Add option to choose different voice backend depending on if a cutscene is active or not

The problem that I'd be hoping to solve is this. I'm perfectly happy with local Kokoro voices that I can run for free on my PC. The downside with Kokoro, at least on my PC, is that for each new line of dialogue it causes a split second of lag on my system. This is completely unnoticeable during cutscenes, but during gameplay, it can be jarring.

With a feature like this to conditionally use Kokoro voices for cutscenes, and something hosted like Amazon Polly for TTS during non-cutscenes, I could not sacrifice on game performance, while still saving money by generating my own TTS voices during cutscenes where I won't notice lag spikes.

It's just an idea I had that I wanted to write down somewhere where I wouldn't forget.