Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Cohere Transcribe: Speech Recognition

cohere.com

Post reply on HN
Print view View original

61–70 of 77 posts

Re: Cohere Transcribe: Speech Recognition

#61

Post by roflcopter69 »

geooff_ wrote: post #33
bluegatty wrote: post #5

Earlier quoted context omitted.

can u comment on overall quality? their models tend to be a bit smaller and less performant overall.

My baseline was Jina, A Chinese model provider. I had major issues with their reliability. I have no comparison to provide in terms of offline metrics as I had to do an emergency migration because their inference service has extended downtimes. My experience with Cohere and interacting with their sales engineers has been boring, I say that is the most flattering way possible. Embeddings are a core service at this poi…

[dead]

Re: Cohere Transcribe: Speech Recognition

#62

Post by catlifeonmars »

bartman wrote: post #15
gruez wrote: post #10

> Limitations >Timestamps/Speaker diarization. The model does not feature either of these. What a shame. Is whisperx still the best choice if you want timestamps/diarization?

Even in the commercial space, there’s a lack of production grade ASR APIs that support diarization and word level timestamps. My experiences with Google’s Chirp have been horrendous, with it sometimes skipping sections of speech entirely, hallucinating speech where the audio contains noise, and unreliable word level timestamps. And this all is even with using their new audio prefiltering feature. AWS works slightly b…

I wonder if you could run multiple models and average out the timestamps, kind of like how atomic clocks are used together and not separately

Re: Cohere Transcribe: Speech Recognition

#63

Post by ChrisMarshallNY »

Bolwin wrote: post #50
mnbbrown wrote: post #46

Ran it over our internal dataset of ~250 recordings of people saying british postcodes (all kinds of accents, etc) - it's competitive for sure! Soniox (stt-async-v4): 176/248 (71.0%) ElevenLabs (scribe_v2): 170/248 (68.5%) AssemblyAI (universal-3-pro): 166/248 (66.9%) Deepgram (nova-3): 158/248 (63.7%) AssemblyAI (universal-2): 148/248 (59.7%) Cohere (transcribe-03-2026): 148/248 (59.7%) Speechmatics (enhanced): 134/…

Try two newlines between each one

That, or add 4 spaces before each line (renders as a 
).

Re: Cohere Transcribe: Speech Recognition

#64

Post by ChrisMarshallNY »

I remember Dragon Dictate. You had to spend ages, training it, and it still did a suckass job.

I recently was interviewed for a podcast, and she published it on Apple Podcasts. Apple does a transcript of the podcast. I assume it’s some kind of AI (not sure if it’s the same engine as Siri -which I’m not too thrilled with).

It made quite a few errors (not too bad -but errors, nonetheless), but the thing that annoyed me the most, is that it didn’t differentiate between speakers.

Re: Cohere Transcribe: Speech Recognition

#65

Post by mcbetz »

gruez wrote: post #10

> Limitations >Timestamps/Speaker diarization. The model does not feature either of these. What a shame. Is whisperx still the best choice if you want timestamps/diarization?

Mistral Voxtral has timestamps and diarization and does a good job for German (have not tested for other languages yet).

Re: Cohere Transcribe: Speech Recognition

#66

Post by walthamstow »

yorwba wrote: post #60
mnbbrown wrote: post #46

Ran it over our internal dataset of ~250 recordings of people saying british postcodes (all kinds of accents, etc) - it's competitive for sure! Soniox (stt-async-v4): 176/248 (71.0%) ElevenLabs (scribe_v2): 170/248 (68.5%) AssemblyAI (universal-3-pro): 166/248 (66.9%) Deepgram (nova-3): 158/248 (63.7%) AssemblyAI (universal-2): 148/248 (59.7%) Cohere (transcribe-03-2026): 148/248 (59.7%) Speechmatics (enhanced): 134/…

Is the human baseline 248/248?

Assuming all the accents are British, I doubt it. I probably couldn't get all 248 myself.

Re: Cohere Transcribe: Speech Recognition

#67

Post by aitchnyu »

ChrisMarshallNY wrote: post #64

I remember Dragon Dictate. You had to spend ages, training it, and it still did a suckass job. I recently was interviewed for a podcast, and she published it on Apple Podcasts. Apple does a transcript of the podcast. I assume it’s some kind of AI (not sure if it’s the same engine as Siri -which I’m not too thrilled with). It made quite a few errors (not too bad -but errors, nonetheless), but the thing that annoyed me…

You mean the ones designed to work on 64MB of RAM and CPU? I downloaded too many speech recognition and TTS shareware as a kid.

Re: Cohere Transcribe: Speech Recognition

#69

Post by Tsarp »

dinakernel wrote: post #4

My worry is that ASR will end up like OCR. If the multi modal large AI system is good enough (latency wise), the advantage of domain understanding eats the other technlogies alive. In OCR, even when the characters are poorly scanned, the deep domain understanding these large multi modal AIs have allows it to understand what the document actually meant - this is going to be order id because in the million invoices I h…

ASR has already proved its usefulness. Dictation tools are a prime example. Ever since whisper came out, usefulness for AST models running locally suddenly became a thing. Opened up soo many variants

https://superwhisper.com

https://carelesswhisper.app

https://macwhisper.com

Re: Cohere Transcribe: Speech Recognition

#70

Post by mnbbrown »

walthamstow wrote: post #66
yorwba wrote: post #60

Earlier quoted context omitted.

Is the human baseline 248/248?

Assuming all the accents are British, I doubt it. I probably couldn't get all 248 myself.

They are all transcribed by multiple blinded "accent natives". But yes, your point is valid - going to see if I can tease out the "single person accuracy".
Post reply on HN