Skip to content

feat(transcription): one speaker label per voice across the parts of an episode - #542

Merged
chhoumann merged 66 commits into
masterfrom
feat/transcription-speaker-merge
Nov 16, 2026
Merged

chhoumann merged 66 commits into
masterfrom
feat/transcription-speaker-merge

Conversation

@chhoumann

@chhoumann chhoumann commented Nov 14, 2026 •

Copy link
Copy Markdown
Owner

Summary

OpenRouter takes at most 18 MB and 1,800 s a request, so PodNotes uploads a long episode in parts, and each part numbers its speakers on its own. On the 54 min E231 in 3 parts, those labels agreed with a whole-file run 0.344 of the time. With this change, Grok STT 2.0 gives each voice one number across all parts of an episode. Builds on #540.

  • After the parts, PodNotes sends one more request, the reel. It holds one clip per part-local speaker label, from that label's longest turn of at least 2.6 s, starting 0.2 s in and at most 9 s long. PodNotes cuts the clips at MP3 frame starts and joins them with 1.5 s of silent frames. Each speaker in Grok's answer gets credit for the seconds it speaks over each clip. Part labels whose clips go to one speaker become one, and a label with no clip keeps its own number.
  • mergeSpeakers on the model profile turns the reel on. It is on for x-ai/grok-stt-2.0 and off for elevenlabs/scribe-v2 and microsoft/mai-transcribe-2.
  • The reel runs only with speaker labels on, more than one part and MP3 audio. Audio decoded to WAV parts keeps the restart warning. A failed reel request keeps the transcript and the restart warning and never fails the job. A merged transcript has no restart warning.

Speaker accuracy of the lab implementation (diarize-lab/scripts/step0.mjs) through OpenRouter:

Episode Reference Parts as they come With the reel Whole-file Grok
E231, 3 parts at the product's caps whole-file Scribe v3 0.344 0.995
Planet Money, 300 s parts human transcript 0.367 0.949 0.942
Up First, 300 s parts human transcript 0.753 0.830 0.831

Scribe v2's reel heard Planet Money's 10 voices as 4 (0.378 to 0.414), so it stays off. MAI-Transcribe 2 is not measured. This PR's code has not run against the live API, because this session has no key.

Verification

  • npm run test passes 2,888 tests. The new tests send MP3 fixtures whose frames carry a voice and a frame index through the requestUrl seam. The mock provider hears each frame's voice and numbers voices in the order they speak in each request, as a diarizing model does. On the base, the 4 tests for new behavior fail, for example with 1 Ann, 2 Bob, 3 Bob, 4 Ann where 1 Ann, 2 Bob, 2 Bob, 1 Ann is expected.
  • Each plant below failed its test (1 failed) and passed once restored (1 passed), through npm run plant:
Plant Caught by
The note ignores the matched labels Grok STT 2.0 numbers each voice once across two parts... (3 Bob for 2 Bob)
No silence between clips the reel is the stream's own frames...
A label without a clip takes the speaker matched last a voice whose turns are all under 2.6 s... (2 Cy for 3 Cy)
Every turn gets a clip, short ones too the same test (the reel held Cy's frames)
A reel failure ends the job, as it does for a part a reel OpenRouter refuses keeps the parts' labels...
The reel runs for every model the Scribe v2 and MAI-Transcribe 2 tests, one run each
The reel runs with labels off with speaker labels off, Grok STT 2.0 sends no reel...
The reel runs for one part an episode in one part sends no reel
  • lint, format:check, typecheck, build, test:tooling (48 passed) and docs:build pass. npm run test:e2e -- tests/e2e/journey-resilience.test.ts tests/e2e/journey-transcript.test.ts passes 26 tests. npm run size reports 106,873 gzip bytes. The budget rises by the measured 808 bytes, against 106,065 for the base built the same way, and keeps the base's 97 bytes of headroom.
  • Obsidian 1.14.4 on Linux, in this worktree's isolated instance, with the verify skill's fixture provider and no real key. The fixture now decodes Grok's MP3 uploads with Chromium's decoder and hears a 300 Hz tone and a 700 Hz tone as two speakers. The local file Fixtures/Two Voices.mp3 is 70 min of 8 kbps, 8 kHz mono, with the two tones taking turns every 45 s. With Grok and labels on, Transcribe current episode from audio sent three part POSTs of about 1.87 MB each, then one reel POST of 82,291 bytes that Chromium decoded to 61.56 s, which is six 9 s clips and five gaps of 21 silent frames. The second part starts with the high tone and numbers the voices the other way. The note still numbers all 47 low-voice paragraphs 1 and all 47 high-voice paragraphs 2, and the paragraph at each part boundary runs on across it. No notice stayed. On Scribe v2 the same file went as three multipart POSTs with no reel, the note numbered six speakers, and the restart warning stayed. dev:errors reported none.

