Raphaël Doursenaud

Thanks for explaining the context of your report. I'll leave it open for that specific purpose keeping in mind that the issue is most certainly upstream in the `rtmidi` code....

Hi and thanks for proposing a fix. Could you please share an offending file or at the very least a console output when loading the file with `debug=True`: ```py import...

Thanks for the suggestion. That's something I'm willing to explore for `mido` 2.0. I agree that the best implementation would be to have a `bleak` backend. Feel free to try...

> `'set_tempo'` meta-messages, pitch-bend messages, and pitch-bend range messages are inside `MidiTrack` objects. So it seems like each track has its own tempo changes, pitch-bend changes, and pitch-bend range changes,...

> > * 0 & 1: The `Set Tempo` affects the whole since there is only a single track (format 0) or multiple synchronous tracks (format 1). > > If...

_Edit: spelling and clarifications._ Regarding tempo, since you insist (I mean that in a good way), what is specified is: "[...]tempo information should always be stored in the first MTrk...

Agreed! > Regarding the playback of Standard MIDI Files on a computer: do pitch-bend events and pitch-bend range events in any track affect the pitch of the same channel in...

Just to be clear: for the choice deferred to application implementer I was talking about the relationship between tracks and outputs, not pitch. Furthermore, pitch bends were considered an effect...

I'll use [PEP 287 – reStructuredText Docstring Format](https://peps.python.org/pep-0287/) as this is the closest we have to a standard.

This is one level of abstraction above what mido is currently providing. Nothing prevents one from building a RPN/NRPN interpreter above mido though. Don't get your hopes high unless there...