News for hackers. Stuff that matters.
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by wahnfrieden on Tuesday July 22, 2025 @04:24AM (#44644511)
whisper MUST be combined with silence detection / VAD
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by pferde on Tuesday July 22, 2025 @04:44AM (#44644614)
Ah, the good old "you're holding it wrong".

What good is a speech recognition tool that literally hears imaginary voices?

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by bmacho on Tuesday July 22, 2025 @06:46AM (#44645320)
> What good is a speech recognition tool that literally hears imaginary voices?

Well, if it is supposed to work after silence detection, then it is good for speech recognition I guess. It's like blaming a wheel why is it circular, you can't sit on it. It's a part of a larger machine.

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by dumbfounder on Tuesday July 22, 2025 @09:05AM (#44646385)
Just lay the wheel on its side and it makes a fine seat.
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by zettabomb on Tuesday July 22, 2025 @06:27AM (#44645221)
Considering that if you DO use VAD (voice activity detection), it's the best open weights voice recognition model by a very wide margin, it's quite good. I'd be willing to be that commercial products that "don't have this problem" are using VAD as well, and that this is well known to them. But Whisper is just the weights, and I suppose a simple reference implementation, not a full product.
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by xandrius on Tuesday July 22, 2025 @07:52AM (#44645740)
So if a tool has a process to have it perform at its best then it's a problem?

Do you also moan that before applying glue to a surface or it won't stick? Or if you need to drill a guiding hole before making a larger one in wood? Or that you need to use truly prime numbers for a security key to actually be safe?

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by wahnfrieden on Tuesday July 22, 2025 @07:28AM (#44645578)
Yes, you are holding it wrong. The good of it is that it does not output imaginary voices when used with VAD.

Show us a technology with better results that does not use VAD. If you can’t, then I’m not sure what you’re arguing against except superficialities so inconsequential that I can’t comprehend the condescension. The results speak for itself

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by Xmd5a on Tuesday July 22, 2025 @05:05AM (#44644752)
faster-whisper has a min_silence_duration_ms option
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by wahnfrieden on Tuesday July 22, 2025 @07:29AM (#44645587)
There are much higher quality VAD solutions available
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by DANmode on Wednesday July 23, 2025 @07:30PM (#44665173)
Please name a couple to get someone started who's hacking on webapps?

I'd really appreciate it.

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by DANmode on Wednesday July 23, 2025 @11:23PM (#44666555)
(as would future readers, I'm sure)
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by DANmode on Thursday July 24, 2025 @01:50AM (#44667371)
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by wahnfrieden on Thursday July 24, 2025 @04:15AM (#44668292)
I last used silero but haven’t kept up with stage of the art so didn’t mention it
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by nhecker on Tuesday July 22, 2025 @12:50PM (#44649819)
>imaginary voices

On the other hand, I can imagine that when things get quiet and the signal-to-noise ratio gets close to zero, random background audio (or randomness introduced in the transcription model) will be enough to tickle a critical number of neurons and elicit hallucinations.

The related thought exercise is this: Try scanning across the band with an AM or sideband radio, and after a while your brain will start to wonder "was that a voice I just heard, or music perhaps?" when in reality it was just environmental static.

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by DANmode on Tuesday July 22, 2025 @05:06AM (#44644753)
What's VAD?
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by maxbond on Tuesday July 22, 2025 @05:11AM (#44644780)
Voice Activity Detection (it predicts whether a short clip contains speech, eg to mute your microphone when you aren't speaking).
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by cmiles74 on Tuesday July 22, 2025 @08:31AM (#44646064)
If that's truly the case then they should make it part of the product, IMHO.
Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by wahnfrieden on Tuesday July 22, 2025 @12:23PM (#44649330)
How is it not the case? It is unusable without VAD or editing. I don't understand what you're questioning

I agree their products could be better "end to end" integrated. Meanwhile there is a continuously-improving field of work for detecting speech (which Whisper is incapable of). They offer official "cookbooks" with guidance on an approach they recommend: https://cookbook.openai.com/examples/whisper_processing_guid...

> At times, files with long silences at the beginning can cause Whisper to transcribe the audio incorrectly. We'll use Pydub to detect and trim the silence.

(Official OpenAI quote)

Re:Complete silence is always hallucinated as "ترجمة نانسي ق... by DANmode on Tuesday July 22, 2025 @06:00PM (#44653491)
What's a good starter VAD lib, and if you know, the best implementation of something like this to use in a browser-based app?

Say if I wanted to use it for Voice Nav, or Voice Input, but not piss off random people speaking the wrong language.