Self-echo VAD false trigger
open2 independent reportsacross 2 sourceson 2 stackslast reported 2026-08-23
What this is
The agent's own synthesized speech bleeds into the audio input and voice-activity detection mistakes it for the caller starting to talk, triggering a false interruption.
Status open: reported and still open where it was reported.
Evidence
1967xbill2026-08-23livekitsileroGitHub issuesgithub.com/…/6956
...livekit-plugins-silero==1.6.10 * livekit-plugins-openai==1.6.10 * WebRTC browser voice session * Silero VAD used for turn detection/interruption ### Checkpoints tested Previous checkpoint: text 597d30b3ec076608d059477bb14cfeffdf951bf5cae370d38f65d33bbfe82004 Current bundled checkpoint: text 1a153a22f4509e292a94e67d6f9b85e8deb25b4988682b7e174c65279d8788e3 The...
First seen 2026-08-23 · our copy fetched 2026-08-24T14:42:21Z · SHA-256 of that copy begins 33fe3914c037b2f3… · profile
PranayKumarJain2026-02-10—Hacker Newsnews.ycomb…id=46957802
...duplex audio: if you don’t subtract your own TTS audio (or at least gate VAD while TTS is playing), you’ll get false user-starts. Even a crude ‘TTS playing → raise...
First seen 2026-02-10 · our copy fetched 2026-07-25T17:50:19Z · SHA-256 of that copy begins 7b3c299ade3943aa… · profile
Quotes run to 30 words around the reported failure and link to the post.Add your call to this entryTake a quote down within 24 hours.
What is measurable
| Signal | hotato trust --stereo ./call.wav maps to this mode as detector kind echo_correlated_activity. Printed as a crosstalk line carrying the coherence and the lag: the caller channel looks like a lag-shifted copy of the agent's own audio, so the barge-in may be leaked output rather than a caller. |
|---|---|
| Establishes | That the caller channel carries a lag-shifted copy of the agent's own audio, with the lag and the correlation printed beside the finding.What else it establishes, and what your own trace settles
What your trace settles
|
| Tier | Tier 1 · dual-channel audio: caller and agent separated at capture, so overlap and timing are facts of the recording, exact to the sample. hotato is MIT and runs locally, getting started. |
Status history
The record has carried one status since it opened: open, reported and still open where it was reported. A move lands here with the link, the quote and the date that show it.
Discussion
Reader discussion of this mode: what fixed it and what did not. The reports above are attributed and linked; a comment is a comment.
Open this discussion on GitHubThe record
Ids, digests, every link this entry stands on, and how to take a quote down
| Atlas id | VFA-2026-0011 |
|---|---|
| Mode id | pat_8534cd0ace8a24cfc4955e32 |
| Content digest | sha256:8534cd0ace8a24cfc4955e32a71e86a1a37c128e0ef75a7a970debfdb78c4e1f |
| First reported | 2026-02-10 |
| Last reported | 2026-08-23 |
| First seen by the atlas | 2026-02-10T10:32:39Z |
| Last re-checked | 2026-08-23T20:25:37Z |
| Reported on | livekitsilero |
| Sources | GitHub issues, Hacker News |
| Scope | 2 independent public reports across 2 sources. This entry states where the mode has been reported; where else it can occur is what the next report settles. |
| References |
|
| Machine surfaces | feed.jsonfeed.atomfeed.schema.jsonatlas.llms.txt |
| Removal | If you wrote something quoted on this page and want it gone, one email to atlas-remove@hotato.dev does it. The removal path takes the quote, the handle and the link off this page, out of the feeds and out of the llms.txt within 24 hours, and tombstones the id so the entry cannot quietly return at a new address. |