I have discovered the work ‘Gesture Co-ordinates' (2006) by Max Kazemzadeh who takes audience gestures and then translates them into digital codes that provide audio visual feedback to the viewer. As the audience moves around the space, their gestures are turned into pixel patterns on a screen which is overlaid with their projected image. This relates to my project in that human gestures are being encoded into a new form. However, in this interactive situation, the viewer cannot help but modify their gestures to ‘produce’ certain effects with the animation. I want to expose unconscious gestures and let the audience perceive their own emotional response from them as with the music I compose. There is sound in this work, however it performs the function of audio wallpaper accessory to the visually encoded material.

Another interesting Kazemzadeh work is 'Dial-logs' where "sound and image are crossed through a file conversion process to explore relational patterns of digital identity.
Quote from http://www.maxkazemzadeh.com/: Mouse and Man (Mickey and Arnold) takes the image of Mickey Mouse and Arnold Swartzenaegger and converts them to individual sound files. And conversely, the digitally recorded words "Mickey" and "Arnold" are converted from two sound files to two separate images. Those images are projected over sillohuettes of the iconic figures where drawings are made using only those digitally converted sound file artifacts that overlap the drawn contour of the figure. Clear audible relationships can be made from the sound to its respective image.

Having not actually heard the sound file it is difficult to comment on this work but I think the idea of turning sound to picture and vice versa is very interesting.