It's been a while since my last blog post, but we've been working behind the scenes here! A lot of this album is going to be about getting the software right, and having a strong audio interpretation of the images to work from.
I thought that I'd hand this blog post over to Greg to explain a bit more about the software, and the preliminary stages of where it's at. Here he's talking about an audio conversion of Bridget Riley's 1962 painting "Blaze 1".
Over to you Greg:
"I've started looking at how I can generate sounds from the pictures that Madeleine has taken. I don't have any previous experience in image processing so I'm starting from scratch.
When I'm presented with a problem like this I often try to do the dumbest thing that I think will work, and then build from there.
The first image Madeleine has given me is quite monochromatic, and is essentially a spiral (sorry Ms Riley). Therefore I thought I could "traverse" the image in various directions, looking at the colors of pixels, and generate sounds based on those colors.
With this image I can traverse left-to-right, and generate a graph of the color value at each position:
Then I could say, when the value is large, play some white noise, and when the value is small, play nothing:
It's simple, but it's a thread to pull on."