Hi,
does anyone know if there exists a research project which tries to analyze music in digital form and then tries to rebuild it completely on a mathematical basis in that form that it ouputs the instructions for feeding a softsynth like in werkzeug?
I'm not talking about ogg/mp3 compression here...
You mean
Markov chains (which is based on note data, so, e.g. a midifile)?
Hmm i think that's not exactly what i meant but i'll have to read through it first to get a understanding...thanks for the link!
btw, I wouldn't know how to analyse a random piece o' music in order to rebuild it with some softsynth. While I don't know how what features that FR synth has (I don't expect anything out of the ordinary though), I'm sure the music is not converted from a normal piece of music to their format. It's just a tune using a synth. The synth is prolly somewhat advanced, with a major role for DSP.
Yes sure. I was thinking of how to get any audiofile converted, stored and interpreted in a way that you're independent of using a specific softsynth or sequencer. This for sure is a major research project but it would lead to much better results than ogg/mp3 if it's done well. I just named werkzeug in order trying to explain what i roughly want as an output.
Unattached to this: If i remember right kb also offers his softsynth lib on their site. I don't know if there is any visual interface for the synth or sequencer involved.