AlgoComp

Audio snapshots from DarwinTunes.org at various intervals during evolution.  As of 2012 we are at 3500 generations and still going strong... you can take part in this interactive experiment here: http://game.darwintunes.org

DarwinTunes

DarwinTunes medley at 0 generations

DarwinTunes medley at 400 generations

DarwinTunes medley at 250 generations

DarwinTunes medley at 900 generations

DarwinTunes medley at 150 generations

Dr Bob MacCallum from Imperial College London describes the DarwinTunes experiment and its results delivered in classic Open University style.

DarwinTunes - evolution of music commentary

Here we have a selection of the latest loops from http://darwintunes.org/evolve-music

DarwinTunes medley at 3520 generations

DarwinTunes medley at 3630 generations

The latest snapshot from http://darwintunes.org/evolve-music

Please leave timed comments!

DarwinTunes medley at 5170 generations

For the first time we have made a medley completely automatically (previously, loops were picked by Bob for the medleys).  For each 100 generation period a consistently highly rated loop (across multiple raters) is chosen by a simple confidence interval-based statistic.  Most loops scored more than 4 (I like it) on average.  Reply to the timed comments if you would like more detailed statistics for particular loops.  Homologous recombination was enabled from around 3300 generations.

DarwinTunes - the first 7200 generations

The final chapter of the original DarwinTunes experiment.  Here are some randomly selected loops from the final population.

DarwinTunes-medley-at-8700-generations

Bob is a human who loves music and has been doing live mixes on a lockdown radio station he set up with friends. Most are on Mixcloud (https://www.mixcloud.com/Glastonauts_Live/) but he'll start putting them here too.

DarwinTunes (2009-2016) was an experiment to see what kind of music we get as a result of evolution by natural selection applied to a computer-generated four bar loop.  The "synth patches" are evolved also - using additive synthesis (just sine waves).  The only human input was the choice of tempo, time signature, effects rack and few other minor things.

DarwinTunes medley at 5170 generations

uncoolbob aka DarwinTunes