OpenAI just released new audio AI models today, designed to significantly upgrade / replace Whisper and introduce advanced text-to-speech capabilities. These models seem more natural, human-like voices and have improved transcription accuracy which has been a problem with Whisper.
I’m curious if LingQ has plans to integrate these models soon and get rid of Whisper. It would also be fantastic to have more realistic AI voices available in LingQ, making listening practice and content creation even more immersive and enjoyable.
It would be cool if LingQ could adapt these. Can any staff member at LingQ here spill the tea about any developments on LingQ’s end?
This would be awesome! I already use it myself and then import the audio, but it would be really great if it were built into LingQ in the TTS button. I also prompt it to speak more clearly, suitable for someone learning a language (French for me), and I think it works quite well!
I agree. I hope LingQ rolls this out to subscribers. I think people will have a better experience with it than using Whisper, which seems to have issues occasionally
Yes the transcription AND the new tts. https://www.openai.fm/ Playing with it right now for chinese and spanish and it sounds like a real person reading out the text. Genuinely incredible.
Yea, I was playing with it too. It’s extremely cool, you can make it with different voices, personalities and accents. I remember seeing a comment from @ericlaycock about how LingQ was looking at Google Gemini for languages like Korean. I wonder about their thoughts on this new model. A lot of cool new technology out there.