< Back to IRCAM Forum

Pipo.peaks to mubu.additive

MuBu
#1

Dear developers,

Is it already possible to use the analyses values of pipo.peaks directly into mubu.additive?

If I understand correctly I will in the first place need phase information in order to do so and I don’t see how to extract that information in the pipo.peaks example nor an attribute that points me in the direction. Is there a way? Or would there be a way to generate the sound without phase information through mubu.additive? (like the way you can do it in gbr.addpartials)

Secondly I guess the index information that is present in the matrices in the mubu.additive help file is to keep the same (almost) peak frequencies from bin to bin together. Is pipo also able to generate this index information? And again does mubu.additive also has a mode where you don’t need this?

I could also ask this for the residual data, but for me that is less important at the moment.

Best, Hans.

1 Like
#2

Hi Hans, for some time now I wanted to look at mubu.additive synthesis and make an example. Thanks for the motivation to do so now!
Some first answers:

  • you don’t need the phase (or index) for synthesis (gbr.peaks or gbr.harms did never output the phase), mubu.additive will complain but work nonetheless without Phase or Index columns.
  • pipo.peaks has a bug in the frequency (domain) calculation, you can work around this by setting the sr as domscale, but the hop size plays into it, too.
  • mubu.additive wants linear amplitudes, but the pipo.peaks examples output log amplitudes (not even dB)
    More later…
#3

Hi Diemo,

That would be great. I will experiment myself for a little bit using your answers as well.

#4

I will hijack this thread:
What pipo modules should I use for analysing the partials and residual of a source file, which I want to resynthesize?

#5

pipo.peaks for the partials, residual not implemented

#6

Here’s an example of re- and pure synth: additive-resynthesis-example.maxpat (142.3 KB)

#7

Yes! Great. I’ll try it out next week.

Best, Hans.