Srgy,
We have made contact before at the old BabylonJS forum. I have been integrating this repo with an animation system of mine. I have done a lot of breaking of things into classes, using Typescript. The reason I have done so, in addition to understanding the code, is so that I could make a sub-class of Preset
, which I call VoiceFont
. I have also added a method to Player
, to queue zones of this preset type.
This class has an additional property in each of its zone's, phoneme
. The class also has additional methods for creating / testing / trimming of zones, as well as saving of the "instrument". None of which this repo has needed to do.
I was making the buffer of vowel zones, about 2 seconds & ahdsr = true
. The consonant zones, I was making "to size", which is very small, 0.04 to 0.08 seconds. This is about the size of AFTER_TIME
itself.
Which brings me to, what is AFTER_TIME
? If this were set to zero, would a series notes of real instruments just run together? If so, I could probably add an additional property to zones with the default value of 0.05, and explicitly set those in VoiceFont
to 0. I do not want any gaps for my "instruments".
This is almost starting to run, but the consonant "notes" seem really short to almost not sounding.
FYI, I have also integrated Carnegie Mellon University's ARPABET 10k word phonetic database. I can just type in what I want & syllables are generated. For a look at very old version of my animation system, see & click Talk
. Back then I was using sound files & trying to voice sync.
And yes, I plan to sing if possible, so should I be adding 0.05 gaps between my syllables to keep in sync with the real instruments?
Thanks,
Jeff