If you’re looking for a cool robot voice, there’s this great website i’m using.
Edit: it isn’t great, you can’t make recordings from it.
http://www.oddcast.com/home/demos/tts/tts_example.php?sitepal
Use the voice Dave (US) and the effect Reverb to get a full-on 1990’s metallic voice thing. I am having so much fun with it.
It’s really cool.
(But beware to use this in a game though. Their terms of service are pretty specific on how you can use it)
Ah!
But I’ve only used a recording made from the website, which I believe doesn’t appear to have any terms and conditions. I guess sicher is looking at the app integration section
Caustic (Free) has an amazing Vocoder tool where you can do this with a sample of your own voice (or any sample). Pretty amazing: http://www.singlecellsoftware.com/caustic
Watch this for 30 seconds or so for an idea:
https://youtu.be/VlDO0uZ_Y4Q?t=160
It won’t sound anything like you when you’re done and shouldn’t have any licensing issues.
The terms explicitly prevents using recordings:
…you may not: (i) use the Vocalware Product not in conformity with these Terms and our other written instructions and specifications; (ii) copy, download, store or export any Audios created - whether on to your servers, or by permanently integrating the Audio with your own media or otherwise; the Audio may only be accessed through the Vocalware API with each audio being delivered for playback from the Product servers each time an Audio is played …
Google just updated their nlp cloud suite with text to speech that sounds amazing. I have not found any use restrictions but there is a limit to how much you can run through their cloud servers to generate voice, and there is a cost associated per use.
If anyone finds any restrictions on how the generated sound files can be used, please post it here!
Fwiw. I talked to someone via customer service email for this tool and they said there were not restrictions to using the finalized sound files thereafter for free. This may have changed but when I spoke to them six months ago it was not a restriction so far as anyone knew. Their goal is more real time dynamic web apps vs. set in stone software. The trick would be creating a web service that made it easy to push up your game’s script and methodically pull down final sound files.
I had contemplated writing a web service to do just this. But it’s on back burner for now while other projects move forward.
The voice quality especially of the ML based ones is pretty good. Certainly serviceable for educational games that want to provide text-to-speech functionality to reach users who might have varying qualities of reading capacities.
There is also Amazon Polly which provides the same thing. I attended a presentation where the service was used for prototyping the voice acting in Hitman:
“They have used Amazon Polly to voice-over NPCs in the last Hitman game. While voice actors play an important role in creating an immersive experience in games, they can be a quite expensive resource and not an ideal cost in the earlier development stages. Amazon Polly with its million characters per month in the Free Tier, allows IO Interactive to quickly iterate with text to evaluate how the final result feels, saving thousands of dollars on voice actors, whom they bring in during the later stage of the game development.”
Yes I saw that and has the same result for re-using saved .wav files from the service. Unlike Google Amazon polly says it’s legal right in their FAQ. I should really build this service 