
+    audio = silence(4.0)
+    kept, result = trim_pcm(audio, threshold_db=-40.0, guard=GUARD)
+
+    assert result.all_silence
+    assert not result.applied
+    assert kept == audio, "an all-silence recording must not become zero-length"
+
+
+def test_digital_silence_at_the_measured_field_floor_is_detected():
     """
-    MP3 encoder delay makes the container duration longer than the audio
+    The -91 dBFS floor is the whole reason this needs no field calibration.
-    silencedetect timestamps. Without this correction the trailing-region test
+    Detection must not depend on the threshold being tuned to a noise floor.
-    needs a slack epsilon big enough to clip real speech.
     """
-    assert _parse_duration(FFMPEG_STDERR) == pytest.approx(4.70 - 0.050113)
+    audio = silence(1.0, level=1) + speech(1.0) + silence(1.0, level=1)
+    for threshold in (-70.0, -60.0, -50.0, -40.0):
+        _, result = trim_pcm(audio, threshold_db=threshold, guard=GUARD)
+        assert result.applied, f"threshold {threshold} should still find the speech"
+        assert result.lead == pytest.approx(0.75, abs=0.05)
 
 
-def test_duration_is_none_when_absent():
+def test_audio_with_no_silence_at_either_end_is_left_alone():
-    assert _parse_duration("no duration here") is None
+    audio = speech(3.0)
+    kept, result = trim_pcm(audio, threshold_db=-40.0, guard=GUARD)
+
+    assert not result.applied
+    assert kept == audio
+    assert result.duration_before == pytest.approx(result.duration_after)
 
 
-def test_silence_regions_are_parsed():
+def test_an_empty_buffer_is_handled():
-    regions = _parse_silences(FFMPEG_STDERR)
+    kept, result = trim_pcm(b"", threshold_db=-40.0, guard=GUARD)
-
+    assert kept == b"" and not result.applied and not result.all_silence
-    assert len(regions) == 2
-    assert regions[0] == (pytest.approx(0.0), pytest.approx(2.05361))
-    assert regions[1] == (pytest.approx(3.56367), pytest.approx(4.63102))
 
 
-def test_a_region_with_no_silence_end_is_still_parsed():
+def test_thresholds_default_to_settings():
-    """Older FFmpeg simply stopped reporting at EOF — keep handling that."""
+    audio = silence(1.0) + speech(1.0) + silence(1.0)
-    regions = _parse_silences(FFMPEG_STDERR_OPEN_TAIL)
+    _, result = trim_pcm(audio)
+    assert result.applied
+    assert settings.trim_silence_threshold_db == -40.0
+    assert settings.trim_silence_guard_seconds == 0.25
+    assert result.lead == pytest.approx(1.0 - settings.trim_silence_guard_seconds, abs=0.05)
 
-    assert regions[-1][1] is None
+
+def test_a_scan_that_gives_up_leaves_the_audio_untouched_and_says_so():
+    """
+    Refusing to guess is the point: an untrimmed upload is always better than a
+    wrongly-truncated one, and better than dropping a call as "all silence"
+    without having actually looked at all of it.
+    """
+    long_silence = silence(audio_trim.MAX_SCAN_SECONDS + 5.0)
+    kept, result = trim_pcm(long_silence, threshold_db=-40.0, guard=GUARD)
+
+    assert result.scan_truncated
+    assert not result.all_silence
+    assert not result.applied
+    assert kept == long_silence
 
 
 def test_trim_result_reports_total_trimmed():
-    result = TrimResult(path=None, lead=1.9, tail=0.35)
+    assert TrimResult(lead=1.9, tail=0.35).trimmed_seconds == pytest.approx(2.25)
-    assert result.trimmed_seconds == pytest.approx(2.25)

drb-edge-node/tests/test_call_recorder.py

@@ -1,13 +1,17 @@
 """
-Unit tests for the CallRecorder pre-roll ring buffer and per-call accumulator.
+Unit tests for the CallRecorder: PCM ring buffer, per-call accumulator, the
+continuous voice-activity signal the segmenter reads, and the single encode.
 
 No FFmpeg and no PulseAudio: chunks are pushed through _ingest() with a patched
-clock, which is exactly what the capture loop does at runtime. Silence trimming
+clock, which is exactly what the capture loop does at runtime, and the MP3
-is disabled by default here and exercised separately with a stubbed trimmer.
+encoder is replaced with a stub that writes the raw PCM it was handed. That stub
+is also how "exactly one encode per call" is asserted — the old design captured
+MP3 and then re-encoded it to trim, so every upload was double-encoded.
 """
 import asyncio
 import itertools
 import time
+from array import array
 from typing import List
 from unittest.mock import patch
 
@@ -15,7 +19,7 @@ import pytest
 
 from app.config import settings
 from app.internal import call_recorder as recorder_mod
-from app.internal.audio_trim import TrimResult
+from app.internal import pcm
 from app.internal.call_recorder import (
     CallRecorder,
     MAX_RECORDING_BYTES,
@@ -25,11 +29,42 @@ from app.internal.call_recorder import (
 )
 
 T0 = 1_700_000_000.0
-CHUNK_INTERVAL = 0.1  # seconds of audio per synthetic chunk
+
+# One synthetic chunk carries exactly CHUNK_INTERVAL seconds of audio AND
+# arrives CHUNK_INTERVAL apart, so arrival-timestamp arithmetic (slicing) and
+# byte-offset arithmetic (trimming) agree with each other.
+CHUNK_INTERVAL = 0.1
+CHUNK_SAMPLES = int(pcm.SAMPLE_RATE * CHUNK_INTERVAL)
+CHUNK_BYTES = CHUNK_SAMPLES * pcm.FRAME_BYTES
+
+SPEECH_LEVEL = 4096   # -18 dBFS, the measured field average
+FLOOR_LEVEL = 1       # -90.3 dBFS, the measured digital-silence floor
+
+
+def block(level: int, samples: int = CHUNK_SAMPLES) -> bytes:
+    return array("h", [level, -level] * (samples // 2)).tobytes()
+
+
+VOICE = block(SPEECH_LEVEL)
+QUIET = block(FLOOR_LEVEL)
 
 
 @pytest.fixture
-def recorder(tmp_path, monkeypatch):
+def encodes(monkeypatch):
+    """Replace the one encode with a stub that writes the PCM it was given."""
+    calls: List[tuple] = []
+
+    async def _encode(audio: bytes, path):
+        calls.append((audio, path))
+        path.write_bytes(audio)
+        return True
+
+    monkeypatch.setattr(recorder_mod, "encode_mp3", _encode)
+    return calls
+
+
+@pytest.fixture
+def recorder(tmp_path, monkeypatch, encodes):
     monkeypatch.setattr(settings, "trim_silence", False)
     r = CallRecorder()
     r._recordings_dir = tmp_path
@@ -37,64 +72,51 @@ def recorder(tmp_path, monkeypatch):
     return r
 
 
-def ingest(recorder, start: float, end: float, marker: bytes = b"A", index: int = 0) -> int:
+def ingest(recorder, start: float, end: float, chunk: bytes = VOICE) -> None:
-    """Feed one chunk every CHUNK_INTERVAL seconds over [start, end) through _ingest."""
+    """Feed one chunk every CHUNK_INTERVAL seconds over [start, end)."""
     stamps: List[float] = []
-    chunks: List[bytes] = []
     ts = start
     while ts < end:
         stamps.append(ts)
-        chunks.append(marker + str(index).encode() + b";")
-        index += 1
         ts = round(ts + CHUNK_INTERVAL, 6)
-
+    if not stamps:
+        return
     with patch("app.internal.call_recorder.time.time", side_effect=stamps):
-        for chunk in chunks:
+        for _ in stamps:
             recorder._ingest(chunk)
-    return index
 
 
-def fill(recorder, start: float, end: float, marker: bytes = b"A", index: int = 0) -> int:
+def duration_of(path) -> float:
-    """Alias kept for readability where the accumulator is not the point."""
+    return pcm.seconds(len(path.read_bytes()))
-    return ingest(recorder, start, end, marker=marker, index=index)
 
 
 def timestamps(recorder):
     return [ts for ts, _ in recorder._buffer]
 
