Can i use 11Labs as TTS with voice cloning?

[DavideLentini]

Hello, I have a question. Would it be possible to use 11Labs as TTS with the voice cloning option?
https://elevenlabs.io/docs/product-guides/voices/voice-cloning/instant-voice-cloning#creating-an-instant-voice-clone
Thank you for your help.

[kaloprojects]

Good idea .. but propably not with 11Labs (one reason: 11Labs TTS is far more extensive than STT)… but Hume AI TTS is on my todo List already, seems same (or even more?) powerful , flexible API/json, supporting Natural (and emotional!) Voices and cloning too. … and all for half the costs ;)

but might need some more time, first have to spend some investigstions in upcoming weeks/month, will let the folks know when I come closer to this idea ☺️

[palashis92]

Hello, I have a question. Would it be possible to use 11Labs as TTS with the voice cloning option? https://elevenlabs.io/docs/product-guides/voices/voice-cloning/instant-voice-cloning#creating-an-instant-voice-clone Thank you for your help.

Yeah, you can. I’ve added ElevenLabs TTS with voice cloning, but as Kalo mentioned, it’s really expensive. Also, ElevenLabs doesn’t provide any streaming link in the API response, so what I’m doing right now is downloading the audio file to ESP32’s LittleFS and then playing it from there. Because of that, the delay becomes around 5–6 seconds.

@kaloprojects do you know why the ESP takes 3–4 seconds just to download a 200–300 KB file? And is there any way to make it faster—like under one second?

update i reduced the delay around 70-80%.

[RWBRWB]

do you know why the ESP takes 3–4 seconds just to download a 200–300 KB file? And is there any way to make it faster—like under one second?

I'm doing alot of radio station streaming and the reason your seeing 2-3-4-5 seconds download times is because the max download speed maxes out around 100-110 KB's speed wise and a lot of times it ends up being lower. It's a bottle neck of how the code is setup, its technically capable of much faster speeds but you have to go full ESP IDF to attempt it.

You can start playing the audio file before it's finished by streaming to a buffer and then have the audio decoder / player start playing data form that buffer before it's fully downloaded.