Basically, in the near future, I'm going to try to unravel all of the words and phrases I can from the Annie Coleman mp3s on Librivox and create a sort of virtual actor with these... which I may eventually tie in to animations and/or make available here or on Turbosquid for game designers, or anyone else who has a use for such a thing.
After I do this with all the stuff from Annie, I may try to do some with some other open source stuff from others that post a lot of works on librivox or other public domain type places. I think just doing this with audio for now is what I should be doing since it's a little easier than doing it all with video where the images throw in a whole new level of complexity...
Quickly combined sounds for wav files, where I mix words, phrases or phonemes don't sound totally realistic or authentic... mainly due to timing, and the fact that Annie used various recording setups (maybe her voice is closer to mic, she was on another computer or something?, etc.)... but that's all more or less fixable to a degree, just takes some working over the tempo, timing, sound levels, and various other sound engineering principles. I sort of like this idea of creating an entire library of sounds based on librivox recordings like this since it does create a more or less "voice actor" for you free... since it's public domain... and since it's actually a real person that did the recording, most of it will sound better than old computer software things that try to make text in to speech automatically. At least the sentences and short phrases should sound better.
I used to play around with Voiceany demo and similar type text the speech software long ago, but it just sounded computery more or less. With Anne's voice, even though the recordings aren't perfect, and there needs to be some work done on timing and volume levels and things, there's much more potential for greatness to come out of the remixing. If I add all of that to my own unusual vocal recordings (see grababeat blog by looking at my profile) and maybe, eventually some midi recordings (now that I'm going to eventually start practicing midi keyboard playing more), I see a whole slew of opportunity here... maybe an entire remix based on public domain stuff culture will come in to being (it already does exist somewhat... but not nearly at the level that low level video or imagery swapping does with youtube, flickr and that sort of thing now)..
Once I get good with Annie's stuff on Librivox, I may move on to other folks voices over there that have recorded as much or more than her?!!... future is unlimited.
Happy Day to You.
Sunday, December 2, 2007
Subscribe to:
Posts (Atom)