Grok note at the first part boundary

Scribe v2 note with the restart warning

Real OpenRouter key (lead, after review)

Obsidian 1.14.4, Linux, build of 5eebf37b, Grok STT 2.0 (recommended), speaker labels on, the 54 min, 39 MB local E231 interview (two voices), palette command, a pass-through request log:

  • One catalog GET, three part POSTs (18.1, 17.0 and 16.9 MB of JSON), then one reel POST of 994,643 bytes. Done in 48 s, Transcription completed and saved., no restart warning, dev:errors none.
  • The note has exactly two speakers (60 turns of 1:, 65 of 2:), alternating 1 2 1 2 through all three parts. The same episode on feat(transcription): transcribe with OpenRouter #540 got new labels in each part.
  • Identity holds across parts: the host's introduction in part 1 ("My name is ...") and his sign-off in part 3 ("Amazing ... good to have you on") are both 1:.

Release impact

None for settings, storage, the API or URIs. A Grok transcript of an episode over 18 MB or 1,800 s costs one more OpenRouter request of about a minute of audio. OpenRouter is unreleased, so no released behavior changes.

Note

Add speaker-reel matching to merge speaker labels across transcription parts

  • Multi-part transcription now sends an extra speaker-matching request for models with mergeSpeakers: true (Grok STT 2.0 only). The reel is built from each speaker's longest turn (≥2.6s, trimmed to 9s) and the response maps part-specific labels to one shared number in transcribe.ts and cues.ts
  • Reel construction lives in speakers.ts and split.ts: turns group same-speaker words within 2s gaps, and cutReel assembles MP3 clips with generated silent frames between them
  • If the reel request fails with a TranscriptionError, part transcripts still return and unmapped labels keep the existing label-restart warning
  • The multipart label-restart warning now fires only for labels without an equivalence mapping, not for any labeled words
  • Raises the gzip budget for root main.js from 106,562 to 107,399 bytes in size.mjs to cover the reel code

Macroscope summarized 5958579.


Note

Medium Risk
Adds an extra provider request and MP3 slicing on multi-part Grok jobs; failures degrade to prior per-part labels rather than failing the job.

Overview
Multi-part OpenRouter transcriptions can now keep one speaker number per voice across the whole episode when the model supports it (enabled for Grok STT 2.0 only via mergeSpeakers).

After all parts finish, the pipeline may send an extra speaker reel: a short MP3 built from each part-local speaker’s longest turn (≥2.6 s, trimmed, clipped at frame boundaries, separated by 1.5 s silence). The reel response is mapped back to part labels through speakerKey / SameSpeaker, and transcriptFromParts applies that mapping when numbering cues. Scribe v2 and MAI-Transcribe 2 keep mergeSpeakers: false because reel matching was unreliable in measurement. A failed reel still saves the transcript and falls back to the existing “labels restart in each part” warning only for labels that were not merged.

Verify tooling’s fixture provider decodes Grok MP3 uploads (tone-based diarization) so e2e can assert reel requests. Bundle gzip budget rises to 107,399 bytes for the new reel code.

Reviewed by Cursor Bugbot for commit 5958579. Bugbot is set up for automated code reviews on this repo. Configure here.

gpt-4o-transcribe-diarize rejects requests with more than 1400 s of audio,
but MP3 chunks were split by size alone. Any MP3 under about 120 kbps put
more than 1400 s in a 20 MB chunk, so every chunk failed.

