you just reverse-engineered an assumption ![/quote]
I reverse-engineered like 98% of the user data. but also got support from Elektron devs.
most of the work was actually painstakingly mapping out sound parameters (A4Params).
before 1.1 there was a lot of arcane usage of the sysex dump menu, but this is not necessary anymore… take a look at the A4Request & A4SysexMessage classes.
there’s also various tools & helpers in the framework, such as a markov chain implementation…
even models for the envelope & LFO behaviour… the sequencer is completely modelled and can run synced to the AF, generating MIDI notes… a useful MIDI parser… etc., etc.[/quote]
what are you doing with markov chains? the other day i was reading about how they’re used to crack “good” passwords by starting with dictionaries of statistically relevant words and going through all the l33t spelling people try (if i understood it right) and they’re cool if you use them to try out a bunch of different musical elements as the “dictionaries”. seems like i see them used for just pitch a lot, which i dunno…
i’d like try to use your sysex to get more into speech synthesis on the a4, it’s definitely capable and it’d be so much easier to not do it by hand. i’m not even sure overbridge will work for this because a lot of plug in hosts don’t seem to like updating at audio rates…