Quote:
Originally Posted by
Bydlo
➡️
You can, in a near future, even sing in live situation with different voice skins, just like a guitar pedal does or like a synthesizer. Od course latency wont be a problem in a near future.
That's an interesting thought.
TC Helicon already makes live voice processors with real-time Autotune built-in.
Of course, different AI voice skins would require a LOT more processing power. And as of now, it won't be able to do that in real-time, forget about low-latency.
But I think that is also because developers aren't particularly concerned about real-time or low latency.
Same way everybody in Silicon Valley just knew you need $20B worth of server farms to power LLM's, until DeepSeek showed you can do it for a fraction of the cost and now they are all scrambling to adjust.
Likewise, It's only a matter of time before someone builds a Vocal AI that can do all that in real-time.