I would welcome any way to do it for sure. I guess my method of splitting audio files and creating curves for those would work theoretically. It feels a bit hacky though.
Or if you wanted to create some nice tool to do that maybe you could have something like Papagayo (http://my.smithmicro.com/papagayo.html). You would open a dialogue sound file but instead of text you could drag in curve names (maybe only getting curves with a certain prefix or curves with some extra phoneme checkbox ticked), these curves would then give you some nice auto detected (from amplitude I guess) default start and end duration.
Then when you want to create the next curve (for the next phoneme) you would just drag the curve name to the start of the next phoneme, and when you’re dragging the curve name around you would preview the audio where the pointer is so you could easily tell where the curve should start. You could also have a right click option on the start/end markers to have split or shared start/end markers, for example you’ve got some words with silence in between, use split start/end markers. Or you’ve got a long word without silence, you could use shared start/end markers so the end of every marker is the start of the next one so you could have some nice crossfading.
A tool like this sounds awesome. Just too bad it doesn’t exist! :rolleyes: