midilib does not handle tempo changes when calculating beats_per_minute
. The tempo events are correctly handled when reading/writing/moving them around, it’s just the functions that answer questions about the current tempo that are wrong. See github.com/jimm/midilib/issues/8. The method beats_per_minute
and related methods like pulses_to_seconds
will have to take into account the possibility of more than one tempo event. They will probably have to take new arguments specifying where in the sequence the beats or pulses are being requested. For example we could have +beats_per_minute(at_seconds: 0.0, at_beat: 0.0)+ (where the two keyword args are mutually exclusive). Perhaps there should be separately named public methods like beats_per_minute_at_seconds
and beats_per_minute_at_beat
as well.
-
print_decimal_numbers
andprint_channel_numbers_from_one
should be associated with sequence, or perhaps track, not event. -
Method to translate event’s time_from_start to number of milliseconds from start.
-
Swing quantizing. (Implied by list email from Carl Youngblood <carl.youngblood@gmail.com>)
-
Implement key signature in SeqReader.
-
Ignore unknown chunks in MIDI files. See the MIDI file spec.
-
Format 2 files(?).
-
Write better docs.
-
Tests for Noah Thorp’s midilib bug fixes.