Skip to content

Ensure resampling is skipped for empty audio frames #1044

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
lukasIO merged 2 commits into from
Feb 12, 2026
Merged

Ensure resampling is skipped for empty audio frames #1044

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
38d3e79
Ensure resampling is skipped for empty audio frames
lukasIO Feb 11, 2026
03114de
Create fair-needles-obey.md
toubatbrian Feb 11, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions .changeset/fair-needles-obey.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
---
"@livekit/agents": patch
---

Ensure resampling is skipped for empty audio frames
5 changes: 5 additions & 0 deletions agents/src/stt/stt.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -351,6 +351,11 @@ export abstract class SpeechStream implements AsyncIterableIterator<SpeechEvent>
}
}

if (frame.samplesPerChannel === 0) {
this.input.put(frame);
return;
}

if (this.resampler) {
const frames = this.resampler.push(frame);
for (const frame of frames) {
Expand Down
12 changes: 10 additions & 2 deletions agents/src/utils.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -651,14 +651,22 @@ export function resampleStream({
let resampler: AudioResampler | null = null;
const transformStream = new TransformStream<AudioFrame, AudioFrame>({
transform(chunk: AudioFrame, controller: TransformStreamDefaultController<AudioFrame>) {
if (chunk.samplesPerChannel === 0) {
controller.enqueue(chunk);
return;
}
if (!resampler) {
resampler = new AudioResampler(chunk.sampleRate, outputRate);
}
for (const frame of resampler.push(chunk)) {
controller.enqueue(frame);
}
for (const frame of resampler.flush()) {
controller.enqueue(frame);
},
flush(controller) {
Copy link
Copy Markdown
Contributor Author

@lukasIO lukasIO Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This moves the flushing to the flush of the transform stream.

I don't know if the previous implemenation was intended to always flush after every insertion?

@toubatbrian do you happen to remember why this code path is flushing after each chunk?

Copy link
Copy Markdown
Contributor

@toubatbrian toubatbrian Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here is the original PR: #541

What are the difference between flushing for each frames v.s. flushing at the very end? I guess technically they are the same but maybe flushing on every frame would incur some more performance costs?

if (resampler) {
for (const frame of resampler.flush()) {
controller.enqueue(frame);
}
}
},
});
Expand Down