just watched image metrics’ “Faceware” vid recently and decided to do a study of my own in blender.
here in this vid i tracked 5 mouth points and used those trackers to drive shapes, some of the shapes werent excactly phonemes themselves but were designed to create phonemes by combining them together.
phonemes tracked:
A,O,U,FV,BM,EC
the shapes are currently symetrical, will try a non symetrical test soon
thanks for the reply Atom, actually i figured that out afterwards.
at first what i did is i constrained bones to the trackers and used those bones to drive shapes, but it was unmanagable, had to guess numbers for the drivers, like how far a certain bone could go from its local location. The shapes werent excactly driving up to 100%, I wasnt sure if id get the same result every time id track or perhaps track someone else so i did quick solution.
and the solution was, yeah a special rig with rulers :yes:
what it does is it captures specific bone transformations (like loc in the x axis or scaling) and transfers it to a much more managable set of bone drivers, this way i could keep track of the min and max transforms of a certain bone and use it to push the drivers all the way up.
thanks, it was kinda hard for me to grasp as well, but as i go along, i tried solving small problems and eventually answers for bigger ones came as well. anyways its still under testing, will update BA with tests soon. :spin:
im sorry i recently set the vid to private, currently made a freelance portfolio that’s linked to it and didnt want clients to see me and my dusty old room anyways i made a password for it if you guys still ever wanna see it.
password: “retargetting” w/o the " "
just click on the video link again and a password should prompt you this time :yes:
wow, pretty badass, nice work. This rig/work could result in some real helpful ideas for other blender users in the future, not to mention some pretty solid animations!