# Query: Triplets Duration

• Jul 25, 2021 - 10:23

Triplets seem to have too much duration in PRE. In this example triplets are scored over eighth notes.

• If I set the duration to 2000 in PRE then a triplet almost fills the beat allocation instead of taking 2/3 of it.
• If I set the duration to 3000 then it overlaps the next next triplet.

What am I doing wrong here?

Romance - LR.mscz

I don't know why, but the required number for a quarter (= x3 triplet) seems to be 2334 (duration multiplier) or 167 (duration) on average. // In the lower left setting section
I guess the triplet calculation doesn't follow the rules in PRE.

Estimated results I found:
Duration:
one triplet: 0
two triplets: 83 (83.333...)
three triplets: 167 (166,666...)

Duration (multiplier):
one triplet: 1000
two triplets: 1667 (+666,666...)
three triplets: 2334 (+666.666...)

I think some mixed calculation was done based on another note value.

In reply to by Ziya Mete Demircan

I think that the 666.6667 factor is derived from the tuplet's ratio of number of normal notes to number of actual notes.

In this example there are 3 actual tuplet notes in the space of 2 base notes so the factor 1000 x 2/3 = 666.667 is applied for each 1000 stretch above the base note's 1000 duration. So for a x1.5 length triplet I would expect to see a duration of 1000 + (500/1000) x 666.667 = 1333.

If this idea is correct then 5 tuplets over 2 base notes should use a factor of 1000 x 2/5 = 400. I'll test this in Musescore and if it turns out to be correct then I'll code it into the plugout.

So, the calculation should be similar to the following.

```      (Ratio)
%.  Tup Base
1000 / 3 x 2 = 666.666... // 1000 + (666.667 x 2) = 2334
1000 / 5 x 4 = 800        // 1000 + (800 x 4) = 4200
1000 / 7 x 4 = 571.428... // 1000 + (571.428 x 6) = 4428
1000 / 9 x 8 = 888.888... // 1000 + (888.889 x 8) = 8111```

In reply to by Ziya Mete Demircan

Yes, that makes sense. It's a bit of an empirical solution since I don't understand why the usual PRE duration is not valid for tuplets but as long as it works then I'm happy - and it's codeable.

It's a bug in Musescore's display of tuplets in PRE so I wasn't doing anything wrong.