MuseScore and MIDI

(Caveat: I am a MuseScore user like you and I am not entitled to speak for anybody in the MuseScore community, so the following are mostly my ideas and perspectives; however, they take some advantage of my frequentation of the community and of some familiarity with the code and its recent history.)

Thanks for your comments and for highlighting some of the areas in which further development would be useful.

There is a sizeable part of the community (including some/most of the core developers) which sees MuseScore primarily as a notation program, i.e. a program to create written scores to print, distribute and so on.

Playback is perceived as a rather secondary tool, useful for aural check of the input music and perhaps as a teaching support, but not much more than that. Priority is given to correcting, completing and improving the notation part. This is why you have seen relatively little discussion about MIDI support in these fora.

This does not mean that there will be no develoment on the MIDI part; in fact, I think the next major version (2.0) will include some aural support for some of the articulations; it mostly means that MuseScore will never become a MIDI sequencer and the notation part will (always?) have the priority.

Thanks,

M.

I agree that 'most notation software users are interested in the exact representation of music for printing'; and to that end, it largely determines the method that is chosen to transcribe notes into or from a performance. Many people use realtime input, to then use various software features to remove musical performance characteristics in order to get 'accurate' transcriptions of notes - the subtleties of 'musical' performance to which I refer generally include: not playing in time, not playing in tune, not playing note durations precisely and so on - all the qualities that humanize the blobs and sticks on a page and make them 'musical'. For this reason, step-time is considered by many to be an especially efficient method of transcribing notes.

To look at this another way, a conventional way of teaching staff notion is to sustain the myth that certain symbols such as crotchet/quaver (whole/half note, etc) determine the duration of a sound, when in most cases, they generally determine when the next note occurs. To illustrate, two different instruments (clave, cymbal) might play a succession of crochets at the same speed but the 'sound' bears no resemblance - the cymbal sound is actually stopped by the start of another note ... Between these two extremes are innumerable various that may be determined by the nature of: the instrument, the sound it makes, the player, the expressive articulations that may be given to any particular note or phrase, the style and interpretation of a piece of music, etc.. For this reason, when using music technology to capture, transcribe or realise a musical performance, it is wise to consider them as quite different - for the sake of quality of output, efficiency (time and effort) ... and maintaining one's sanity.

So, real-time software provides features (such as quantize) to ensure input becomes transcribed into something 'readable' even though what is played back from the score then, doesn't perhaps, sound as musical; score writers such as sibelius/finale work the other way round by allowing performance articulations to be ascribed to notations so that the realisations do not sound robotic/machine-like. Digital Performer can do both in that a score transcription from realtime input is an approximation - the score can be corrected to 'look right' without necessarily changing the performance / note data and vice versa.

In defence of MuseScore, what you are asking is (at this stage perhaps) unreasonable given the current development time, and the number of other features that have a greater priority, particularly as you say 'most notation software users are interested in the exact representation of music for printing'. For such capabilities you we are talking about require massive investment in: time (eg Motu released V1 around 1986; Coda not long after that); and money (you know what sibelius costs) and customers have grown to expect 'value for money' .... Indeed, if you were really concerned about articulation, perhaps you wouldn't be writing scores for MIDI but for real people playing real instruments! I also create much of my stuff using MIDI but your reference to 'faithful MIDI output' can not be applied to all MIDI equipment. This is in part due to the different interpretations and implementations of the MIDI spec by manufacturers and the various models they create; but also in the way that the MIDI data can be radically transformed from instrument to instrument according to the nature and quality of the sound chosen to play that MIDI note data. I could painstakingly focus on precise MIDI note data (or to use your term - note articulations) to no avail because of inappropriate selection of a sound on a MIDI device.

I agree with you that all things considered it would be nice to wish for technology to be increasingly expressive and hope that the developers will present use with increasingly clever tools to harness our musical creativity. However, we shouldnt overlook the fact that the fallibilities do not solely rest with the technology (or the programmers!). At best, musical notation is never more than an approximation - its one of its most important qualities ....