omtree2bachtree bug or feature?

[Julien_Vincenot]

Hello,

One of my student came to ask me how to reproduce some examples from Ferneyhough based on the proportional tree notation from the old PatchWork / OM / PWGL days.

I directed him towards omtree2bachtree (with all the precautions about how bach behaves differently from those)
but still we couldn't wrap our head around some cases. Especially when setting up a (3 8) time signature, or bigger than (4 4), the splitting of the bar would not work as I expect, or produce results with values bigger than the bar.
I'm not 100% sure since I didn't use those much in a while, but in my memory in PWGL you could split any bar with a same pattern and obtain the same amount of beats, correct?

test.jpg (217.38 KiB) Viewed 53819 times

Also, makes me think of a feature request : would it be imaginable to have the "split" message accept a pattern of divisions instead of standard division by a single number 2, 3, 4 etc. ? For instance here we could use split 7 6 3 on the initial 3/8 value, then keep splitting resulting values.

Thanks !

Julien


<pre><code>
----------begin_max5_patcher----------
1968.3oc6b01aihCD9yo+JP74t8v175ocOo62QTUjCwIgs.NBbZZuU6+8arM
j.ASZRnM6d8fz.DaiGOOyiGyPw9G2MwdN+EVos0eZM0Zxjeb2jIpjjILo52S
rynuDmRKUEyNiUVRWwruWmmf8hPk9TKOKWqGg5IvZJxBagzed7w5htgJhWmj
uZVAKVnEI16g.2nnH3KA6DERht2B4F8fCbffkGvvdqGqpg7sYI4oLgpkfqRL
YgR974e+KnP6CkjuUTWTTUp5jDutgoEuss0ixb94c2I2c+fg.2JHno9eB0G4
UoptpCnnKWi8uIZbLOKikK5nwe8ae6DpGIx6gH4lePHniQnPLndNNRsjn1ic
5SWQFzUOy5py6pIjXEVwh8gyOI4MJ7ABr45365FR7cw96UuH+K2TR9c1TBcJ
ChP9ANDL1ELjHBPXCHWqoDeKMk5dhmzT5Gn5ApUnvqvyiyMw3ky1ARqipNmF
u9gMz3mrvus2Fr1aiVU6WIMX07tbcrJqR5yrEyf5.ZLynBQQx7sB83NS1qpS
relUTlvygpeNqPlYnCwQwIz.06G0nZrptCVcwiVowwqYzJ2e8TlhjCdBLovt
.UQt46D.NwIA9AP2cD4AeXKHvKB7D3FF3cukO5hYSnfd7A7qkybJ7ZWAcyY0
EyIT0Ey8hAEx+05hABQvyOAj3pfBbXT8fDW3PE9Gp6BZFSvJlwxoySYM4JmI
ZMmlu555E8lpYkkmf75143JzZbvuGpsj0WFyK16CMMI+oEulSyRhKE7xTtnY
CJlmWJJnfWk4LZVYRdILpDqU25MI.1wVj.EKeUcKLeaZ5AprYNZUlK44BP7p
qz9uKRnomgUwSeOmPWwN2OJBEntKNB10KBQPfW8C8oDzkKAMtNBIu84rfsjt
MUjygtX.FTZTOR47MUYL0Fjj5OncXCCxzMo8WVF8Ilnfwf6TCToDvluKQrts
yQY9qo4KfF2ploKQmk.jeDK4.WKngcjlCsePBv0zwRJyO80hVV4fFNEi4o7B
sxov1C6PMn4MYz8b2CQ0FcdBvT.5RkFUUuX+FUmpQWHRh2lpZ1FXfR.nL4eT
..RNlrN446pQyjzDwqGSpfljV9G0OROhP8lcylpjCVa1+ZNOm8WGrfccSe+a
7c+ktaMOkMS0ka1BpfNyY+VUcU0cTSfrk5eR9RdCFkNCUeD3HMaCnwaWvrX4
OyR4aTotmAZqZkvwkayikXZ6beh8p8QbVacGW3jYuT8Q5lX1xTNU36tOQ804
blkC4D53f7Ci7aIqx5F74JqFMS1lBVIDpCsqZsp.XlsR40xb5l1IsfmAdxd+
Uz.rOb+itgFD1.TWAKaCufllwWvzpurOxyGYsgdtr7EcLo6RVHVqZFaKL.Wq
YIqVKTrosBd67.m7qRUNAWxgtlrEJehULwCEKlu40cqYv.B.ari7owPeuRcy
lt.7qsqHQbnouuf3iY2R4ZYl59aCw1n8djXORraTPhQhcJGZBIkByDa3tr5l
4mTG1U2uk84yECiBcCbbOIURNr8kQiNSgG4E35gC+n4Znqhq4Z1IpTW5mr0S
1eRcidgzMmQVVGVlWudz50a1nmrQOYWDGyueOYmvK1nGrQOXmC6J3X1k94gX
lYodPwuIw5yFvdc2ra3w.6xjT1It0Ci49qDdCcvHh2Qtj+sg2F0wq3l97G1M
mao6PejCBgctEtC+eQPjHGiiGJ+eSax32s2wnKqdgVjYnEMBsC0cEx7C0S9J
hLBsCDZM+XkjuoQiP6.gVyOEEj6HzNXn07iNP95+MBsCDZMGwr7sHcDZGHzF
XFZCFg1ACsglg1vQncvPajYnMZDZGJzhMGMFdLZrg+Z.XNZL7XzXCGZMGMFd
LZrg6PvbzX3wnwFNzZNZL7XzXCGZMGMFdLZrgCsliFCOFM1vgVyQigGiFa3P
q4nwviQiMbn0bzX3wnwFLzRLGMFYLZrAAsMqqrLnLk8NQDpJP6qR1tZdrQ8s
lWrn0IM+O8pea.pOdHmLZwSrh1m0.VDzBQOFJcK6Ms68+pn+IS+Hex0O2O45
m2mX86X4cz6jUVxhj30z7bVZWedMm+X0Sbp13wZvaW2bZuyUSfNj8QsqT1yr
zpAST+mANjE5OB1e9Y95F4E3S7Bckm6z5Eeo8nJ3OrZFMVyev0b6i8Oc.i4a
yqmbquuKdMXqoSm5Z49nbEQ.97fb4PvpNI3vTH8oUKSBxSlV+sYZ0mq9VsjJ
n1rZKflW6WPGJd2J5gFsBUowO1XYZPN6b2iIjqZNBiplbv58p4ru2krNNHer
ymZ9X2ZJoR1OkTuAquC7L47DFKOWdh8.lB0XerdZRSvdjPeG2.IYNvQOgpCI
QN9AxIPs7wbdAyy8dVCHvWJzc+G3RgfpdsaLMvqMLJzRldaySIeaQbcKqRIs
bZL0wKEI452ivCkA2tPvPt51D1Hk3cUxxmjGpqjQCVxnyQxjtR14CWmQ8f1N
e35beRdXns2YnyUBdPxw8LjC4cPNjyPNduCxAe1cPFpjTuhM2BnCQtUBx4pD
j1aJcylJ2uJoHkALr120KoDg2q9YRt9mdpeVvdNot79pTnEvXUBvG+1B8PYu
3qWRkrkOOlh7sIUcr.sCDoZHS0p1vlpEHC0Hq28y69WfB99ON
-----------end_max5_patcher-----------
</code></pre>

