< Back to IRCAM Forum

OM how to order huge cluster chords from SDIF analysis?

if you have a sequence of chords (i.e. a “chord-seq” object), you have to separate each of them into a “chord” object, and then use the arpUp option in the second menu down on the left (see the image). If you have to analyze only one chord, you don’t have to use “chord-seq”, but just “chord”.
Francesco Vitale



Could you tell me how to load the sdif file in OM? Did you transfer the data from wave to sdif by SPEAR? I did tried to transfer the data in SPEAR but fail to load sdif in OM. When the sdif file is loaded, is it just connect to “sdif->chord-seq”? Thanks!


How do I separate my chord-seq object into chord objects? I’m totally new to this and don’t know the next step. My patch so far is very simple: SDIF-file–>SDIF-chord-seq—>Score/Chord-Seq

Foresun, see my attached screenshot for the patch. You can see it, it’s very simple. Also the tree is specified above.

you can separate the chord-seq in chord objects connecting to the first output slot of chord-seq the function “nth” (to be applied for each chord in the chord-seq), or the functions “first”, “second”, “third”, etc… (again for each chord in the chord-seq)

I put the first output from “score/chord-seq” to “lisp/nth” and got this error:


To be more precise: I meant the first output slot besides the “self” one, which gives you again the chord-seq object (and that causes the error), i.e. the second one starting from the left. That “chord” output slot should give you the list of chords in the chord-seq.

Hi Francesco,
Here is where I am at, closer? Still getting an error. Thank you for your patience.


Wait, I think I got it!


Hi Vasallo,

You can eventually use sort-list


Hi Vasallo:

Sorry for digging my question more deeper. I use SPEAR to export the file as SDIF 1TRC format. When I loaded it in OM and double click the sdif file, it’s kind of fail. The patch “sdif file——sdif->chord-seq——chord-seq” looks easy, but I suppose something wrong here. Would you like to see the attachments? Thanks!



Is that a chord seq. score? Plug into the leftmost inlet.

Hi Vasallo:

Thanks for the respond. The leftmost plugin doesn’t work either. I guess the problem is about reading sdiffile. I can’t load the sdiffile transcribed from SPEAR.

Hi Foresun,

It could be that trouble with your patch is that the sdif->chord-seq function is in eval-once mode. select the function and type ‘b’.
Just a supposition.
THen the real problem is that the output to this function is connected the second input of the Chord-seq (lmidics) and it should be in the FIRST input of the chord-seq (self).


Hi Karim:

It turns out the problem is non-ASCII characters of my sdif file name. I solved the problem. Thank you!

Hi Foresun:

I´m new to open music. I´m trying to load a sdif file into open music but when I chose the SDIF file I get a :OM > I can not construct a SDIFFILE with these parameters

Seems that you fixed that problem. Thanks.

Hello Vasallo,
The patch I’ve created appears to be the same, however I am not getting the same result.
Do you see where I’ve made my mistake?

It is possible that even though the SDIF file looks the same, its contents be slightly different. If the partials/chords do not start at the exact same time, and they can be considered as distinct chords (of one or more notes each) in the SDIF->Chord-Seq conversion. Could you perhaps attach your patch and/or the SDIF file ?

Thank you for your response!
The wave file is a cymbal roll with yarn mallets (crescendo/decrescendo), so should I have isolated the peak of the crescendo when creating the SDIF?
Thank you,

ride cymbal3.omp (6.4 KB) Ride cymbal- 2019-11-02, 1.19 PM-cs.sdif (1.2 KB)

That was the issue. Thank you very much for making me aware of this!

If the file was created with AudioSculpt / pm2, the number of “chords” in the SDIF file depend on the markers that are in place and used for the “chord-sequence” analysis. If you want only one chord, I think there should just be no markers on your sound.