Spaces:
Runtime error
voice tone different
Hello, bro. I wanted to ask you something. When I use text-to-speech multiple times, I've noticed the voice tone isn't consistent. Sometimes it sounds one way, and other times it sounds completely different. Is there a way to keep the voice tone stable?
Yes , this is happening while you are using alloy as voice because this voice analyse prefered voice(male or female) for your line and switch to that automatically. But if you are using another voice this issue may not be coming.
Yes , this is happening while you are using alloy as voice because this voice analyse prefered voice(male or female) for your line and switch to that automatically. But if you are using another voice this issue may not be coming.
I've already tried other voices besides Alloy, but the voice tone still changes. Could this be a bug, and do you have a way to fix it?
I looked it up online and it turns out that OpenAI TTS doesn’t have a seed parameter or a 'voice style lock', so we can’t customize a stable voice across multiple requests. I’m working on an audiobook, which requires a lot of requests, so it seems this TTS model isn’t suitable for me. Do you have any solution?