There's a lot of data loss and guessing with STT/TTS.
An STT model might misrecognize a word, but an audio LLM may understand the true word because of the broad context. A TTS model needs to guess the inflection and it can get it completely wrong, but an audio LLM could understand how to talk naturally and with what tone (e.g. use a higher tone if it's interjecting)
Speaking of interjection, an STT/TTS system will never interject because it relies on VAD and heuristics to guess when to start talking or when to stop, and generally the rule is to only talk after the user stopped talking. An audio LLM could learn how to conversate naturally, avoid taking up too much conversation time or even talk with a group of people.
An audio LLM could also produce music or sounds or tell you what the song is when you hum it. There's a lot of new possibility
I say "could learn" for most of this because it requires good training data, but from my understanding most of these are currently just trained with normal text datasets synthetically turned into voice with TTS, so they are effectively no better than a normal STT/TTS system; it's a good way to prove an architecture but it doesn't demonstrate the full capabilities
You need a lot more power. I found gpt4o struggles doing basic OCR of printed text by hallucinating alot, while tesseract engine (old skool) gets it perfect. You need the model to be powerful enough to do everything.
You can work around this by the way by sending the output through a checking stage.
Interesting, I've actually been using gpt4o extensively for OCR and didn't encounter any significant issues - could I ask you to please give an example of an image of (otherwise legible) text that it hallucinates on?
Personally, I'm very much looking forward to using a speech model like OpenAI's advanced voice mode to learn language. It can already do things like speak quickly or slowly which traditional TTS systems can't. Also, in theory a speech model could tell me if my pronunciation is accurate. It could correct me by repeating my incorrect pronunciation and then providing the correct pronunciation. I don't actually know how capable OpenAI's advanced voice mode is in this regard because I haven't seen anyone actually test this but I'm extremely curious to try it myself. If other voice models can achieve this then it will be an incredible tool for language learning.
Lots has been written on this subject, check out OpenAI’s papers on -O for example.
Latency is a big one due to batching. You can’t really interrupt the agent, which makes actual conversation more clunky. And yes, multimodal has better understanding. (I haven’t seen analysis of perception of emotions, has anyone seen analysis of this capability for GPT-O?)
Essentially, there's data loss from audio -> text. Sometimes that loss is unimportant, but sometimes it meaningfully improves output quality.
However, there are some other potential fringe benefits here: improving the latency of replies, improving speaker diarization, and reacting to pauses better for conversations.
Yeah that's the point. Without punctuation, no one can tell what inflection my "really" above should have, but even if it'd been "Really?" or "Really!", there's still room for interpretation. With a bet on voice interfaces needing a Google moment (wherein, prior to Google, search was crap) to truely become successful (by interpreting and creating inflection, cadence, emotion, as you mentioned), creating such a model makes a lot of sense.
So it's not STT -> LLM -> TTS? If I scream Chewbacca noises as input, will the model recognize it as nonsense, or will it interpret it with some lousy STT as some random words?
It's not, but it probably won't recognize it as nonsense. According to the paper,
> we construct a dataset named InstructS2S-200K by rewriting existing text instruction data and performing speech synthesis
It has only been trained on questions spoken by TTS, it has never seen (heard) nonsense. Most likely it'll just hallucinate that you asked some question and it'll generate some answer instead of asking if you're good. There's just not many audio datasets with real voices, there's no audio version of StackOverflow to be scraped
I used to have fun with that. Set Google Translate to Chinese (Or some other language I don't speak, though tonal languages seemed to work better), make some vague noises into it, and get out coherent but crazy phrases in English.
The speed looks very nice. I just recently setup LMStudio + AnythingLLM to try out local voice chat and its still a little slower than I'd like but the PiperTTS voices are nicer than this.
Obviously it doesn’t sound human but that’s extremely impressive for an 8B model. Compared to the Moshi model also on the front page now, this model seems to be more coherent, but maybe less conversational?
Can this play sounds that can't be represented in text? Ie. "make the noise a chicken makes"
Bwaaaaaak bwakbwakbwak
I see you like dubstep
your pseudo... a person of taste for sure.
If it can create sounds associated with any nonphonetic word spellings, I can't see why it would struggle with onomatopoeia.
Almost certainly not. Sounds like an old school vocoder, made to produce human speech and nothing else.
Can it understand those sounds? And distinguish correct and incorrect pronunciation of words and accents?
As in, a cluck?
But can it both say the word cluck, and make a clicking sound?
asking the real questions
Very interesting question actually
I’m not clear on the virtues or potential of a model like this over a pure text model using STT/TTS to achieve similar results.
Is the idea that as these models grow in sophistication they can properly interpret (or produce) inflection, cadence, emotion that’s lost in TTS?
There's a lot of data loss and guessing with STT/TTS.
An STT model might misrecognize a word, but an audio LLM may understand the true word because of the broad context. A TTS model needs to guess the inflection and it can get it completely wrong, but an audio LLM could understand how to talk naturally and with what tone (e.g. use a higher tone if it's interjecting)
Speaking of interjection, an STT/TTS system will never interject because it relies on VAD and heuristics to guess when to start talking or when to stop, and generally the rule is to only talk after the user stopped talking. An audio LLM could learn how to conversate naturally, avoid taking up too much conversation time or even talk with a group of people.