 
-def markers(path) -> List[str]:
-    return path.read_bytes().decode().strip(";").split(";")
-
-
-def indices(path) -> List[int]:
-    return [int(m[1:]) for m in markers(path) if m[1:].isdigit()]
-
-
 # ---------------------------------------------------------------------------
 # Ring buffer trimming (pre-roll duty only)
 # ---------------------------------------------------------------------------
 
 def test_idle_buffer_keeps_only_the_rolling_window(recorder):
-    for offset in range(0, int(RING_BUFFER_SECONDS) + 20):
+    ingest(recorder, T0, T0 + RING_BUFFER_SECONDS + 20)
-        with patch("app.internal.call_recorder.time.time", return_value=T0 + offset):
-            recorder._ingest(b"x" * 16)
 
-    assert len(recorder._buffer) <= RING_BUFFER_SECONDS + 1
+    assert recorder.buffered_seconds <= RING_BUFFER_SECONDS + CHUNK_INTERVAL
-    assert min(timestamps(recorder)) >= (T0 + RING_BUFFER_SECONDS + 19) - RING_BUFFER_SECONDS
+    newest = max(timestamps(recorder))
+    assert min(timestamps(recorder)) >= newest - RING_BUFFER_SECONDS
 
 
 @pytest.mark.asyncio
 async def test_ring_buffer_is_trimmed_even_while_recording(recorder):
     """
-    The ring buffer serves PRE-ROLL only. An open recording must no longer pin
+    The ring buffer serves PRE-ROLL only. An open recording must not pin it —
-    it — that was the mechanism that made call length depend on buffer size.
+    that was the mechanism that made call length depend on buffer size.
     """
-    index = ingest(recorder, T0, T0 + 2.0)
+    ingest(recorder, T0, T0 + 2.0)
     await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
-    ingest(recorder, T0 + 2.0, T0 + 2.0 + RING_BUFFER_SECONDS + 10, index=index)
+    ingest(recorder, T0 + 2.0, T0 + 2.0 + RING_BUFFER_SECONDS + 10)
 
-    assert recorder.buffered_seconds <= RING_BUFFER_SECONDS + 1
+    assert recorder.buffered_seconds <= RING_BUFFER_SECONDS + CHUNK_INTERVAL
     # ...and the audio the ring buffer dropped is safe in the accumulator.
     assert recorder._active is not None
     assert recorder._active.chunks[0][0] == pytest.approx(T0 + 1.0 - PRE_ROLL_SECONDS, abs=CHUNK_INTERVAL)
@@ -110,19 +132,16 @@ async def test_call_longer_than_the_ring_buffer_is_captured_whole(recorder):
     grant = T0 + 1.0
     end = grant + call_length
 
-    index = ingest(recorder, T0, grant)
+    ingest(recorder, T0, grant)
     await recorder.start_recording("call-long", start_epoch=grant)
-    ingest(recorder, grant, end + 1.0, index=index)
+    ingest(recorder, grant, end + 1.0)
 
     rec = await recorder.stop_recording(end_epoch=end)
     assert rec is not None and rec.path is not None
 
-    kept = indices(rec.path)
+    captured = duration_of(rec.path)
-    # Contiguous: no hole anywhere in the middle of a 65s call.
+    assert captured > RING_BUFFER_SECONDS, "call length must not be clamped by the ring buffer"
-    assert kept == list(range(kept[0], kept[-1] + 1))
+    assert captured == pytest.approx(call_length + PRE_ROLL_SECONDS, abs=2 * CHUNK_INTERVAL)
-    span = (kept[-1] - kept[0]) * CHUNK_INTERVAL
-    assert span > RING_BUFFER_SECONDS, "call length must not be clamped by the ring buffer"
-    assert span == pytest.approx(call_length + PRE_ROLL_SECONDS, abs=2 * CHUNK_INTERVAL)
 
 
 @pytest.mark.asyncio
@@ -130,8 +149,12 @@ async def test_accumulator_stops_growing_at_the_memory_ceiling(recorder, caplog)
     """A runaway call must not be able to exhaust RAM on a Pi."""
     await recorder.start_recording("call-runaway", start_epoch=T0)
 
-    big = b"z" * 64_000
+    # Silent blocks on purpose: this test is about bytes, not content, and the
+    # all-zero fast path keeps it from spending seconds in the RMS loop.
+    big = b"\x00" * 64_000
     needed = (MAX_RECORDING_BYTES // len(big)) + 5
+    # An unbounded clock: patching time.time patches it for everything running
+    # inside the block, not only for our calls.
     ticks = itertools.count()
     with caplog.at_level("WARNING", logger="drb-edge-node"):
         with patch("app.internal.call_recorder.time.time",
@@ -145,73 +168,133 @@ async def test_accumulator_stops_growing_at_the_memory_ceiling(recorder, caplog)
     assert any("memory ceiling" in r.message for r in caplog.records)
 
 
+def test_the_byte_ceiling_can_never_truncate_a_legal_call():
+    """
+    PCM costs 44.1 KB/s where MP3 cost 2 KB/s, so this had to be re-derived.
+    The TIME cap must always bite before the BYTE cap, or a long pursuit would
+    be silently cut short by a memory limit.
+    """
+    assert MAX_RECORDING_BYTES > MAX_RECORDING_SECONDS * pcm.BYTES_PER_SECOND
+    # ...and it still has to be a deliberate, bounded number on a Pi.
+    assert MAX_RECORDING_BYTES <= 48 * 1024 * 1024
+
+
+def test_the_ring_buffer_memory_cost_is_bounded():
+    assert RING_BUFFER_SECONDS * pcm.BYTES_PER_SECOND < 2 * 1024 * 1024
+
+
+# ---------------------------------------------------------------------------
+# Voice activity — the signal the segmenter starts and stops on
+# ---------------------------------------------------------------------------
+
+def test_digital_silence_produces_no_voice_marks(recorder):
+    ingest(recorder, T0, T0 + 5.0, chunk=QUIET)
+
+    activity = recorder.audio_activity()
+    assert activity.last_voice_epoch is None
+    assert activity.voice_onset_epoch is None
+
+
+def test_voice_onset_is_the_arrival_of_the_first_non_silent_chunk(recorder):
+    ingest(recorder, T0, T0 + 2.0, chunk=QUIET)
+    ingest(recorder, T0 + 2.0, T0 + 3.0, chunk=VOICE)
+
+    activity = recorder.audio_activity()
+    assert activity.voice_onset_epoch == pytest.approx(T0 + 2.0)
+    assert activity.last_voice_epoch == pytest.approx(T0 + 3.0 - CHUNK_INTERVAL)
+
+
+def test_a_gap_shorter_than_the_silence_timeout_does_not_start_a_new_run(recorder, monkeypatch):
+    """Back-and-forth inside the window is ONE run, hence one recording."""
+    monkeypatch.setattr(settings, "call_silence_timeout", 3.0)
+    ingest(recorder, T0, T0 + 1.0, chunk=VOICE)
+    ingest(recorder, T0 + 1.0, T0 + 2.5, chunk=QUIET)
+    ingest(recorder, T0 + 2.5, T0 + 3.5, chunk=VOICE)
+
+    assert recorder.audio_activity().voice_onset_epoch == pytest.approx(T0)
+
+
+def test_a_gap_longer_than_the_silence_timeout_starts_a_new_run(recorder, monkeypatch):
+    monkeypatch.setattr(settings, "call_silence_timeout", 3.0)
+    ingest(recorder, T0, T0 + 1.0, chunk=VOICE)
+    ingest(recorder, T0 + 1.0, T0 + 6.0, chunk=QUIET)
+    ingest(recorder, T0 + 6.0, T0 + 7.0, chunk=VOICE)
+
+    assert recorder.audio_activity().voice_onset_epoch == pytest.approx(T0 + 6.0)
+
+
+def test_activity_snapshot_reports_capture_and_recording_state(recorder):
+    activity = recorder.audio_activity()
+    assert activity.capturing is True
+    assert activity.recording is False
+
+    recorder._capturing = False
+    assert recorder.audio_activity().capturing is False
+
+
 # ---------------------------------------------------------------------------
 # Pre-roll and slicing
 # ---------------------------------------------------------------------------
 
 @pytest.mark.asyncio
-async def test_slice_starts_pre_roll_before_the_op25_timestamp(recorder):
+async def test_slice_starts_pre_roll_before_the_detected_onset(recorder):
-    fill(recorder, T0, T0 + 10.0)
+    ingest(recorder, T0, T0 + 10.0)
-    grant_time = T0 + 5.0
+    onset = T0 + 5.0
 
