News for hackers. Stuff that matters.
Re:Show HN: Open-source, native audio turn detection model by remram on Thursday March 06, 2025 @10:14PM (#43287130)
Ok what's turn detection?
Re:Show HN: Open-source, native audio turn detection model by kwindla on Thursday March 06, 2025 @10:25PM (#43287164)
Turn detection is deciding when a person has finished talking and expects the other party in a conversation to respond. In this case, the other party in the conversation is an LLM!
Re:Show HN: Open-source, native audio turn detection model by remram on Thursday March 06, 2025 @10:31PM (#43287185)
Oh I see. Not like segmenting a conversation where people speak in turn. Thanks.
Re:Show HN: Open-source, native audio turn detection model by whiddershins on Friday March 07, 2025 @12:04PM (#43291887)
huh. how is analyzing conversations in the manner you described NOT the way to train such a model?
Re:Show HN: Open-source, native audio turn detection model by remram on Friday March 07, 2025 @01:13PM (#43292557)
Did you reply to the wrong comment? No one is taking about training here.
Re:Show HN: Open-source, native audio turn detection model by password4321 on Friday March 07, 2025 @04:28PM (#43294913)
Speaker diarization is also still a tough problem for free models.
Re:Show HN: Open-source, native audio turn detection model by ry167 on Thursday March 06, 2025 @10:25PM (#43287163)
Detecting when one user of a conversation has finished talking.

It’s a big deal for detecting human speech when interacting with LLM systems

Re:Show HN: Open-source, native audio turn detection model by woodson on Friday March 07, 2025 @01:33AM (#43287779)
It’s often called endpoint detection (in ASR).
Re:Show HN: Open-source, native audio turn detection model by lelag on Friday March 07, 2025 @11:25AM (#43291503)
Yes, weird that they didn't use that term for this project.
Re:Show HN: Open-source, native audio turn detection model by kwindla on Friday March 07, 2025 @12:48PM (#43292289)
I've talked about this a lot with friends.

Endpoint detection (and phrase endpointing, and end of utterance) are terms from the academic literature about this, and related, problems.

Very few people who are doing "AI Engineering" or even "Machine Learning" today know these terms. In the past, I argued that we should use the existing academic language rather than invent new terms.

But then OpenAI released the Realtime API and called this "turn detection" in their docs. And that was that. It no longer made sense to use any other verbiage.

Re:Show HN: Open-source, native audio turn detection model by lelag on Friday March 07, 2025 @01:33PM (#43292804)
Thank for the explanation. I guess it makes some sense, considering many people with no nlp background are using those models now…
Re:Show HN: Open-source, native audio turn detection model by mncharity on Friday March 07, 2025 @06:43PM (#43296042)
Re SEO, I note "utterance" only occurs once, in a perhaps-ephemeral "Things to do" description.

To help with "what is?" and SEO, perhaps something like "Turn detection (aka [...], end of utterance)"... ?