Grouping Ninelet Partials

• Oct 7, 2017 - 00:51

Marching drummer here- so I deal with some stupidly fast rhythms. I've been transcribing something that includes a sixteenth note ninelet (9 notes over 2 quarter note beats, in case there's confusion). The only partials of the ninelet that are played are the 1st, 4th, 5th, 6th, and 7th (1st attached picture). I want to notate the ninelet so that there is a dotted eighth, then four sixteenths, and then another dotted eighth. However, for some reason Musescore won't let me include the 9th partial of the ninelet in that second dotted eighth (2nd attached picture). Ideas?

Attachment Size
9let_a.PNG 3.1 KB
9let_b.PNG 2.67 KB


In reply to by QuadSquad1

Very interesting. I can't make the final 16th a dotted 1/8th in your measure either. I deleted only the 9-let in question (select the first note, shift click the last rest press delete) and reentered the 9-let (select first beat, press 6 and ctrl-9) and the notes fit in perfectly.

I have seen similar things to this before. It is a bug I know exists, but I'm not sure it is being worked on, perhaps one of the program contributors can comment on that. From testing other things on your score I believe I now know how to cause it to happen from scratch, but none of this paragraph helps you.

What happens is that when you make a tuplet (that does not divide into 480) you can enter the notes all you like, copy it, cut it, alter it and it works great. Save the score and reopen it (or perhaps let autosave wreak its havoc on your score) and you get difficulties like you are experiencing. MuseScore thinks the last note will go beyond the end of the 9-let, so it won't let you do that. What is a little odd is it WILL let you turn the last rest into a note of the same duration or even 2 notes of half the duration.

In reply to by QuadSquad1

Sorry, that's what I said to do in a round about way.

I did some searching on this bug, and it seems this is being worked on for 3.0. I don't test 3.0 so I can't testify to the bug being removed. When 3.0 starts to get stable I'll test this problem and give feed back via the issue tracker if there are problems with it.

In reply to by Jojo-Schmitz

480/9 = 53.33... This problem happens with 9-lets, 11-lets, 13-lets and so on.

To reproduce the OP's problem
*create any 4/4 measure
*press 6 to make a half rest
*enter 16th notes (or even rests)
*press ESC or N to exit note entry mode
*select the 7th note or rest and press 4 . to try to make a dotted 1/4 rest/note
*result, you get the 1/4 note/rest but it won't make a dotted note/rest because it will end up with a note "beyond the bound of the tuplet."

To make it work right
*create any 4/4 measure
*press 6 to make a half rest
*enter 6 16th notes (or even rests)
*on the 7th 16th note press 4 . for the duration and it will work
*on the 7th note enter 4 more 16th notes/rests so one is beyond the tuplet
*use the back arrow 4 times to put the cursor on the 7th note/rest
*press 4 . to change the duration to a dotted 1/4 note and it works.

To see the corruption after you do the steps to make it work right
*select the dotted 1/4 note you created and press 4 to change it to a 1/4 note
*result is a 1/4 note followed by a 128th, 64th and 32nd rest

As I said, this corruption is actually inside the tuplet rather than outside the tuplet as in #202271: Copy-paste sextuplets/octuplets and their removal leads to corruption

In reply to by Jojo-Schmitz

My suspicion is that 3/9 is less than .5 while 8/9 is greater than .5 and a rounding issue arrises.

Since I don't program, I'm not sure when a repeating decimal causes this problem. I'm not sure what special cases have be set up to prevent this from happening. What I do know, is that when MuseScore puts a Tick into the xml this happens and when there is no tick, this doesn't happen. I've looked at a few xml scores where it worked and where it didn't and I noticed this trend.

In reply to by mike320

I think your assessment is pretty good. If a duration can't be represented using an exact division of 480 (number of "ticks" in a quarter note), then there is a discrepancy that we attempt to deal with here and there by various "fudge factors" - like instead of checking if two durations are equal, checking if there within 5 ticks of being equal, etc. No doubt we miss some places where a fudge factor would be needed, and other places the fudge factor is probably either too big or too small, leading to errors.

For 3.0 I have heard Werner has worked on reducing reliance on such kludges (which does not rhytme with fudge although for years I assumed it did :-), but I have no idea what the current status is.

Do you still have an unanswered question? Please log in first to post your question.