-    await recorder.start_recording("call-1", start_epoch=grant_time)
+    await recorder.start_recording("call-1", start_epoch=onset)
-    assert recorder._active.slice_start == pytest.approx(grant_time - PRE_ROLL_SECONDS)
+    assert recorder._active.slice_start == pytest.approx(onset - PRE_ROLL_SECONDS)
 
-    rec = await recorder.stop_recording(end_epoch=grant_time + 2.0)
+    rec = await recorder.stop_recording(end_epoch=onset + 2.0)
     assert rec is not None and rec.path is not None and rec.path.exists()
 
-    first_ts = T0 + indices(rec.path)[0] * CHUNK_INTERVAL
+    first_ts = recorder._active.chunks[0][0] if recorder._active else None
-
+    assert first_ts is None  # recording closed
-    # A chunk stamped `ts` holds the audio that arrived over [ts - interval, ts],
+    # The audio actually covered must begin at or before the requested slice
-    # so the audio actually covered must begin at or before the requested slice
+    # start — erring early is safe, erring late loses speech.
-    # start — erring early is the safe direction, erring late loses speech.
+    assert rec.audio_start_epoch <= onset - PRE_ROLL_SECONDS + 1e-6
-    assert first_ts - CHUNK_INTERVAL <= grant_time - PRE_ROLL_SECONDS + 1e-6
-    # ...and no more than one chunk of extra pre-roll is dragged in.
-    assert first_ts >= grant_time - PRE_ROLL_SECONDS - 1e-6
 
 
 @pytest.mark.asyncio
 async def test_tail_chunk_straddling_the_end_is_included(recorder):
-    fill(recorder, T0, T0 + 10.0)
+    ingest(recorder, T0, T0 + 10.0)
     await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
 
-    # End halfway through a chunk interval.
     rec = await recorder.stop_recording(end_epoch=T0 + 3.05)
-    last_ts = T0 + indices(rec.path)[-1] * CHUNK_INTERVAL
 
-    assert last_ts >= T0 + 3.05, "the chunk covering the end instant must be kept"
+    # The chunk covering the end instant must be kept, so the captured audio
+    # reaches past the requested end rather than stopping short of it.
+    assert rec.audio_end_epoch >= T0 + 3.05
 
 
 @pytest.mark.asyncio
 async def test_max_recording_seconds_caps_the_slice(recorder):
-    index = fill(recorder, T0, T0 + 1.0)
+    ingest(recorder, T0, T0 + 1.0)
     await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
-    ingest(recorder, T0 + 1.0, T0 + MAX_RECORDING_SECONDS + 60, index=index)
+    ingest(recorder, T0 + 1.0, T0 + MAX_RECORDING_SECONDS + 60)
 
     rec = await recorder.stop_recording(end_epoch=T0 + MAX_RECORDING_SECONDS + 50)
-    last_ts = T0 + indices(rec.path)[-1] * CHUNK_INTERVAL
 
-    assert last_ts <= T0 + 1.0 + MAX_RECORDING_SECONDS + CHUNK_INTERVAL
+    assert duration_of(rec.path) <= MAX_RECORDING_SECONDS + PRE_ROLL_SECONDS + CHUNK_INTERVAL
 
 
 # ---------------------------------------------------------------------------
-# Tail wait — the fix for recordings that ended mid-word
+# Tail wait — still needed for control-channel-derived ends (tgid splits)
 # ---------------------------------------------------------------------------
 
 @pytest.mark.asyncio
 async def test_stop_waits_for_captured_audio_to_reach_the_call_end(recorder, caplog):
     """
-    PulseAudio → FFmpeg → encoder → muxer → our pipe read has latency, so at the
+    A tgid_change close pads past a control-channel timestamp that is ~now, so
-    instant a call ends the newest captured chunk is OLDER than the end epoch.
+    the audio it asks for has not been captured yet. Slicing immediately would
-    Slicing immediately cuts the last word off. stop_recording must wait for it.
+    cut the last word off.
     """
-    index = ingest(recorder, T0, T0 + 4.0)
+    ingest(recorder, T0, T0 + 4.0)
     await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
 
     async def late_tail():
         await asyncio.sleep(0.15)
         with patch("app.internal.call_recorder.time.time", return_value=T0 + 4.6):
-            recorder._ingest(b"TAIL;")
+            recorder._ingest(block(SPEECH_LEVEL))
 
     task = asyncio.create_task(late_tail())
     with caplog.at_level("INFO", logger="drb-edge-node"):
@@ -219,10 +302,14 @@ async def test_stop_waits_for_captured_audio_to_reach_the_call_end(recorder, cap
     await task
 
     assert rec is not None and rec.path is not None
-    assert b"TAIL" in rec.path.read_bytes(), "the late-arriving tail must be in the file"
+    # The slice covers the chunks stamped T0+0.8 .. T0+3.9 (32 of them) plus the
+    # one that arrived late — and that last one is where the final word of the
+    # transmission lives. Without the wait it would have been cut.
+    chunk_seconds = pcm.seconds(len(VOICE))
+    assert duration_of(rec.path) == pytest.approx(33 * chunk_seconds, abs=0.01)
+    assert duration_of(rec.path) > 32 * chunk_seconds
     assert any("Waited" in r.message and "tail" in r.message for r in caplog.records), \
         "a tail wait must be observable in the field logs"
-    assert index  # sanity: the pre-roll fill actually ran
 
 
 @pytest.mark.asyncio
@@ -245,6 +332,10 @@ async def test_tail_wait_is_bounded_and_warns_when_audio_never_arrives(recorder,
 
 @pytest.mark.asyncio
 async def test_no_wait_when_the_buffer_already_covers_the_end(recorder, caplog):
+    """
+    An audio-driven close derives its end epoch from audio that is already
+    buffered, so the common path must never pay the tail wait at all.
+    """
     ingest(recorder, T0, T0 + 10.0)
     await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
 
@@ -261,7 +352,7 @@ async def test_no_wait_when_the_buffer_already_covers_the_end(recorder, caplog):
 
 @pytest.mark.asyncio
 async def test_pre_roll_earlier_than_buffer_start_is_clamped_and_warned(recorder, caplog):
-    """A grant older than anything buffered must still produce a file, loudly."""
+    """An onset older than anything buffered must still produce a file, loudly."""
     ingest(recorder, T0 + 5.0, T0 + 10.0)  # buffer only covers T0+5 onwards
 
     with caplog.at_level("WARNING", logger="drb-edge-node"):
@@ -269,36 +360,87 @@ async def test_pre_roll_earlier_than_buffer_start_is_clamped_and_warned(recorder
     rec = await recorder.stop_recording(end_epoch=T0 + 8.0)
 
     assert rec is not None and rec.path is not None and rec.path.stat().st_size > 0
-    assert markers(rec.path)[0] == "A0", "slice should begin at the buffer head, not fail"
-    # Buffer head is T0+5.0, requested slice start is T0-PRE_ROLL: everything in
-    # between is audio we can never recover, and the number must be reported.
     assert rec.clamped_seconds == pytest.approx(5.0 + PRE_ROLL_SECONDS, abs=CHUNK_INTERVAL)
     assert any("BUFFER CLAMP" in r.message for r in caplog.records)
 
 
 @pytest.mark.asyncio
-async def test_no_buffered_audio_returns_none(recorder):
+async def test_no_buffered_audio_returns_none(recorder, encodes):
     await recorder.start_recording("call-1", start_epoch=T0)
     assert await recorder.stop_recording(end_epoch=T0 + 2.0) is None
+    assert encodes == [], "nothing to encode means no encoder subprocess"
 
 
 @pytest.mark.asyncio
 async def test_start_epoch_omitted_falls_back_to_now(recorder):
     now = time.time()
-    fill(recorder, now - 5.0, now)
+    ingest(recorder, now - 5.0, now)
 
     await recorder.start_recording("call-1")
     assert recorder._active.slice_start == pytest.approx(now - PRE_ROLL_SECONDS, abs=1.0)
 
