Trill playback varies (excessively) with tempo

• Aug 17, 2021 - 03:40
Reported version
3.6
Type
Functional
Frequency
Once
Severity
S4 - Minor
Reproducibility
Always
Status
active
Regression
No
Workaround
No
Project

Create 3 measures at 3 different tempos, e.g. q=40, q=120 and q=180. Add trilled notes to all 3 measures (or even just one trill across 3 measures)
The trill starts of somewhat uncomfortably slowly in the q=40 bar, feels a bit fast in the q=120 bar and at q=180 doesn't sound human/musical at all (and note that tempos of 1/2 note = 120 or greater are not uncommon, and the playback is even faster in those cases).
While I wouldn't say a human performer wouldn't make any adjustment to the speed of a trill based on the tempo at all, generally I'd expect they'd play them about as fast as it is musically permitted by the instrument being performed on, and from what I can judge only at about q=80...130 do trills sound at all musical (using the piano sound anyway). NB grace notes don't seem to have this issue - unless they're before a note that is a trill, in which case they slow down significantly for slower tempos (!).

But I will say at least in seem cases the trill wasn't affected by a tempo change, I just haven't worked out when yet (it was happening consistently in one score, but on creating a new one I couldn't reproduce it).
Also as noted elsewhere, adding an accidental to a trill line is ignored by the playback - it appears to the only way to get playback with an accidental not in the key signature is to create a two-note tremolo (or write it out in full!).