My take on it is an actual dynamics analyzer and effector that first analyzes the tracks dynamics and sonic qualities and then writes those attributes to a file that is parralell to the audio. Like the stripe in traktor or abletons helper files.

The dynamics block contains user defined systems to write the attributes at various rates and levels of detail. So basically its re-writeing the audio waveform but in descriptions of how it deals with the audio . For instance a multi-band key analysis could be written to the dynamics file. For obvious reasons. How this information is analyzed is up to the user because the user sets up band definition with groups of eqs and crossovers to define the parts of the song. Gain is another quality that could be defined at this time. I think you can see how this pans out. The analyzation is set-up for the audio to be played more efficiently by virtul ‘equipment’ that uses the pre-ananlysis information to make decisions–because the analysis is defined by the user it is very customizable and so there for the behavior of the processing.

I have thought that once this other information starts developing midi is the obvious next description to take place. the audio and the dynamics analyzation can create midi events and also be output as different streams of midi that describe the user’s wishes for the processing of the audio.

This would necssitate the available of a midi system inside the program that deleivers events to whatever modules the user defines. In a simple example in traktor there might be a situation where a certain audio dynamic, key, or user defined midi event(cue) creates a string of dynamics adjustments in another module. Like the bass becomes stronger in one channel and the dynamics are adjusted in another channel. This seems like alot of set-up and complication just to switch the basses in a mix. But .. once this language developes in the user and interface I can see the system developed where the langauge is copied and pasted between songs and arrangements. A similar example of this might be the clip envelopes in Ableton and how you can manipulate two pieces of audio with the same set of directions. this is, however, different.

These would really re-invent the song on an auditory level and there would be no actual sample to tie source material to actual output. Is there a system that does this already?

sounds awesome to me! I am not aware of anything that does that but it would be great or triggering synths in ableton live and remixing tracks on the fly.

I agree … this possability does exist .. but a sample is still a sample even if its warped or not. Ive been toying with the ideas of a dynamics block and analyzer that outputs midi. It would work on some of the same principles as this. I also have been using the pyrolator midi device to create midi from an audio signal. In such a scenario the audio recording could be seperated as with this software and then analyzed by a device that outputs midi to a synthesizer set-up. These wouild really re-invent the song on an auditorey level and there would be no actual sample to tie source material to actual output. Is there a system that does this already?

GET ON THE CRACK

:)