I love the voice quality, and have been talking with a bunch of other game devs about how this and other TTS solutions have been making remarkable strides recently (also visited your GDC booth this past week!). Some years ago now, I worked on an experiment that auto-generated a gaming-centric TV show on Steam[1], but one of the big hurdles was that TTS was pretty flat (Amazon Polly); we couldn't get as expressive as [2] for instance. A few years ago, you could get emotive performances from TTS, but you needed to put in a lot of post-processing work from an audio engineer (e.g., Sonantic's TTS[3]). Stuff like ElevenLabs/PlayHT etc. seem like they'd solve that part of the problem.
As an independent game dev, I think we'll use TTS for placeholder VO a bunch - the writers can try out a pile of different material, and we only have to have a VO actor record at the end. And the current $600/year subscription for your "Ultra Realistic Voices" is a steal when used for that part of production. But for smaller studios, the pricing structure can make it tough to evaluate a new tool properly. What I really want to do is to spend 6 months having someone play around with the tech, integrating it into our toolchains, testing it out on playtesters, and so forth (and the 5,000 word free version won't do that for us). That $600 to try it out really isn't unreasonable, but when I'm also testing alongside Polly, ElevenLabs, Altered.ai, Uberwhatsit, Murf, and whatever other subscriptions, it's easy enough to say, "okay, well, maybe we don't need to add one more."
I'm not sure what the solution is, but I think smaller studios, who will be the ones to experiment with/benefit from this tech most in the coming year will give it a pass because we're all penny-pinchers.
Thanks for the feedback. We certainly want to support gaming studios of all sizes and are working with few of them to understand the workflows. What we've seen is not everyone wants a high-fidelity clone (which cost more), most of their voiceovers can be done with zeroshot cloning (quick clones that don't cost much).
As an independent game dev, I think we'll use TTS for placeholder VO a bunch - the writers can try out a pile of different material, and we only have to have a VO actor record at the end. And the current $600/year subscription for your "Ultra Realistic Voices" is a steal when used for that part of production. But for smaller studios, the pricing structure can make it tough to evaluate a new tool properly. What I really want to do is to spend 6 months having someone play around with the tech, integrating it into our toolchains, testing it out on playtesters, and so forth (and the 5,000 word free version won't do that for us). That $600 to try it out really isn't unreasonable, but when I'm also testing alongside Polly, ElevenLabs, Altered.ai, Uberwhatsit, Murf, and whatever other subscriptions, it's easy enough to say, "okay, well, maybe we don't need to add one more."
I'm not sure what the solution is, but I think smaller studios, who will be the ones to experiment with/benefit from this tech most in the coming year will give it a pass because we're all penny-pinchers.
[1] https://store.steampowered.com/labs/ultracast
[2] https://cdn.cloudflare.steamstatic.com/store/labs/ultracast/...
[3] https://venturebeat.com/games/sonantic-uses-ai-to-infuse-emo...