allow recording of interpretations

[josephlarralde]

Interpretations should be recorded automatically so that we can add an "export" button to the interface that would allow the users to download their last interpretation in MIDI file format.

[josephlarralde]

note : it seems that the json-midi-encoder lib allows to do that with a representation similar to midi-json-parser

[NyuBlara]

Though I understand the benefit of it being an "automatic" recording, how would we detect the "end" of an interpretation ? I.e., how would we deem an interpretation to be complete and ready for download (taking into account the fact that the user is free to loop any section of the piece) ?

[josephlarralde]

Indeed, "automatic" doesn't sound like the best idea to implement. Let's discuss this with Jean

[rblard]

Discussions with Jean led to the following model :

• Permanently record (starting from, either the first note in perform mode, or the first note overall), storing the note events with their dts (not the commands : looping points may change)
• After a fixed duration (that can be modified by the user in the menu evoked in #32), stop recording and display a log message in a "chat history" : "Interpretation of file [FILE] recorded at [TIME]"
• When clicking such a log, the interpretation is loaded back directly into the performer (through a MIDI inbetweener ? or through a new library method to load NoteEvents in directly ?) and the user can act on it (listen or perform it again)
• There is a download button for each log. It saves a MIDI file to the user's system.