 
+# ---------------------------------------------------------------------------
+# Encode — exactly once, at save time
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_audio_is_encoded_exactly_once_per_call(recorder, encodes, monkeypatch):
+    """
+    The old pipeline captured MP3 and then re-encoded it to trim, so every
+    upload was double-encoded. Capture is PCM now and MP3 happens once, after
+    trimming, at save time.
+    """
+    monkeypatch.setattr(settings, "trim_silence", True)
+    ingest(recorder, T0, T0 + 1.0, chunk=QUIET)
+    ingest(recorder, T0 + 1.0, T0 + 3.0, chunk=VOICE)
+    ingest(recorder, T0 + 3.0, T0 + 6.0, chunk=QUIET)
+
+    await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
+    rec = await recorder.stop_recording(end_epoch=T0 + 6.0)
+
+    assert rec is not None and rec.path is not None
+    assert len(encodes) == 1, "exactly one encode per recording"
+    encoded_audio, encoded_path = encodes[0]
+    assert encoded_path == rec.path
+    # What was encoded is the TRIMMED audio, not the raw slice.
+    assert pcm.seconds(len(encoded_audio)) < 5.0
+
+
+@pytest.mark.asyncio
+async def test_a_failed_encode_leaves_no_file_and_no_recording(recorder, monkeypatch):
+    async def _fail(audio, path):
+        return False
+
+    monkeypatch.setattr(recorder_mod, "encode_mp3", _fail)
+    ingest(recorder, T0, T0 + 5.0)
+    await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
+
+    assert await recorder.stop_recording(end_epoch=T0 + 3.0) is None
+    assert list(recorder._recordings_dir.glob("*.mp3")) == []
+
+
+def test_encoder_command_contract_matches_what_c2_expects():
+    """
+    /upload has always received mono MP3 at 22050 Hz / 16 kbps, and Whisper
+    consumes it downstream. The single encode must not quietly change that.
+    """
+    assert recorder_mod.MP3_SAMPLE_RATE == str(pcm.SAMPLE_RATE) == "22050"
+    assert recorder_mod.MP3_BITRATE == "16k"
+
+
 # ---------------------------------------------------------------------------
 # Silence trimming and timing metadata
 # ---------------------------------------------------------------------------
 
-async def _recorded(recorder, end_offset: float = 3.0):
+async def _recorded(recorder, lead_silence=1.0, voice=2.0, tail_silence=1.0):
-    ingest(recorder, T0, T0 + 10.0)
+    start = T0
-    await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
+    ingest(recorder, start, start + lead_silence, chunk=QUIET)
-    return await recorder.stop_recording(end_epoch=T0 + end_offset)
+    ingest(recorder, start + lead_silence, start + lead_silence + voice, chunk=VOICE)
+    ingest(recorder, start + lead_silence + voice,
+           start + lead_silence + voice + tail_silence, chunk=QUIET)
+    await recorder.start_recording("call-1", start_epoch=start + 0.5)
+    return await recorder.stop_recording(end_epoch=start + lead_silence + voice + tail_silence)
 
 
 @pytest.mark.asyncio
@@ -306,12 +448,12 @@ async def test_trimming_is_off_when_the_setting_is_off(recorder, monkeypatch):
     monkeypatch.setattr(settings, "trim_silence", False)
     called = False
 
-    async def _never(*args, **kwargs):
+    def _never(*args, **kwargs):
         nonlocal called
         called = True
-        return TrimResult(path=None)
+        return b"", None
 
-    monkeypatch.setattr(recorder_mod.audio_trim, "trim_silence", _never)
+    monkeypatch.setattr(recorder_mod.audio_trim, "trim_pcm", _never)
     rec = await _recorded(recorder)
     assert rec is not None and not called
 
@@ -325,41 +467,31 @@ async def test_trim_shifts_the_audio_bounds_but_not_the_call_bounds(recorder, mo
     """
     monkeypatch.setattr(settings, "trim_silence", True)
 
-    async def _trim(path, **kwargs):
+    rec = await _recorded(recorder, lead_silence=1.0, voice=2.0, tail_silence=1.5)
-        return TrimResult(path=path, lead=1.9, tail=0.4, duration_before=3.3,
-                          duration_after=1.0, applied=True)
 
-    monkeypatch.setattr(recorder_mod.audio_trim, "trim_silence", _trim)
-
-    rec = await _recorded(recorder)
     assert rec is not None and rec.path is not None
-    assert rec.lead_trimmed == pytest.approx(1.9)
+    assert rec.lead_trimmed > 0.0 and rec.tail_trimmed > 0.0
-    assert rec.tail_trimmed == pytest.approx(0.4)
+    guard = settings.trim_silence_guard_seconds
-    # Untrimmed slice was [T0+0.75, T0+3.0]; the audio now starts 1.9s later and
+    # Slice began at T0+0.25; speech begins at T0+1.0, so the audio now starts
-    # ends 0.4s earlier, which is exactly what downstream needs to map an audio
+    # one guard margin before the speech.
-    # offset back to wall clock.
+    assert rec.audio_start_epoch == pytest.approx(T0 + 1.0 - guard, abs=3 * CHUNK_INTERVAL)
-    assert rec.audio_start_epoch == pytest.approx(T0 + 1.0 - PRE_ROLL_SECONDS + 1.9, abs=CHUNK_INTERVAL)
+    assert rec.audio_end_epoch == pytest.approx(T0 + 3.0 + guard, abs=3 * CHUNK_INTERVAL)
-    assert rec.audio_end_epoch == pytest.approx(T0 + 3.0 - 0.4, abs=CHUNK_INTERVAL)
+    assert rec.audio_end_epoch > rec.audio_start_epoch
 
 
 @pytest.mark.asyncio
-async def test_all_silence_recording_is_dropped_and_logged(recorder, monkeypatch, caplog):
+async def test_all_silence_recording_is_dropped_and_logged(recorder, monkeypatch, caplog, encodes):
     monkeypatch.setattr(settings, "trim_silence", True)
-    seen = {}
+    ingest(recorder, T0, T0 + 5.0, chunk=QUIET)
-
+    await recorder.start_recording("call-1", start_epoch=T0 + 1.0)
-    async def _trim(path, **kwargs):
-        seen["path"] = path
-        return TrimResult(path=path, duration_before=4.0, duration_after=4.0, all_silence=True)
-
-    monkeypatch.setattr(recorder_mod.audio_trim, "trim_silence", _trim)
 
     with caplog.at_level("WARNING", logger="drb-edge-node"):
-        rec = await _recorded(recorder)
+        rec = await recorder.stop_recording(end_epoch=T0 + 4.0)
 
     assert rec is not None
     assert rec.all_silence is True
     assert rec.path is None, "an all-silence recording must not be uploaded"
-    assert not seen["path"].exists(), "the file must be cleaned up, not left on disk"
+    assert encodes == [], "and must not be encoded either"
     assert any("no speech" in r.message for r in caplog.records)
 
 
@@ -369,7 +501,7 @@ async def test_all_silence_recording_is_dropped_and_logged(recorder, monkeypatch
 
 @pytest.mark.asyncio
 async def test_second_start_is_rejected_while_recording(recorder):
-    fill(recorder, T0, T0 + 5.0)
+    ingest(recorder, T0, T0 + 5.0)
     assert await recorder.start_recording("call-1", start_epoch=T0 + 1.0) is True
     assert await recorder.start_recording("call-2", start_epoch=T0 + 2.0) is False
     assert recorder.is_recording
@@ -381,6 +513,21 @@ async def test_stop_without_start_is_a_noop(recorder):
     assert not recorder.is_recording
 
 
+@pytest.mark.asyncio
+async def test_discard_drops_the_audio_without_writing_anything(recorder, encodes):
+    """The orphan-audio path: unattributed audio must never reach a file."""
+    ingest(recorder, T0, T0 + 5.0)
+    await recorder.start_recording("call-orphan", start_epoch=T0 + 1.0)
+
+    await recorder.discard_recording()
+
+    assert not recorder.is_recording
+    assert encodes == []
+    assert list(recorder._recordings_dir.glob("*.mp3")) == []
+    # ...and the recorder is immediately reusable.
+    assert await recorder.start_recording("call-next", start_epoch=T0 + 2.0) is True
+
+
 @pytest.mark.asyncio
 async def test_split_then_immediate_restart_keeps_both_slices(recorder):
     """
@@ -405,22 +552,26 @@ async def test_split_then_immediate_restart_keeps_both_slices(recorder):
 # FFmpeg invocation
 # ---------------------------------------------------------------------------
 
