Software

Starting to use Processing, nice picture API…

The rest are custom images drawn one pixel at a time.

Images are generated from “.fa” files, which are divided into sections of 100,000 bp each, and breaking these into all possible 8bp sections. The pictures take a couple hours to generate with this setup. But to be honest, the pictures are basically the same, and can come out in seconds if I change the resolution to 2bp sections. I’m using 8bp because they look better, and its the most my computer can handle.

The algorithm is “counting”, the most complicated math is distributing values over the colors.

The code is either C# or Ruby, and most of the processing is through a DSL type language, which looks like this:

pattern prefix c:/path/to/patterns
grid 1000 c:/path/to/grid.data
1-1000 1-1000
grid close

Something like that creates the raw data, then a separate command line processes it into an image.

It actually gets a bit more complicated because of memory issues, depending on resolution. For the images I’m posting, I’m generating raw data and images at 200×200, then putting images into a single larger image.

The DSL approach allows some flexibility in data used to generate the image.