Repository navigation
fix(codecs): respect finite WAV data sizes - #8455
Draft
peterolkhov wants to merge 1 commit into
Draft
peterolkhov wants to merge 1 commit into
peterolkhov wants to merge 1 commit into
Conversation
The inline decoder treated footer chunks as PCM after a bounded data chunk, adding non-audio samples and extra playback. Limit framing to the declared payload while preserving streaming size markers and per-file reset.
|
|
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The inline WAV decoder reads through the end of the file after a
dataheader. A valid footer chunk therefore becomes PCM, adding nonzero samples and extra playback to an otherwise silent clip.Limit positive finite data payloads before framing and resampling. Zero and
0xFFFFFFFFkeep their existing unknown-length behavior, and a fresh RIFF header resets the budget for the next WAV.The regressions fail on the unmodified base and pass with the fix, covering whole and split input, a partial final frame, and public TTS emitter bytes and duration.
make checkpasses;make unit-tests PYTEST_ARGS='-n 2 -ra'passes with 7,180 passed and five cloud/platform-dependent skips on macOS/Python 3.13. The local room fixture used the official LiveKit server v1.13.11 built from its pinned source.AI assistance: OpenAI Codex (GPT-6.1 Sol, high reasoning) investigated, implemented, and tested this change. No human code review is claimed.
Draft pending the contributor agreement.