< Back to IRCAM Forum

Partiels, batch exporting


Is there any possibility to batch export the result of many analysises choosing start and end time for all files, as possible in exporting the analysis result of one single file?
I have searched for that, but without success. It could be very useful.




not sure it answers your specific question, but did you watch that ?

Have a sweet summer.


Hi Greg,

Thank you for reply. Unfortunately it is not the exact process I am looking for: I want to export analysises of a group of files, but only relatively, to say, to time from 2’’ to 3’’ in each file.

I know it is possible to select exact time portion in exporting the analysis of a single file, but this option is missing in batch exporting.

To be more exact, I want to calculate the centroid mean value for a group of files, excluding the final silence, where the values are out of range. And I would prefer not to have to prevently edit each file cutting out final silence.

Sorry for long explanation, only for hoping to be clearer.

Thanks, have a nice time, ciao


Hello Fabio,

No, it is not currently possible to define the time range for exporting analyses during batch processing. As audio files do not necessarily have the same duration, there is the question of whether to select a valid time range or rather cut the beginning and/or the end, and use absolute or relative times to the file duration. For the time being, this use case has not yet arisen, and there are so many possibilities that it’s hard to know which way to go.

for the moment, I suggest cutting off the end of audio files as this should be relatively quick to do, and there are numerous tools on the Internet for this (or with Reaper ~20 seconds max). Or command line tools, like ffmpeg for example.


Hi Pierre (I suppose…),

Thank you for reply. I will provide as you suggest. My qyestion was relative to some SOL samples, that have a similar length, to extract mean centroid value using SDIF in OpenMusic, selecting only central duration zone.

Good Summer to you all, and thanks for replying even in there hot days…



Depending on what you want to do with OM, instead of removing the end of the audio before the analysis, you can perhaps ignore the results after a defined time directly in OM?
Or you can simply remove the end of the results after the analysis. I try to avoid SDIF most of the time but this is really easy to do with CSV (or JSON) and then use the CSV file in OM.
These are only workarounds but it might be better than cropping the audio files. Let me know if you need help with this option.


Hi Pierre,

Thank you for suggestion, indeed I have not realized I can work on sdif or cvs files directly inside OM, even if I use OM intensively enough…

I wil work around that, thanks again, ciao


Hi Pierre,

Always me…

Just if useful, here is a OM patch to calculate the centroid mean value of a SDIF file exported from Partiels. I have tested it on Linux Fedora 32. The used audio file is “Vn-art_harm-A5-mf-4c-N.wav” from FullSOL2020 database:

I have selected from 2 to 3 seconds. Next step will be to structure this process inside a omloop and saving list of file names together with their centroid mean value. (not today… :slight_smile: )

Thanks again for suggestion, ciao


1 Like