2013-12-23, 06:50 PM
(This post was last modified: 2013-12-23, 06:58 PM by Holy_Diver.)
I would find a place with public domain recordings. See what the waveforms look like on those recordings, and see if they could be somehow synthesized. Or if that didn't work out, just modulate the public domain stuff. Help improve the public domain.
Also, don't record speech. If you must, use a made up language. The future is in hardware accelerated speech synthesis. Until there are good open libraries for that, text is where it's at. Games with speech are awful laughing stocks. Friends don't let friends record voice tracks for their video game projects...
Unless you can hire million dollar actors like Disney does (like the biggest budget games don't) your speech will be a blight on your game. Even primitive synthesized speech would be an improvement. It's easier to forgive a machine. So much easier. And besides, you want your game to be accessible to everyone, translating vocals is too much work, and too much data to store anywhere.
Also, don't record speech. If you must, use a made up language. The future is in hardware accelerated speech synthesis. Until there are good open libraries for that, text is where it's at. Games with speech are awful laughing stocks. Friends don't let friends record voice tracks for their video game projects...
Unless you can hire million dollar actors like Disney does (like the biggest budget games don't) your speech will be a blight on your game. Even primitive synthesized speech would be an improvement. It's easier to forgive a machine. So much easier. And besides, you want your game to be accessible to everyone, translating vocals is too much work, and too much data to store anywhere.