4 ms·
Ask HN: What is your recommended speech to text/audio transcription tool?
Currently, I use a GUI for Whisper AI (https://github.com/Const-me/Whisper) to upload MP3s of interviews to get text transcripts. However, I'm hoping to find another tool that would recognize and split out the text per speaker.
Does such a thing exist?
- tmaly 3y agoMicrosoft has a tool that accepts wav or mp3 and transcribes it. But I do not think it can distinguish between speakers. How well does Whisper work in terms of correctness for single speakers?
- elektor 3y agoUsing the large model, it works really well, even in low volume settings/speakers mumbling. Some of my transcripts are pharma related and Whisper stumbles on the drug names, but I’m pretty understanding of that.
- java_beyb 3y agowhat you're looking for is called diarization. almost all enterprise STTs do that, you can find individual libraries on GitHub too. fine-tuning whisper is a nightmare, I don't know what the interviews are for, but again most enterprise STTs offer customization. you can add medical terminology. ---Google, Amazon and Nuance have medical models but either expensive or not available for personal projects.
- elektor 3y agoThanks for that! Searching for diarization really helped me narrow down for what I was looking for.
- tikkun 3y agoFor an end user application, Otter.ai is the best I've seen - I wish there was a better faster one built on top of Whisper, but there isn't a good one that I've seen. If you're looking for an API - then check AssemblyAI, Google Cloud transcription, Deepgram. I have a list here: https://llm-utils.org/List+of+AI+APIs https://llm-utils.org/List+of+AI+APIs
- solardev 3y agoDescript.com was pretty good at it when I tried it, but it's pretty expensive: https://www.descript.com/transcription https://www.descript.com/transcription We ended up using Otter.ai, which if I remember correctly didn't have as good a speaker separation model, but it was good enough for the price: https://otter.ai/ https://otter.ai/ There's also the much more expensive, human-powered Rev: https://www.rev.com/ https://www.rev.com/