< Back to IRCAM Forum

real time analysis

Dear All,

For my new piece, I was searching for real time partial tracking analysis library/objects. In sound box, there is this iana~ object, which is very good documented. However, there are couple of articles about real time partial tracking analysis, where the gabor library often mentioned. There is also ejies library for the similar processes. Which one would you recommend me?

What I plan to do, is to analyze the partials, noisiness, loudness (perceptual) of the sound, which comes from live-input and process them independently from each other. Interpolation to another sound is also which I will try to do. One instrumental noisy sound will be in this case interpolated to another noisy sound of a live instrument. Ftm has machine learning part as well, as I understood correct. It would be also interesting to develop an intelligent interpolator, which creates an algorithm by analyzing the behavior of the decisive partials and interpolate them not linear but according to partials’ characters of both sounds.

Ftm looks difficult to use for me if I compare with iana~ . I would very much appreciate your suggestions.

Thank you in advance for your responses.
Best,
Onur Dulger

Hi Onur,

There is no free lunch here: There is no magic object that does everything as you like so at some point you’ll end of putting things together in a Max patch. Here are some additional pointers on top of Iana~ that might be helpful on the analysis side:

Then you want to synthsize the data and control it… . You might want to look at MuBu which has imported a lot interesting features from Gabor and FTM into an integrated framework including analysis and resynthesis: http://forumnet.ircam.fr/product/mubu/
I’d start playing around with MuBu first and then forward some quesitons to their User Group.

Dear Arshia,

Thank you for your helpful suggestions. I will check the Mubu and Ircamdescriptor~. Actually I was searching for some documentation of gabor, similar to what iana~ has. However, I will continue with MuBu in this case.

Thank you again

Best,
Onur