This project aims to investigate links between images and audio by turning an image into an audio image. I do this by taking an image, processing it to find objects in the image. Then, taking each object and turning it into audio and recombining the resulting audio to create a small audio segment with characteristics which come from the image.
Implementation uses openCV, a open source computer vision toolkit as well as CDP, composers desktop project, to synthesise audio. I segment the image into parts which contain characteristics such as its position and colour. This information is then used to synthesize audio segments. Then the audio segments are put together into one file in various ways. This brings up several problems this project intends to solve.
How do we appropriately find segments in an image? How do we represent height in a two channel audio system? How do we represent colour in audio?
Name: Daniel Ward