-def test_ffmpeg_command_reads_pulse_and_flushes_packets(recorder):
+def test_capture_command_asks_for_raw_pcm_not_mp3(recorder):
     cmd = recorder._ffmpeg_command()
     joined = " ".join(cmd)
 
     assert "-f pulse" in joined
     assert "drb_sink.monitor" in joined, "must address the monitor explicitly, not 'default'"
-    # Without -flush_packets the mp3 muxer buffers 32 KB (~16 s at 16 kbps) before
+    assert cmd[-1] == "-" and cmd[-2] == "s16le", "capture must emit raw PCM on stdout"
-    # writing, which would destroy the ring buffer's timestamp resolution.
+    assert "mp3" not in joined, "MP3 now happens once at save time, not in the capture"
-    assert "-flush_packets" in cmd
+    assert "-ar" in cmd and str(pcm.SAMPLE_RATE) in cmd
-    assert cmd[-1] == "-" and cmd[-2] == "mp3", "must emit MP3 on stdout for /upload"
+    assert "-ac" in cmd and str(pcm.CHANNELS) in cmd
 
 
-def test_memory_ceiling_covers_the_longest_allowed_call():
+def test_read_chunk_is_finer_than_the_pre_roll(recorder):
-    """The cap must bound RAM without ever being able to truncate a legal call."""
+    """
-    bytes_per_second = 16_000 // 8
+    Chunk size is both the ring buffer's timestamp resolution and the window
-    assert MAX_RECORDING_BYTES >= MAX_RECORDING_SECONDS * bytes_per_second
+    silence detection runs over, so it has to stay well under the pre-roll.
+    """
+    chunk_seconds = pcm.seconds(recorder_mod.READ_CHUNK_BYTES)
+    assert chunk_seconds < PRE_ROLL_SECONDS / 4
+    assert recorder_mod.READ_CHUNK_BYTES % pcm.FRAME_BYTES == 0
 
 
 # ---------------------------------------------------------------------------
@@ -481,4 +632,3 @@ def test_capture_exit_with_no_stderr_captured_is_generic_warning(recorder, caplo
         recorder._log_capture_exit()
 
     assert _log_levels(caplog) == ["WARNING"]
-    assert MAX_RECORDING_BYTES <= 8 * 1024 * 1024, "must stay small enough for a Pi"

drb-edge-node/tests/test_metadata_watcher.py

@@ -1,15 +1,30 @@
 """
-Unit tests for the event-driven MetadataWatcher state machine.
+Unit tests for the MetadataWatcher segmentation state machine.
 
-Call START comes from OP25 `call_log` entries (stamped with OP25's own
+TWO MODES, both covered here:
-time.time()); call END comes from the srcaddr != 0 -> srcaddr == 0 transition in
+
-`channel_update`.  All OP25 HTTP calls are mocked — no running services required.
+  AUDIO MODE (production) — a recording STARTS at voice onset heard in the
+  captured audio and STOPS after settings.call_silence_timeout seconds of
+  silence heard in the same audio. The OP25 console supplies only the LABEL
+  (talkgroup/alias/rid), resolved at CLOSE time from a rolling history, and the
+  forced SPLIT when the talkgroup changes with no silence between calls.
+
+  CONSOLE FALLBACK (capture down) — the older state machine: `call_log` grants
+  start segments, the srcaddr != 0 -> 0 edge plus call_idle_timeout ends them.
+  Tests that wire no audio provider exercise this path, which is exactly the
+  behaviour a node falls back to when PulseAudio is not producing audio.
+
+All OP25 HTTP calls are mocked — no running services required.
 """
 import pytest
 from unittest.mock import AsyncMock, patch
 
 from app.config import settings
+from app.internal.call_recorder import AudioActivity
 from app.internal.metadata_watcher import (
+    ATTRIBUTION_LOOKAHEAD_SECONDS,
+    ATTRIBUTION_LOOKBACK_SECONDS,
+    MAX_SEGMENT_SECONDS,
     MetadataWatcher,
     OP25_OFFLINE_GRACE,
 )
@@ -37,6 +52,7 @@ def clock():
 
 @pytest.fixture
 def watcher(clock):
+    """Console fallback mode: no audio provider wired, capture assumed down."""
     w = MetadataWatcher()
     w._clock = clock
     w.on_call_start = AsyncMock()
@@ -44,6 +60,58 @@ def watcher(clock):
     return w
 
 
+class FakeAudio:
+    """
+    Stands in for CallRecorder.audio_activity.
+
+    Mirrors the recorder's own rule for what starts a new voice RUN: a
+    non-silent chunk more than settings.call_silence_timeout after the previous
+    one. Tests drive it with speak()/quiet() instead of synthesising PCM, so the
+    segmentation logic is tested independently of the detector.
+    """
+
+    def __init__(self, clock):
+        self.clock = clock
+        self.capturing = True
+        self.recording = False
+        self.last_voice = None
+        self.onset = None
+
+    def speak(self, at=None):
+        """Mark voice heard now (or at `at`)."""
+        moment = self.clock.now if at is None else at
+        if self.last_voice is None or (moment - self.last_voice) >= settings.call_silence_timeout:
+            self.onset = moment
+        self.last_voice = moment
+        return moment
+
+    def __call__(self) -> AudioActivity:
+        silence = 0.0 if self.last_voice is None else max(0.0, self.clock.now - self.last_voice)
+        return AudioActivity(
+            capturing=self.capturing,
+            recording=self.recording,
+            last_voice_epoch=self.last_voice,
+            voice_onset_epoch=self.onset,
+            silence_seconds=silence,
+        )
+
+
+@pytest.fixture
+def audio(clock):
+    return FakeAudio(clock)
+
+
+@pytest.fixture
+def hearing(clock, audio):
+    """Audio mode: the production wiring, with a controllable audio stream."""
+    w = MetadataWatcher()
+    w._clock = clock
+    w.on_call_start = AsyncMock()
+    w.on_call_end = AsyncMock()
+    w.audio_activity = audio
+    return w
+
+
 def grant(tgid: int, time_: float, tgtag: str = "", rid: int = 101, freq: int = 851_000_000):
     """One OP25 call_log entry (see tk_p25.log_call)."""
     return {
@@ -195,7 +263,12 @@ async def test_op25_unreachable_does_not_start_call(watcher):
 
 
 # ---------------------------------------------------------------------------
-# Call end — srcaddr edge + idle timeout
+# Console fallback: call end — srcaddr edge + idle timeout
+#
+# This is the path a node uses ONLY when PulseAudio capture is not producing
+# audio. It is measurably wrong (the srcaddr edge fires mid-word) but it is all
+# there is when there is no audio to segment on, and it keeps the node
+# reporting radio activity to C2 while the audio path is broken.
 # ---------------------------------------------------------------------------
 
 @pytest.mark.asyncio
@@ -452,9 +525,11 @@ async def test_different_tgid_grant_splits_recording(watcher, clock):
     assert ended["call_id"] == first_id
     assert ended["tgid"] == 1111
     assert ended["end_reason"] == "tgid_change"
-    # The outgoing segment ends exactly where the new one begins — no tail pad,
+    # The outgoing segment is padded PAST where the new one begins. The buffered
-    # or it would swallow the first moments of the new talkgroup.
+    # audio lags control-channel timestamps by ~1.5s, so ending exactly at the
-    assert ended["ended_at_epoch"] == split_time
+    # split cut the outgoing call's last words. The overlap is correct — the
+    # audio stream really does hold one call's tail then the next call's start.
+    assert ended["ended_at_epoch"] == split_time + settings.call_tail_pad_seconds
     assert watcher.on_call_start.call_args[0][0]["started_at_epoch"] == split_time
 
 
@@ -482,7 +557,8 @@ async def test_multiple_call_log_entries_in_one_poll(watcher, clock):
     assert ended["tgid"] == 1111
     assert ended["transmissions"] == 2
     assert ended["started_at_epoch"] == t0
-    assert ended["ended_at_epoch"] == t0 + 0.9
+    # Split close is padded past the new grant — see the tgid_change test above.
+    assert ended["ended_at_epoch"] == t0 + 0.9 + settings.call_tail_pad_seconds
 
