Skip to main content

Synfire vs. Genetic Evolution

Posted

I am a trained chemist and do music just as hobby. I just had an insight I want share. Might feel a bit strange in this music dominated community. Is hopefully embraced, because to me its extremly encouraging about Synfire and its potential to explore new territories.

The insight is that Synfire's concepts (encoding building block by figures, late rendering into notes based on harmonic context and parameters) is basically extremly close to natural genetic principles which has created the enormous variation in living nature.

Take figures. They encode pitch relation but are not rendered MIDI yet until a seperate rendering step is done. Same in nature, where DNA takes the same role like figures. Biologic functions (= final music) are driven by proteins, encoded by DNA. DNA is expressed (= rendering) into proteins which happens by genetically hard rules (harmony) and also epigentic environmental factors (parameters). The enormous variety of nature is based on small building blocks on the level of DNA and its environment. Wonders of evolution (new music) happens by variation (happy accidents) of DNA and selection on basis of a fitness criteria (= sounds good).

I may continue on and on, it's really fascinating to see the similarities, but will stop here intentionally.

Good and since many, many years successful design principles, Andre !


Sun, 2022-04-24 - 13:44 Permalink

The challenge with genetic algorithms is always the fitness criterium ("sounds good"). Only a human can decide that. Or some sophisticated neural network, which however will only deem those examples "good" that are similar to what it has been trained on.

It looks like the only source of negative entropy in the universe is biological life (us). There is not much else that can be creative.