Hi,
I am having a coder build a home assistant, and we have been workign with the TTS-1 API. I was very surprised that it did not display emotions like the chatgpt APP does, and would really love to see a push in the TTS-1 to have SSML enabled , like we have on the app. As of now, it is so bland that it seems useless. We are looking for a new TTS until we can have what we hear in the App (which is great). I hope this little drop in the pound from a non coder pushes toward a better TTS-1 !
I feel like a better TTS would be more impactful than other things like Sora, that are compute intensive and is not “ready” yet for a greater impact.
AVM is nice but it is limited in which prompts can be said by speaking rather than typing, and the TTS available to listen the response is not bad, but definetly not near what AVM does.
A better TTS (keeping the costs reasonably) opens not only path for a huge improvement in ChatGPT itself, but also on app development as it would make the content more lively and rich.