Evolvotron_complexity
http://freaknet.org/alpt/src/misc/evolvotron_complexity
This script is based on `evolvotron' a wonderful program which mutates
algorithmically generated images. You can get it here:
http://www.bottlenose.demon.co.uk/share/evolvotron/index.htm
This script mutates an initial random image in only one direction, that of
the minimum redundancy, or in another words that of the maximum complexity.
*** Usage
evolvotron_complexity [new [function.xml]]
evolvotron_complexity [frame_number]
1) Copy the script into an empty directory and go there
mkdir ev; cp evolvotron_complexity ev/ ; cd ev/
2) Launch the script for the first time
./evolvotron_complexity new
In this way it starts from a random function and runs forever (until you
kill it).
The evolution can also start from an existing function, f.e. a function
generated by evolvotron:
./evolvotron_complexity new function.xml
All the generate images and function will be saved in the current
directory with the good_NUMBER.xml and good_NUMBER.ppm file name.
3) If you've halted the script, you can continue from the previous mutation
by simply running:
./evolvotron_complexity.
If you want to continue from a specific frame (let's say good_17.xml),
specify it:
./evolvotron_complexity 17
4) Render the final high quality image
cat good_NUMBER.xml | evolvotron_render -s 1024 1024 rendered_image.ppm
Where good_NUMBER.xml is the function of the final image
5) Enjoy ^_^
Note: After some time you'll see that the output images will degenerate. If
you let the degenerated images evolve, they will be more and
more like a /dev/urandom image:
cat /dev/urandom | rawtoppm 1024 768 > TEST.ppm; display TEST.ppm
When you've reached this point, you won't go further very much, so
it's better to stop the script, choose the really last good frame,
increasing resolution and continuing from there. Ex:
- We are at good_37.xml, but it is already degenerated. Let's
continue from good_20.xml
- ./evolvotron_complexity 20
For more info on this problem, read the "Parameters" section below.
AlpT (@freaknet.org)
Links
http://freaknet.org/alpt/src/misc/evolvotron_complexity