For diarization, split MP3 at frame boundaries into the fewest equal parts
that stay under both 20 MB and 1400 s. Whisper keeps its size-only chunks.

Fixes #532
A chunk that OpenAI rejected with a status such as 400 or 401 was sent
three times, although each send fails the same way. A long diarized
episode went up three times before the error showed.

Retry only the statuses openai-node retries: 408, 409, 429 and 5xx, plus
network errors.

Refs #532
gpt-4o-transcribe-diarize takes up to 0.58 s per second of audio, measured
over 81 requests, so the 1,072 s parts a 54 min episode got would take about
10 min each, at the request timeout. Parts are now at most 600 s, which the
model answers in under 6 min.

Refs #532
Each test passes on the hand-written code. Together they catch wrong
derivations that no existing test sees: the -1 SecretStorage ID suffix
that released versions accept, the provider an export error names, the
kind a settings key row adopts a secret for, and the key migration
notices.
CREDENTIALS in src/types/Credentials.ts holds each kind's settings
field, export key and label. The credential repository, settings
decoder, settings transfer, settings tab and key migration notice loop
over it. IPodNotesSettings takes its key fields from it, so a kind with
no default in DEFAULT_SETTINGS fails typecheck. The plaintext key map
from before 2.19.0 stays fixed.
The published transcript renderer is the reader of the format that stays
when the diarization folder goes. The default template now lives only in
DEFAULT_SETTINGS, so src/constants.ts imports no service code.
src/services/transcription/ turns an episode into a transcript with any
provider. A provider is one file: its models, its request, and how to read
its answers. OpenAI (gpt-transcribe) and Deepgram (nova-3) are the first two.

The engine owns what providers share:

- audioSource.ts decides whether a provider may be given the episode's
  address. Only an address that a saved public feed lists, that carries no
  credential, and that names a public host passes. All other audio is
  uploaded.
- split.ts is audioChunker.ts with the caps as arguments and a start time
  on each part.
- http.ts sends a request at most three times, and only after a 429, a 5xx,
  or a lost connection. Its failure sentences never include the provider's
  own text, which can quote a private address.
- cues.ts turns words or text into the shapes of a published transcript.

Nothing calls the engine yet. Lint rejects an import of the folder's files
from outside it, except its index.
TranscriptionService keeps its queue, the published-transcript path, the
progress notice, and note saving. It calls the engine for audio.

- OpenAI transcribes with gpt-transcribe. whisper-1 and
  gpt-4o-transcribe-diarize, which OpenAI shuts down on 2027-02-26, are
  gone. A tooling test searches the built main.js for them.
- With speaker labels on and OpenAI selected, the transcript is saved
  without labels. A notice that stays until clicked names the providers
  that label speakers.
- Deepgram transcripts go through renderTranscriptBody, so each paragraph
  has a timestamp. Speakers are numbered in the order they first speak.
- Deepgram gets the address of a public episode and downloads it itself.
- A failure notice stays until clicked. Parts go out one at a time, and a
  401, 402, 404, or 400 stops the job before the next part.

Deletes src/services/diarization/, openaiTranscription.ts, the Whisper
worker pool, and createBinaryChunkFiles. getEpisodeAudioBuffer no longer
returns a basename, because uploads are named for their format only.
net.mjs answers OpenAI with { text } and Deepgram with the words of two
speakers, and logs the content type of each request. fixtures.mjs serves a
feed at priv3.supercast.com, an address that PodNotes keeps private. The
transcripts recipe covers OpenAI in parts, and Deepgram by address and by
upload.
The marketplace review test called TranscriptionService.waitForRetry, which
the engine replaced. An aborted run now rejects through waitForLifecycle, so
the test aborts a pending wait there and expects the same Error.
An aborted job no longer downloads the episode or decodes it into parts.
Until now only its requests were refused.