 
 @pytest.mark.asyncio
@@ -575,3 +651,499 @@ async def test_end_never_precedes_start(watcher, clock):
 
     payload = watcher.on_call_end.call_args[0][0]
     assert payload["ended_at_epoch"] >= payload["started_at_epoch"]
+
+
+# ---------------------------------------------------------------------------
+# AUDIO MODE — boundaries from the audio, label from the console
+#
+# This is the production path. Everything above this line is the fallback that
+# only runs when PulseAudio capture is down.
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_audio_onset_starts_the_recording(hearing, clock, audio):
+    """Voice onset opens the segment, and the recorder is told to slice to it."""
+    onset = audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now, tgtag="Police Dispatch")],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+
+    assert hearing.is_active
+    hearing.on_call_start.assert_called_once()
+    payload = hearing.on_call_start.call_args[0][0]
+    assert payload["driver"] == "audio"
+    # The recorder slices the ring buffer back to THIS epoch, so it has to be
+    # the audio onset, not the grant and not our detection time.
+    assert payload["started_at_epoch"] == onset
+    assert payload["tgid"] == 1234
+    assert payload["tgid_name"] == "Police Dispatch"
+    assert payload["attributed"] is True
+
+
+@pytest.mark.asyncio
+async def test_a_grant_with_no_audio_does_not_start_a_recording(hearing, clock):
+    """
+    The grant fires 0.84-1.62s before anyone speaks. Opening on it is what put
+    seconds of dead air at the head of every recording.
+    """
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now, tgtag="Fire")],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+
+    assert not hearing.is_active
+    hearing.on_call_start.assert_not_called()
+
+
+@pytest.mark.asyncio
+async def test_recording_closes_after_the_configured_silence(hearing, clock, audio):
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now)],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+
+    clock.advance(0.5)
+    last_voice = audio.speak()
+    await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=555)]))
+
+    # Quiet, but not for long enough yet.
+    clock.advance(settings.call_silence_timeout - 0.5)
+    await tick(hearing, update(channels=[channel()]))
+    assert hearing.is_active
+    hearing.on_call_end.assert_not_called()
+
+    clock.advance(0.6)
+    await tick(hearing, update(channels=[channel()]))
+
+    assert not hearing.is_active
+    payload = hearing.on_call_end.call_args[0][0]
+    assert payload["end_reason"] == "audio_silence"
+    assert payload["tgid"] == 1234
+    # The audio ends where the silence run began plus the threshold, so the
+    # trim can measure and strip exactly that run.
+    assert payload["ended_at_epoch"] == pytest.approx(last_voice + settings.call_silence_timeout)
+
+
+@pytest.mark.asyncio
+async def test_a_false_srcaddr_drop_mid_speech_does_not_end_the_recording(hearing, clock, audio):
+    """
+    THE BUG THIS REARCHITECTURE EXISTS FOR. srcaddr can reset to 0 while someone
+    is still talking; under the old design that started the idle timer and the
+    window closed on top of live speech (recording 0ff35b20: "-1.61s lead,
+    -0.00s tail" — nothing left to trim because the cut landed mid-word).
+    """
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now)],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+    call_id = hearing.active_call_id
+
+    # The control channel now says the call is over. It is wrong; the audio
+    # keeps arriving. This runs well past call_idle_timeout, which is what
+    # would have closed the segment before.
+    for _ in range(10):
+        clock.advance(0.5)
+        audio.speak()
+        await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=0, hold_tgid=1234)]))
+        assert hearing.is_active, "a false srcaddr drop must never end a recording"
+
+    assert (clock.now - hearing._started_at) > settings.call_idle_timeout
+    assert hearing.active_call_id == call_id
+    hearing.on_call_end.assert_not_called()
+
+
+@pytest.mark.asyncio
+async def test_silence_close_logs_the_measured_trailing_silence(hearing, clock, audio, caplog):
+    """
+    call_silence_timeout can only be tuned from the real trailing silence, so
+    the measured number has to reach the log — the audio-mode counterpart of
+    the "measured control-channel idle" line.
+    """
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+
+    with caplog.at_level("INFO", logger="drb-edge-node"):
+        clock.advance(settings.call_silence_timeout + 0.25)
+        await tick(hearing, update(channels=[channel()]))
+
+    lines = [r.message for r in caplog.records if "measured trailing silence" in r.message]
+    assert lines, "audio closes must log the measured silence for later tuning"
+    assert "3.25s" in lines[0]
+
+
+@pytest.mark.asyncio
+async def test_same_tgid_grant_continues_one_audio_recording(hearing, clock, audio):
+    """Back-and-forth on one talkgroup must stay a single call/recording."""
+    onset = audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now)],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+    call_id = hearing.active_call_id
+
+    # The other party keys up on the SAME tgid while audio is still flowing.
+    clock.advance(1.0)
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now)],
+        channels=[channel(tgid=1234, srcaddr=777)],
+    ))
+
+    assert hearing.active_call_id == call_id, "same tgid must not open a new call"
+    hearing.on_call_start.assert_called_once()
+    hearing.on_call_end.assert_not_called()
+
+    clock.advance(settings.call_silence_timeout + 0.5)
+    await tick(hearing, update(channels=[channel()]))
+
+    hearing.on_call_end.assert_called_once()
+    payload = hearing.on_call_end.call_args[0][0]
+    assert payload["call_id"] == call_id
+    assert payload["started_at_epoch"] == onset
+    assert payload["transmissions"] == 2
+
+
+@pytest.mark.asyncio
+async def test_different_tgid_splits_even_with_no_silence_between(hearing, clock, audio):
+    """
+    Back-to-back calls on two talkgroups with no gap. Pure audio segmentation
+    would merge them into ONE file under ONE label, which corrupts correlation.
+    The console talkgroup change has to force the cut.
+    """
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1111, clock.now, tgtag="Fire")],
+        channels=[channel(tgid=1111, srcaddr=1)],
+    ))
+    first_id = hearing.active_call_id
+
+    clock.advance(2.0)
+    audio.speak()  # still talking — no silence anywhere in this test
+    split = clock.now
+    await tick(hearing, update(
+        call_log=[grant(2222, split, tgtag="EMS")],
+        channels=[channel(tgid=2222, srcaddr=2)],
+    ))
+
+    hearing.on_call_end.assert_called_once()
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["call_id"] == first_id
+    assert ended["tgid"] == 1111
+    assert ended["end_reason"] == "tgid_change"
+    # Padded past the split: buffered audio lags the control channel, so cutting
+    # at the exact grant timestamp clipped the outgoing call's last words. The
+    # overlap between the two slices is correct.
+    assert ended["ended_at_epoch"] == split + settings.call_tail_pad_seconds
+
+    assert hearing.is_active and hearing.current_tgid == 2222
+    assert hearing.active_call_id != first_id
+    assert hearing.on_call_start.call_count == 2
+    assert hearing.on_call_start.call_args[0][0]["started_at_epoch"] == split
+
+
+@pytest.mark.asyncio
+async def test_an_unlogged_tgid_change_also_splits_and_reopens(hearing, clock, audio):
+    """
+    OP25's call_log deque is capped at 10, so grants get dropped. If our only
+    receiver is plainly on another talkgroup the segment is over — and in audio
+    mode a new one must open immediately or the audio would be dropped on the
+    floor until the next voice run.
+    """
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1111, clock.now)],
+        channels=[channel(tgid=1111, srcaddr=1)],
+    ))
+    first_id = hearing.active_call_id
+
+    clock.advance(1.0)
+    audio.speak()
+    await tick(hearing, update(channels=[channel(tgid=3333, srcaddr=7)]))  # no call_log
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["call_id"] == first_id
+    assert ended["end_reason"] == "tgid_change_unlogged"
+    assert hearing.is_active and hearing.current_tgid == 3333
+
+
+@pytest.mark.asyncio
+async def test_max_length_backstop_closes_and_reopens(hearing, clock, audio):
+    """
+    A talkgroup that never goes quiet must not produce an unbounded recording —
+    but the audio must not be dropped either, so the segment is immediately
+    reopened rather than simply abandoned.
+    """
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+    first_id = hearing.active_call_id
+
+    clock.advance(MAX_SEGMENT_SECONDS / 2)
+    audio.speak()
+    await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=5)]))
+    assert hearing.active_call_id == first_id
+
+    clock.advance(MAX_SEGMENT_SECONDS / 2 + 1)
+    audio.speak()
+    await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=5)]))
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["call_id"] == first_id
+    assert ended["end_reason"] == "max_length"
+    assert hearing.is_active, "a still-live transmission must not be dropped at the cap"
+    assert hearing.active_call_id != first_id