[danieleghisi]

Hello, indeed this looks weird. It's a bad bug. Fixed for next release.
(that said, I usually achieve Ferneyhough-like rotations with bach.flat -> bach.rot -> bach.reshape)

[danieleghisi]

As for the split message, what you want is a different kind of splitting: split 3 splits in 3 parts (3 = number of parts); split 3 4 would split in TWO parts, with proportions 3:4. I guess it must be another message instead.

[danieleghisi]

Finally I've settled on
split @proportions 1 3 4
which feels quite natural. It'll be in next whatever!

[Julien_Vincenot]

Thanks Daniele that looks fantastic, I'm looking forward to it !
Indeed @proportions is a good choice I think, to make the distinction.

Also thank you for mentioning bach.reshape, I discovered it only recently but forgot its name.
I think it instantly made it in my top favorite of bach objects ever…!

[Julien_Vincenot]

I think half the problem is not what I expected.
Somehow I never realized working with bach.score that tuplets that are multiple of 2 cannot be recognized as tuplets,
unlike what was possible in the PW(OM)GL family.

What someone in a rather old post was calling a binary tuplet :
viewtopic.php?f=2&t=401&p=1435&hilit=bi ... lets#p1435

Is there any change on that front since 2016 (hidden option to force the old PW(OM)GL behavior perhaps), or is this distinction still deeply ingrained in bach.score?

[danieleghisi]

Ah, I see: you'd like to have 2-in-the-space-of-3 marked as a tuplet, right?
No this is still not possible right now. 2-in-the-space-of-3 quarters can be written as dotted quarters, and that's how bach writes them.
I see no easy way on the spot to mark them as |-----2------|, then. Sorry...

[Julien_Vincenot]

It's ok, good to know at least that's how it is !

I guess people who still rely heavily on the old style display could still export the tree notation in OM
before exporting to Finale or Dorico, if needed.