A 429 or a 5xx that asks to wait over a minute is sent once. A sooner send
would only upload the audio to be refused again.
A provider's name was written twice: once in its credential row and once in
its provider file. The row now owns it, so a provider file holds only its
models and its request. Each model gains the name the settings will show.
The provider was derived from the speaker settings: Deepgram with speaker
labels on, OpenAI otherwise. It is now stored as transcript.provider, with
transcript.model beside it, where "" follows the provider's recommended model.
schemaVersion stays 2 and both fields are additions, so a device on 2.24.0
keeps loading the library and keeps the fields on its own saves.

A library that has no transcript.provider gets the provider it ran with:
Deepgram for speaker labels with Deepgram, with or without its key, and OpenAI
otherwise. The migration reads only the stored document, adds no decode
warning, and gives the same answer on every load until a save stores it.

transcript.diarization.provider belongs to versions up to 2.24.0. This version
reads it for that migration alone, never writes it, keeps it in data.json as
found, and leaves it out of settings exports. Importing a 2.24.0 export that
labels speakers with Deepgram selects Deepgram.

A provider or model ID this version does not have is kept as stored and shown
under that name. An unknown provider refuses to transcribe with a notice. An
unknown model runs the recommended model and the notice says so.

The Transcripts settings now read: Transcription provider, the chosen
provider's API key, Model, Speaker labels, Speaker label format, then the path
and template. Speaker labels is disabled for a model that cannot label
speakers. The Diarization provider row is gone.
ElevenLabs is a credential row, one provider file and a registry line. Its key
reference is the new root field elevenLabsSecretId, stored in SecretStorage as
podnotes-elevenlabs-api-key, and a settings export carries the key as
secrets.elevenLabs. The export envelope stays at version 2.

A public episode goes to ElevenLabs by its address, so nothing is downloaded.
Any other episode is uploaded in one request. Word times and speakers reach
the note, and turning speaker labels off asks for none.

ElevenLabs answers 401 for a used-up quota as for a rejected key, and 400 for
a model the key may not use as for an address it could not download, so the
provider reads detail.code. Only the last of these is followed by an upload.
A request waits 10 minutes for its answer.

A new install starts on ElevenLabs with speaker labels on. A library saved up
to 2.24.0 that holds no OpenAI key and has speaker labels off moves to
ElevenLabs too, because it never set transcription up. A library that never
stored the speaker labels choice keeps them off.

The notice for a model without speaker labels now names ElevenLabs first.
A journey now seeds its transcription provider with its key, because a new
vault is on ElevenLabs. The OpenAI journeys choose OpenAI. A new journey runs
the default: a public episode goes to ElevenLabs by its address, no audio is
fetched, and the note has timestamps and speaker labels. The ElevenLabs stub
answers with the words ElevenLabs returns and records what each request named.

The settings tab test walks the new Transcripts rows, and the key recovery
test links the ElevenLabs key, the one a new vault shows.
The fixture provider answers api.elevenlabs.io with ElevenLabs' words, two
speakers with the second 12 s in, and logs the address a request named in
place of a file. The transcript guide follows the new settings rows and adds
the ElevenLabs runs and a 2.24.0 library check.
…choice

The provider row decided for itself whether the stored provider is one this
version has, and the Speaker label format row read the stored speaker setting.
Both now ask transcriptionChoice, as the key rows, the model row and the
Speaker labels row do. The engine no longer exports a type nothing imports.
The Transcripts tests looked each control up by a CSS selector inside its row
and asserted on a hidden row's inline style. They now query by role and by the
row's name, as a user or a screen reader finds a control, so a hidden row is
one the query does not find.
…ings measure

The transcription engine left main.js at 103,795 gzip bytes. The provider and
model settings, their migration and the ElevenLabs provider add 866, which
takes it to 104,661 and over the 104,000 budget. The budget rises by those 866
bytes to 104,866, so the 205 bytes of headroom stay as they were.
…m library

2.24.0 imports the diarization object whole and fills a missing provider
with OpenAI, so omitting the field moved Deepgram libraries to OpenAI.
Imports here keep the provider this library stored.
A library migrated with OpenAI speaker labels on could not turn them off,
so every transcription repeated the notice.
… one

The tests for a provider and a model a newer PodNotes stored used OpenRouter
and Grok STT 2.0, which this version is about to run. They now use AssemblyAI
and universal-3-6-pro.
OpenRouter is a credential row, one provider file and a registry line. Its
key reference is the new root field openRouterSecretId, stored in
SecretStorage as podnotes-openrouter-api-key, and a settings export carries
the key as secrets.openRouter.

