This is a trial from the augmented features test. The way I am tracking the head is using the fiducial marker for vuforia, as in the post before. But this time I try out a sample speech animation. The speech was found online – the choice of audio is intentional as it gives an opportunity to test how well the lipsync works. This trial is using SALSA for lipsync. The result is not optimal.
To hide the real mouth from showing, like in the last test, I tried to hide it using a blur filter. This method could work if the effect was tweaked and the edges of the plane were smoother.
Clay pepper 