Yesterday I had this idea that is, apparently, not new. I have been composing music for about 3 years for a series of stories I have been developing for (now on my second) screen play. My interests in sound generation extend to realistic sound effects in support of the 3D animation and effects that I have been also developing. My wife and I are international nature ecotourists and we have been to all continents (excepting Antarctica) looking for (mostly) birds and other wildlife. I am also a wildlife photographer. It is amazing how (especially song) birds have mastered the ability to make amazing sounds. The idea that came to me is to record these songs and produce a sonogram (using an FFT algorithm) that is represented digitally as an X-Y graph where the X-axis is time, the Y-axis is pitch and the darkness value (luma) represents amplitude. Sonagrams of this type are commonly given in bird field guides and are useful in species identification. However, I have always wondered whether you could take one of these sonagrams and render it back into sound to get something close to the original sound that was captured at the front-end of the pipeline. In my Internet research I turned up MetaSynth. It seems you can paint a sonagram yourself using something like Adobe Illustrator and render it into sound. It even uses one of the chroma channels for panning while, I believe, the luma channel is used for amplitude. You could even use another chroma channel for one of the other surround-sound channels or just RGB for three of the surround sound channels. It might even make sense to use the alpha channel for something. Anyway, MetaSynth only runs on the Mac . Does anyone know of other similar packages (that run on Windows-XP ) that can analyze sound in this way and turn around and produce sound using a modified or even composed sonagram using this graphical technique?
What a powerful technique!