It offers three models, each with the request measured to give word times
and speakers:

- Grok STT 2.0, recommended: base64 JSON with verbose_json and word and
  segment times, routed to xAI with provider.only, which takes diarize from
  provider.options.xai. A request routed elsewhere is answered with no
  speakers.
- ElevenLabs Scribe v2: multipart with diarize beside the model, or JSON with
  input_audio.url for an episode a saved public feed lists.
- MAI-Transcribe 2: multipart with diarize beside the model.

With speaker labels off, no request carries diarize. Sounds Scribe v2 lists
as audio_event items are left out.

OpenRouter takes 25 MB a request and gives the model 60 s, so an upload goes
in parts of at most 18 MB and 1,800 s. Parts of 24 MB were answered in 20 to
31 s.

A 402 says the account is out of credit, a 401 that the key was rejected,
and a 404 that the model is not offered. A 404 whose message names the
account's data policy or zero data retention says the model is not offered
under the account's OpenRouter privacy settings. Each ends the job at the
first part, and none quotes OpenRouter.
…t each part

A note saved with speaker labels asked for and none returned now says so:
"Saved without speaker labels, because Grok STT 2.0 returned none." OpenRouter
answers 200 with no speakers when a request reaches a route that ignores the
speaker option, so this is checked for every provider.

Labels from two parts already name different speakers. An episode sent in
parts now also says that its labels restart in each part, so one person can
have several labels, until labels are merged across parts.
…ll lists

PodNotes reads OpenRouter's public catalog of transcription models, without
the key, when the Model row draws and when an OpenRouter job starts. It keeps
the read in memory for 24 hours. A read that fails is not kept, so the next
one asks again, and the tested list is offered as it is.

The catalog only takes models away. A model it does not list is not offered,
and one whose expiration_date has come is skipped. A model with a later end
date reads "Grok STT 2.0, ends <date>" in the Model row. A catalog that would
leave no model is not followed.

A chosen model the catalog no longer lists stays selected in the Model row as
"<label>, no longer offered". A job runs the first offered model instead and
says "OpenRouter no longer offers ElevenLabs Scribe v2, so PodNotes used Grok
STT 2.0." The row redraws when a read lands while it is open.

A provider gets its catalog through an optional catalog() beside failure(),
so the registry owns the cache and the filter for any provider that has one.
A stub answers OpenRouter's transcriptions with two speakers, the second 12 s
in, and its model catalog with the three models PodNotes offers. It records
what each request named and a hash of the audio it carried, decoded from
base64 for JSON.

J-TR-09 plays a public episode on Scribe v2: one JSON request names the
episode's address, and the episode's audio is never fetched. J-TR-10 plays it
on the recommended Grok STT 2.0: one JSON request carries the audio as base64,
the same bytes the fixture server holds. Both notes have timestamps and
numbered speakers, and the catalog is read once, with no Authorization.

A journey seeds any provider by its credential row, and a model with it.
…eads them

The fixture provider answers openrouter.ai transcriptions with verbose_json
words from two speakers, the second 12 s in, and the model catalog with the
three models PodNotes offers. The log keeps openrouter.ai paths and reads the
address a JSON request names in input_audio.url.