+    assert hearing.on_call_start.call_count == 2
+
+
+# ---------------------------------------------------------------------------
+# Attribution: resolved at close, from a bounded rolling console history
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_orphan_audio_is_discarded_flagged_and_counted(hearing, clock, audio, caplog):
+    """
+    Audio with no console talkgroup anywhere near it — Liquidsoap fallback, a
+    test tone, stray noise, a dropped call_log. It must be impossible to miss
+    and must never be uploaded: an untagged call poisons correlation.
+    """
+    audio.speak()
+    await tick(hearing, update())  # console says nothing at all
+
+    assert hearing.is_active
+    started = hearing.on_call_start.call_args[0][0]
+    assert started["tgid"] is None
+    assert started["attributed"] is False
+
+    with caplog.at_level("ERROR", logger="drb-edge-node"):
+        clock.advance(settings.call_silence_timeout + 0.5)
+        await tick(hearing, update())
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["attributed"] is False
+    assert ended["tgid"] is None
+    assert hearing.unattributed_segments == 1
+    assert any("ORPHAN AUDIO" in r.message for r in caplog.records), \
+        "unattributed audio must be loud in the logs, not silent"
+
+
+@pytest.mark.asyncio
+async def test_a_grant_before_audio_onset_still_attributes_the_recording(hearing, clock, audio):
+    """
+    The common ordering: the console grants the channel, then 0.84-1.62s later
+    (plus pipeline lag) the audio shows up. The lookback has to cover it.
+    """
+    grant_time = clock.now
+    await tick(hearing, update(call_log=[grant(1234, grant_time, tgtag="Fire")]))
+    assert not hearing.is_active
+
+    clock.advance(2.0)
+    audio.speak()
+    await tick(hearing, update())  # console says nothing NOW
+
+    assert hearing.is_active
+    payload = hearing.on_call_start.call_args[0][0]
+    assert payload["tgid"] == 1234
+    assert payload["tgid_name"] == "Fire"
+
+
+@pytest.mark.asyncio
+async def test_a_grant_after_audio_onset_attributes_the_recording_late(hearing, clock, audio):
+    """
+    There is no guaranteed ordering: the console is polled every 500ms, so the
+    grant can land after voice onset. The segment starts unattributed and picks
+    the talkgroup up part-way through — expected, and fine.
+    """
+    audio.speak()
+    await tick(hearing, update())
+
+    assert hearing.is_active
+    assert hearing.on_call_start.call_args[0][0]["attributed"] is False
+
+    clock.advance(0.5)
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now, tgtag="EMS")]))
+    assert hearing.current_tgid == 1234
+
+    clock.advance(settings.call_silence_timeout + 0.5)
+    await tick(hearing, update())
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["attributed"] is True
+    assert ended["tgid"] == 1234
+    assert ended["tgid_name"] == "EMS"
+    assert hearing.unattributed_segments == 0
+
+
+@pytest.mark.asyncio
+async def test_attribution_resolves_from_a_channel_row_when_the_grant_was_dropped(hearing, clock, audio):
+    """A dropped grant is survivable: an active channel row names the talkgroup."""
+    audio.speak()
+    await tick(hearing, update())
+    assert hearing.on_call_start.call_args[0][0]["tgid"] is None
+
+    clock.advance(0.5)
+    audio.speak()
+    await tick(hearing, update(channels=[channel(tgid=4321, srcaddr=99, tag="Sheriff")]))
+
+    clock.advance(settings.call_silence_timeout + 0.5)
+    await tick(hearing, update())
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["tgid"] == 4321
+    assert ended["attributed"] is True
+
+
+@pytest.mark.asyncio
+async def test_console_activity_outside_the_tolerance_does_not_attribute(hearing, clock, audio):
+    """
+    The tolerance is deliberately bounded. A grant from long before the audio is
+    not evidence about this audio, and borrowing it would be worse than
+    admitting the audio is unattributed.
+    """
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+
+    clock.advance(ATTRIBUTION_LOOKBACK_SECONDS + 5.0)
+    audio.speak()
+    await tick(hearing, update())
+
+    assert hearing.is_active
+    assert hearing.on_call_start.call_args[0][0]["tgid"] is None
+    assert ATTRIBUTION_LOOKAHEAD_SECONDS > 0
+
+
+@pytest.mark.asyncio
+async def test_a_directly_stated_talkgroup_is_not_overruled_at_close(hearing, clock, audio, caplog):
+    """
+    The attribution window extends past the audio on purpose, so a neighbouring
+    call's console activity can fall inside it. An inference over that window
+    must never overrule a talkgroup the console stated outright — but the
+    overlap does mean the split logic missed something, so it is logged.
+    """
+    audio.speak()
+    await tick(hearing, update(
+        call_log=[grant(1111, clock.now, tgtag="Fire")],
+        channels=[channel(tgid=1111, srcaddr=1)],
+    ))
+
+    # A second talkgroup is busy on ANOTHER receiver, so no split fires, and it
+    # produces more console observations than ours did.
+    for _ in range(4):
+        clock.advance(0.5)
+        audio.speak()
+        await tick(hearing, update(channels=[
+            channel(tgid=1111, srcaddr=0, hold_tgid=1111),
+            channel(tgid=2222, srcaddr=7),
+        ]))
+
+    with caplog.at_level("WARNING", logger="drb-edge-node"):
+        clock.advance(settings.call_silence_timeout + 0.5)
+        await tick(hearing, update())
+
+    ended = hearing.on_call_end.call_args[0][0]
+    assert ended["tgid"] == 1111, "the console said 1111 directly; inference must not overrule it"
+    assert any("split logic should have fired" in r.message for r in caplog.records)
+
+
+@pytest.mark.asyncio
+async def test_the_same_voice_run_does_not_reopen_a_second_recording(hearing, clock, audio):
+    """A closed run must stay closed; only NEW audio opens the next segment."""
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+
+    clock.advance(settings.call_silence_timeout + 0.5)
+    await tick(hearing, update())
+    assert not hearing.is_active
+
+    # Several more quiet polls must not resurrect it.
+    for _ in range(3):
+        clock.advance(0.5)
+        await tick(hearing, update())
+    assert not hearing.is_active
+    assert hearing.on_call_start.call_count == 1
+
+    # ...but the next voice run does open a new segment.
+    clock.advance(1.0)
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+    assert hearing.is_active
+    assert hearing.on_call_start.call_count == 2
+
+
+# ---------------------------------------------------------------------------
+# Mode changes: capture loss, console loss
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_losing_capture_closes_an_audio_segment(hearing, clock, audio):
+    """
+    An audio-driven segment must never hang open when the audio stops arriving:
+    with no chunks there is no silence to detect, so the mode change is the
+    thing that has to close it.
+    """
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+    assert hearing.is_active
+
+    audio.capturing = False
+    clock.advance(0.5)
+    await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=555)]))
+
+    assert not hearing.is_active
+    assert hearing.on_call_end.call_args[0][0]["end_reason"] == "capture_lost"
+
+
+@pytest.mark.asyncio
+async def test_console_segmentation_takes_over_while_capture_is_down(hearing, clock, audio):
+    """
+    No audio means no recordings, but the node must still report real radio
+    activity to C2 rather than going silent about it.
+    """
+    audio.capturing = False
+
+    await tick(hearing, update(
+        call_log=[grant(1234, clock.now, tgtag="Police")],
+        channels=[channel(tgid=1234, srcaddr=555)],
+    ))
+    assert hearing.is_active
+    assert hearing.on_call_start.call_args[0][0]["driver"] == "console"
+
+    clock.advance(0.5)
+    await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=0, hold_tgid=1234)]))
+    clock.advance(settings.call_idle_timeout + 0.5)
+    await tick(hearing, update(channels=[channel()]))
+
+    assert not hearing.is_active
+    assert hearing.on_call_end.call_args[0][0]["end_reason"] == "idle_timeout"
+
+
+@pytest.mark.asyncio
+async def test_op25_unreachable_still_closes_an_audio_segment(hearing, clock, audio):
+    """Without the console there is no attribution, so there is nothing to keep open."""
+    audio.speak()
+    await tick(hearing, update(call_log=[grant(1234, clock.now)]))
+
+    clock.advance(OP25_OFFLINE_GRACE + 0.5)
+    audio.speak()
+    await tick(hearing, None)
+
+    assert not hearing.is_active
+    assert hearing.on_call_end.call_args[0][0]["end_reason"] == "op25_unreachable"
+
+
+@pytest.mark.asyncio
+async def test_console_history_is_bounded(hearing, clock, audio):
+    """A rolling history that grows without limit would be a slow memory leak."""
+    from app.internal.metadata_watcher import CONSOLE_HISTORY_MAX, CONSOLE_HISTORY_SECONDS
+
+    for _ in range(CONSOLE_HISTORY_MAX + 200):
+        clock.advance(0.05)
+        await tick(hearing, update(channels=[channel(tgid=1234, srcaddr=5)]))
+
+    assert len(hearing._console) <= CONSOLE_HISTORY_MAX
+
+    # ...and old entries age out even when the count is low.
+    clock.advance(CONSOLE_HISTORY_SECONDS + 1)
+    await tick(hearing, update())
+    assert all(e.epoch >= clock.now - CONSOLE_HISTORY_SECONDS for e in hearing._console)

