• Profile picture of Karim
  • Profile picture of Jean


Groupe Public active 5 hours et 39 minutes ago

User group for OpenMusic and computer-aided composition. Visit the Forum for discussions.

merger and tied notes (bug)?

Auteur 2 Utilisateurs souscrits |
Profile photo of ratox

I think we have a bug for merger object applied to voices, when the results include tied notes, they appear in the graphic editor but they are not present in the rhythm-tree accessed from the slot of the resulting voice (so is not further editable)…
i attach a patch….

Thanks for help

Juillet 12, 2018 à 15:10 #27106
Profile photo of Karim

Hi Ratox,

Ok took a look at this issue. In fact it is apparently not a bug. Merger has a very strange non-compositional conception of merging. It is Voice merging and NOT rhythm merging that it does, (hence the very very bizarre ties that i admit are very awkward if used with repeated pitch).

So i send you a patch that does a rhythmic merging workaround. (rests are not welcome for the moment).


Juillet 12, 2018 à 17:26 #27109
Profile photo of ratox

Dear Karim,
thanks for the rapid reply…
neverthless I have to say that, unless I am wrong, the problem is not the ties, wich are metrically correct. Also the result of merger object is more readable musically than your alternative that, even if genial in its simplicity, forces the result to conform to more complex rhythm inside the triplets.

What it seems a bug to me is that we have ties in the voice that are not output in the rhythm tree so there’s a discrepancy between the visualisation of the editor, which is metrically right, and the stored rhythm tree which is wrong. (how does the object build the metrical structure then is for me a mystery).

The best result from a musical point of view is to use merger object and then manually delete all tied notes (which corresponds to the whole original durations, and thus, overlapping) which is pretty tedious. I wonder if there’s an easier way….


Juillet 12, 2018 à 17:57 #27112
Profile photo of Karim

Sorry , but maybe i wasn’t that clear myself :

The output of the voice, (cf. attachment), has a discrepancy with the voice conceptual object itself. A voice object in principle is ONE voice and not two.
That’s where you have the display and the tree structure discrepancy…
It is not supposed to happen.

Now if you are happy with this, please feel free to enhance the code yourself like many others do, with a lot of unhappy results… (merger is one of them!)

If you need to this, start out with om-merger which is originating the discrepancy (in fact, it’s obliterating all continuous-chord object, and transforming them into chord object, where, you got the tree error…


  1. Screenshot_2018-07-12_22-27-45


  2. Screenshot_2018-07-12_22-27-45


Juillet 12, 2018 à 21:28 #27113

Vous devez être connecté pour répondre à ce sujet.

Log in now