The transcript guide adds the OpenRouter rows, its runs (Scribe v2 by
address, Grok by base64, a long MP3 in parts) and the notices that now name
OpenRouter.
Frame-like bytes inside a leading tag set the first frame's duration, and
counting every part in that frame's length sent a 1,728 s episode of 72 ms
frames up in one request. Parts now fit by the summed duration of their own
frames, and each part's start is the sum of the frames before it.
Audio frames inside a tag's declared size were counted and could be cut
through, sending an undecodable first part, and frame-like bytes in a
trailing tag added a part.
…engine

# Conflicts:
#	src/services/TranscriptionService.test.ts
#	src/services/TranscriptionService.ts
#	src/services/audioChunker.test.ts
#	src/services/audioChunker.ts
#	src/services/diarization/openaiProvider.ts
#	src/services/diarization/types.ts
#	src/services/openaiTranscription.test.ts
#	src/services/openaiTranscription.ts
#	tests/mocks/openaiFailures.ts
…h cap

A file in a format other than MP3 under the size cap went whole whatever its
length, so a long .m4a passed a model's maxSeconds. Its length now comes from
its metadata through a media element, without decoding it, and audio over the
cap is decoded into WAV parts that fit both caps. Audio of an unknown length
still goes whole.
A public feed's enclosure with AWS or Google Cloud signing parameters passed the
feed credential check, so Deepgram got the whole signed address. A query
parameter whose name holds signature, credential or token, or Akamai's hdnts,
now uploads the audio instead.
…nload

A WebView may load no metadata for media it does not play, so the length probe
could wait forever, and an unload could not stop it before a decode. The probe
now reads the length as unknown after 5 s or when the job is aborted, and an
aborted job decodes nothing.
…nscription-providers

# Conflicts:
#	AGENTS.md
#	scripts/size-baseline.md
#	scripts/size.mjs
#	scripts/size.test.ts
#	src/persistence/settingsCodec.ts
#	src/settingsTransfer.ts
#	src/types/Credentials.ts
#	src/ui/settings/PodNotesSettingsTab.ts
…nscription-openrouter

# Conflicts:
#	AGENTS.md
#	scripts/size-baseline.md
#	scripts/size.mjs
#	scripts/size.test.ts
…address

An address goes in one request, so Scribe v2 on OpenRouter got a public episode
of any length by address, past the 1,800 s its parts are capped at for
OpenRouter's 60 s limit. An episode the feed says is longer than the model's
cap is now uploaded in parts. One of unknown length still goes by address.
Base automatically changed from feat/transcription-openrouter to master November 16, 2026 09:58
…nscription-speaker-merge

# Conflicts:
#	scripts/size-baseline.md
#	scripts/size.mjs
#	scripts/size.test.ts
Comment thread src/services/transcription/split.ts Resolved
@chhoumann

Copy link
Copy Markdown
Owner Author

Retargeted to master after #540 merged (a152aac), and brought up to date (2ec86af). I merged #540's last head, df7e41c, then recorded master as merged with -s ours. git diff df7e41c3 origin/master is empty. Only the size files conflicted. They keep master's entries and add one for this PR: 831 gzip bytes on top of master's 106,562, measured, so the budget is 107,393. Raw is 328,665 of 330,000.

  • Checks: lint, format:check, typecheck, build, size, test (2,918), test:tooling (48).
  • Obsidian 1.14.4, fake keys, fixture provider. OpenRouter on a 70 min local MP3 through "Play with PodNotes" and "Transcribe current episode from audio":
    • one catalog request, 3 parts of about 1.87 MB, then one 40,243-byte reel request
    • the note gives the same voice **1:** in all three parts, at 0 s, 1,400.362 s and 2,800 s
    • the notice reads "Transcription completed and saved." with no restart warning
    • no runtime errors
  • Not repeated: the real Grok reel against OpenRouter. No key was at hand.

@chhoumann

Copy link
Copy Markdown
Owner Author

FAIL
Verified head 2ec86af by an independent verifier (GPT 6 Astra).

