An AudioGlu Idea

Consider the 8 bit multichannel render potential of a textured surface raster scan. (Also a nice pop up visual too.) If the video surface is 4 times oversampled, and bitstreamed, then an 8 amplitude level makes for 8 waves added in fixed geometric amplitude. Harmonics? The 3 colour channels is 3 left, right and centre perhaps. Alpha volume blending, and rotations as doplar modulation. With some custom audio prospective jigging of the viewport transform matrix, the 4 vector spin of polygon shader supplies the fragment shader to perform tri-granular synthesis of the quite downsampled audio raster read.

A video triple frame buffer could make an occasional used audio frame up for video S&H use.