MichaelBenjamin wrote:come on dudes...
OK!
I’ve not looked into this too deeply but it seems to me the term Procedural Generation needs to be defined. This Wikipedia link seems to sum it up well, according to what I already thought I understood:
https://en.wikipedia.org/wiki/Procedural_generationThis article says there has to be a random component introduced into an algorithm and a seed to get started on. Without the random component the same seed would always produce the same result. Now it occurs to me that if just the seed could produce widely different and maybe unpredictable results (in terms of what the user wants) that could be sufficient in some circumstances.
Randomness produces variation in outcome, like in genetic mutation. But, sticking with genetics, many outcomes are useless so there needs to be a selection process so only “useful” changes are allowed to prosper.
So, coming back to procedural generation for rhythms, as tulamide says a purely random assignment of percussion hits would be of little use. So I think we need a seed which defines the overall type of rhythm and an algorithm that can make use of randomness with a method of constraining the variation to what is musically in context.
Of course I don’t know the methods used in the Beat Map system but my guess is that the x-y system is actually varying quite a few parameters in the algorithm and randomness simultaneously, based on the data in the map. So if you had just 2 sliders or a 2-D joystick it would be insufficient to produce the variations available from the Beat Map. That’s why I personally think it’s a clever system, quite apart from the visual appeal.
So that’s
my ramblings over