An audio LLM could also produce music or sounds or tell you what the song is when you hum it. There's a lot of new possibility
I say "could learn" for most of this because it requires good training data, but from my understanding most of these are currently just trained with normal text datasets synthetically turned into voice with TTS, so they are effectively no better than a normal STT/TTS system; it's a good way to prove an architecture but it doesn't demonstrate the full capabilities
You need a lot more power. I found gpt4o struggles doing basic OCR of printed text by hallucinating alot, while tesseract engine (old skool) gets it perfect. You need the model to be powerful enough to do everything.
You can work around this by the way by sending the output through a checking stage.
So picture -> gpt4o -> out1, picture -> tesseract -> out2, out1,out2 -> llm.
Might work for sound too.
Interesting, I've actually been using gpt4o extensively for OCR and didn't encounter any significant issues - could I ask you to please give an example of an image of (otherwise legible) text that it hallucinates on?
Same, it's perfect at OCR. Generating an image with text in it however… nope!
Speech is inherently easier to represent as a sequence of tokens than a high-resolution image.
Best speech to text is already NN transformer based anyway, so in theory it's only better to use a combined model
Personally, I'm very much looking forward to using a speech model like OpenAI's advanced voice mode to learn language. It can already do things like speak quickly or slowly which traditional TTS systems can't. Also, in theory a speech model could tell me if my pronunciation is accurate. It could correct me by repeating my incorrect pronunciation and then providing the correct pronunciation. I don't actually know how capable OpenAI's advanced voice mode is in this regard because I haven't seen anyone actually test this but I'm extremely curious to try it myself. If other voice models can achieve this then it will be an incredible tool for language learning.
Traditional TTS can certainly be cranked up in speed. Low/no vision users often listen at 2-3x.
Lots has been written on this subject, check out OpenAI’s papers on -O for example.
Latency is a big one due to batching. You can’t really interrupt the agent, which makes actual conversation more clunky. And yes, multimodal has better understanding. (I haven’t seen analysis of perception of emotions, has anyone seen analysis of this capability for GPT-O?)
Essentially, there's data loss from audio -> text. Sometimes that loss is unimportant, but sometimes it meaningfully improves output quality.
However, there are some other potential fringe benefits here: improving the latency of replies, improving speaker diarization, and reacting to pauses better for conversations.
> I’m not clear on the virtues or potential of a model like this over a pure text model
you can't put pure text with keyboard on a robot. it will become a wheeled computer.
actually this is a cool thing as a companion / assistant.
Really
Yeah that's the point. Without punctuation, no one can tell what inflection my "really" above should have, but even if it'd been "Really?" or "Really!", there's still room for interpretation. With a bet on voice interfaces needing a Google moment (wherein, prior to Google, search was crap) to truely become successful (by interpreting and creating inflection, cadence, emotion, as you mentioned), creating such a model makes a lot of sense.
So it's not STT -> LLM -> TTS? If I scream Chewbacca noises as input, will the model recognize it as nonsense, or will it interpret it with some lousy STT as some random words?
It's not, but it probably won't recognize it as nonsense. According to the paper,
> we construct a dataset named InstructS2S-200K by rewriting existing text instruction data and performing speech synthesis
It has only been trained on questions spoken by TTS, it has never seen (heard) nonsense. Most likely it'll just hallucinate that you asked some question and it'll generate some answer instead of asking if you're good. There's just not many audio datasets with real voices, there's no audio version of StackOverflow to be scraped
What about every movie that's been made? Although it might need to stick to those more than 100 yrs old to avoid copyright law?
I used to have fun with that. Set Google Translate to Chinese (Or some other language I don't speak, though tonal languages seemed to work better), make some vague noises into it, and get out coherent but crazy phrases in English.
Does any of the model-runners support this? Ollama, LM Studio, llama.cpp?
The TTS voice in the demo clip sounds remarkably like Ellen McLain (Valve voice actor).
https://en.m.wikipedia.org/wiki/Ellen_McLain
Sounds like it's trained on LJ Speech dataset, which is one of the best datasets and very commonly used
The speed looks very nice. I just recently setup LMStudio + AnythingLLM to try out local voice chat and its still a little slower than I'd like but the PiperTTS voices are nicer than this.
Wish there was training or finetuning code, as finetuning voices seems like a key requirement for any commercial use.
I assume one can't finetune this further?
Any demos showcasing it’s performance?
There's one on Huggingface https://huggingface.co/ICTNLP/Llama-3.1-8B-Omni
Thank you.
Obviously it doesn’t sound human but that’s extremely impressive for an 8B model. Compared to the Moshi model also on the front page now, this model seems to be more coherent, but maybe less conversational?
There is a demo video on the page
Am I the only one who trusts a GitHub repo much less when it has one of those stupid star history graphs on the readme?
That’s silly. People are allowed to be proud of their work.
Thanks for the counterpoint, guess I'm just an old cynic.
Not bad for 3 days training, voice output quality needs some work, it will be interesting to see what effect more training will have.