Open NyuBlara opened 3 weeks ago
After finding such a case in Clair de Lune, it appears that there is actually no desync, although the simultaneous arpeggios sound a little different ; this means this issue is probably much less concerning than previously believed. Still, I'll leave it open.
After finding such a case in Clair de Lune, it appears that there is actually no desync [...]
This was actually due to a deeper bug in the implementation of arps, which, having been fixed, indeed saw Clair de Lune breaking. Although this specific file has been fixed, the algorithm used to do so is flaky at best, and I'm almost certain it won't hold water in other samples, so I'm keeping this open.
I have not yet encountered a test MusicXML file with arpeggios occurring simultaneously on multiple staves. However, when such is the case, the current system (mapping start times to offsets) may prove insufficient.