drb-edge-node/tests/test_pcm.py

@@ -0,0 +1,134 @@
+"""
+Unit tests for the raw-PCM primitives that silence detection rests on.
+
+The whole audio-driven design depends on one field observation: between
+transmissions the captured stream is DIGITAL silence (a PulseAudio null-sink
+monitor), measured at about -91 dBFS — one least-significant bit — not an analog
+noise floor. These tests pin that assumption down in code: a 1-LSB "silent"
+buffer must read as silence at every sane threshold, and speech-level audio must
+never read as silence.
+"""
+from array import array
+
+import pytest
+
+from app.internal import pcm
+
+
+def tone(level: int, samples: int = 1024) -> bytes:
+    """A square wave at +/-level, so RMS == level exactly."""
+    return array("h", [level, -level] * (samples // 2)).tobytes()
+
+
+def zeros(samples: int = 1024) -> bytes:
+    return b"\x00\x00" * samples
+
+
+# ---------------------------------------------------------------------------
+# Format arithmetic
+# ---------------------------------------------------------------------------
+
+def test_capture_format_is_22050_mono_16bit():
+    """MP3_SAMPLE_RATE in call_recorder must match, or the encode resamples."""
+    assert (pcm.SAMPLE_RATE, pcm.CHANNELS, pcm.SAMPLE_WIDTH) == (22050, 1, 2)
+    assert pcm.BYTES_PER_SECOND == 44100
+
+
+def test_seconds_and_byte_offset_round_trip():
+    assert pcm.seconds(pcm.BYTES_PER_SECOND) == pytest.approx(1.0)
+    assert pcm.byte_offset(1.0) == pcm.BYTES_PER_SECOND
+    assert pcm.byte_offset(0.5) == 22050
+
+
+def test_byte_offset_is_always_sample_aligned():
+    """A byte offset that splits a sample would shift every later sample."""
+    for seconds in (0.001, 0.0137, 0.25, 1.7):
+        assert pcm.byte_offset(seconds) % pcm.FRAME_BYTES == 0
+
+
+def test_align_drops_a_trailing_half_sample():
+    assert pcm.align(9) == 8
+    assert pcm.align(0) == 0
+    assert pcm.align(-4) == 0
+
+
+# ---------------------------------------------------------------------------
+# Silence detection
+# ---------------------------------------------------------------------------
+
+def test_exact_digital_zero_is_silence():
+    assert pcm.is_all_zero(zeros())
+    assert pcm.rms_dbfs(zeros()) == pcm.SILENT_DBFS
+    assert pcm.is_silent(zeros(), -50.0)
+    assert pcm.is_silent(zeros(), -90.0)
+
+
+def test_one_lsb_of_dither_is_the_measured_field_floor():
+    """
+    The gap between transmissions measures ~-91 dBFS on a live node, which is
+    exactly 20*log10(1/32768) — a single LSB. It must read as silence at any
+    threshold we would ever configure.
+    """
+    floor = tone(1)
+    assert pcm.rms_dbfs(floor) == pytest.approx(-90.3, abs=0.2)
+    assert pcm.is_silent(floor, -50.0)
+    assert pcm.is_silent(floor, -70.0)
+    assert not pcm.is_silent(floor, -95.0), "an absurd threshold should still be honoured"
+
+
+def test_speech_level_audio_is_never_silence():
+    """Speech on the live node averages about -18 dBFS."""
+    speech = tone(4096)  # -18.06 dBFS
+    assert pcm.rms_dbfs(speech) == pytest.approx(-18.06, abs=0.1)
+    assert not pcm.is_silent(speech, -50.0)
+    assert not pcm.is_silent(speech, -40.0)
+
+
+def test_threshold_is_honoured_exactly_at_the_boundary():
+    # RMS 104 -> -49.96 dBFS, just above a -50 threshold.
+    assert not pcm.is_silent(tone(104), -50.0)
+    # RMS 100 -> -50.30 dBFS, just below it.
+    assert pcm.is_silent(tone(100), -50.0)
+
+
+def test_empty_buffer_counts_as_silence():
+    """
+    "No audio arrived" must never read as "someone is talking" — otherwise a
+    stalled capture would hold a segment open forever.
+    """
+    assert pcm.is_silent(b"", -50.0)
+    assert pcm.rms_dbfs(b"") == pcm.SILENT_DBFS
+
+
+def test_full_scale_is_zero_dbfs():
+    assert pcm.rms_dbfs(tone(32767)) == pytest.approx(0.0, abs=0.001)
+
+
+def test_a_trailing_odd_byte_does_not_break_detection():
+    """Short reads at EOF can leave half a sample; it must be dropped, not skew."""
+    assert not pcm.is_silent(tone(8000) + b"\x00", -50.0)
+    assert pcm.samples(zeros(4) + b"\x01").itemsize == 2
+    assert len(pcm.samples(zeros(4) + b"\x01")) == 4
+
+
+def test_rms_attenuates_an_isolated_click_the_way_peak_would_not():
+    """
+    Why RMS and not peak. A single stray sample in an otherwise silent window is
+    a decoder click, not speech. Peak would score it at its full amplitude and
+    hold a recording open; RMS spreads it over the window and divides it down by
+    sqrt(N) — 30 dB for a 1024-sample window.
+
+    A full-scale click still reads as signal even after that attenuation, which
+    is deliberate: at worst it extends a recording by the silence timeout, and
+    the trim strips the result before upload. Under-detecting speech is the
+    failure that loses words permanently.
+    """
+    moderate = bytearray(zeros(1024))
+    moderate[0:2] = array("h", [1000]).tobytes()  # -30 dBFS peak
+    assert pcm.rms_dbfs(bytes(moderate)) == pytest.approx(-60.4, abs=0.2)
+    assert pcm.is_silent(bytes(moderate), -50.0)
+
+    full_scale = bytearray(zeros(1024))
+    full_scale[0:2] = array("h", [32767]).tobytes()
+    assert pcm.rms_dbfs(bytes(full_scale)) == pytest.approx(-30.1, abs=0.2)
+    assert not pcm.is_silent(bytes(full_scale), -50.0)

drb-edge-node/tests/test_pulse.py

@@ -17,6 +17,8 @@ is proven at the layer that matters.
 import subprocess
 from unittest.mock import Mock
 
+import pytest
+
 from app.internal import pulse
 
 
@@ -53,6 +55,7 @@ def test_is_ready_false_when_daemon_does_not_respond(monkeypatch):
     assert pulse.is_ready() is False
 
 
+@pytest.mark.asyncio
 async def test_wait_until_ready_short_circuits_when_already_live(monkeypatch):
     calls = Mock(return_value=True)
     monkeypatch.setattr(pulse, "_daemon_responds", calls)
@@ -60,6 +63,7 @@ async def test_wait_until_ready_short_circuits_when_already_live(monkeypatch):
     assert calls.call_count == 1
 
 
+@pytest.mark.asyncio
 async def test_wait_until_ready_polls_until_daemon_comes_up(monkeypatch):
     monkeypatch.setattr(pulse, "POLL_INTERVAL", 0.01)
     responses = iter([False, False, True])
@@ -67,12 +71,14 @@ async def test_wait_until_ready_polls_until_daemon_comes_up(monkeypatch):
     assert await pulse.wait_until_ready(timeout=5) is True
 
 
+@pytest.mark.asyncio
 async def test_wait_until_ready_times_out_when_daemon_never_responds(monkeypatch):
     monkeypatch.setattr(pulse, "POLL_INTERVAL", 0.01)
     monkeypatch.setattr(pulse, "_daemon_responds", lambda: False)
     assert await pulse.wait_until_ready(timeout=0.05) is False
 
 
+@pytest.mark.asyncio
 async def test_wait_until_ready_uses_settings_default_when_timeout_omitted(monkeypatch):
     monkeypatch.setattr(pulse.settings, "pulse_wait_timeout", 0.05)
     monkeypatch.setattr(pulse, "POLL_INTERVAL", 0.01)