  • P2, reproduced defect: turnsOf joins same-speaker words across silence. With words at 0–0.25 s and 30–35 s, it selects 0.216–9.216 s instead of the usable five-second turn. The captured reel's first nine seconds decode to exact silence. The note consequently labels the same low voice 1 in part 1 and 3 later. This independently confirms the unresolved finding. Split turns at substantial gaps before choosing a clip.
  • Compared origin/master at a152aac2 against the exact head in isolated Obsidian 1.14.4 on Linux. Ran 25 base and 45 head cases through settings, file-menu playback, palette/direct transcription commands, and the copied verify fixture. Only fake keys were used. Fixtures included a 70-minute two-tone MP3, 40-minute M4A, short MP3, sparse MP3, and http://podnotes.test:27734/feed.xml.
  • Normal Grok multipart MP3 changed from six labels to two consistent voices. Verified the reel body byte-for-byte against six nine-second source clips and five 1.512-second silence gaps. Its decoded duration is 61.56 s. Saved timestamps remain ordered; note-link and direct URI seeking worked.
  • Scribe v2, MAI, multipart M4A, single-part audio, and labels-off cases sent no reel. Scribe v2's public episode used its address without downloading audio. Expected labels and notices matched the base.
  • Missing one reel clip's words preserved its separate label and warning; an empty answer preserved all part labels. Reel HTTP 400/401/402/403/404/408/409/413/422/429/500/502/503/504, both 404 classifications, a disconnected transport, and a timeout all preserved the transcript and restart warning. Timeout used a shortened timer. Retriable failures made three reel attempts; no provider error text leaked.
  • All requested checks passed with Node 22.23.2: npm run lint, format:check, typecheck, build, size, test (2,918 tests), and test:tooling (48 tests). Bundle is 328,665 raw bytes and 107,393 gzip bytes, exactly at the gzip budget.
  • dev:errors reported none. Evidence and repeatable drivers are in this review worktree's .obsidian-e2e-artifacts/review542/. No tracked files changed. The fixture server and isolated Obsidian instance are stopped; the stop dry run reports pids: none.

One speaker's words across a long silence made one turn, so the reel could
clip the silence instead of the voice, and that voice kept two labels.
@chhoumann

Copy link
Copy Markdown
Owner Author

PASS
Verified head f597f71 by an independent verifier (GPT 6 Astra).

  • Compared master a152aac2 against this head in isolated Obsidian 1.14.4 on Linux. Validated 28 base and 47 head cases through settings, file-menu playback, palette/direct transcription commands, captured requests, and saved notes. Only fake keys and the verify fixture provider were used; no live API accuracy claim.
  • The previous silent-clip defect is fixed. The sparse MP3 now contributes its first part's low voice from 30.240–34.848 s, independently decoded as audible low-tone speech. Both voices keep one label across all three parts, with no restart warning. One-second words separated by half-second pauses also merge correctly. For that fixture, the copied detector uses spectral energy to avoid mistaking partially silent high-tone windows for the low voice.
  • Normal and brief-pause 70-minute MP3s send three parts plus one 61.56 s reel. All six clips match the expected source frames byte-for-byte, separated by five 1.512 s silent-frame gaps; independent decoding confirms the intended voice in every clip. Saved timestamps stay ordered. Note-link and direct URI seeking work on both builds.
  • Scribe v2, MAI, the 40-minute M4A split into WAV parts, single-part MP3s, and labels-off cases send no reel and retain their expected labels/notices. Scribe v2 sends the public episode at http://podnotes.test:23486/episode-1.wav by address without downloading its audio.
  • Partial matches preserve the unmatched label and warning; an empty reel answer preserves all six part labels. Reel HTTP 400/401/402/403/404/408/409/413/422/429/500/502/503/504, both 404 classifications, disconnected transport, and timeout all preserve the transcript and warning. Timeout used a shortened timer. Distinct fixture voices never share a label in the validated cases.
  • All required checks pass with Node 22.23.2: npm run lint, npm run format:check, npm run typecheck, npm run build, npm run size, npm run test (2,920 tests), and npm run test:tooling (48 tests). Bundle size is 328,677 raw bytes and 107,399 gzip bytes, exactly at the gzip budget.
  • dev:errors reports none. No tracked files changed. Repeatable drivers, request bodies, clip proofs, notes, and screenshots are under this review worktree's .obsidian-e2e-artifacts/review542/. The fixture server and isolated Obsidian are stopped; the stop dry run reports pids: none.

@chhoumann
chhoumann merged commit f457c5b into master Nov 16, 2026
18 checks passed
@chhoumann
chhoumann deleted the feat/transcription-speaker-merge branch November 16, 2026 10:39
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant