Adding articulations, dynamics, etc is probably the easy 1% of the job, compared with trying to quantize the rhythm - even if the recording actually has a regular rhythm rather than being what some people call “expressive…”
When I’ve done this (for other people) I find it’s usually quicker just to transcribe the audio by ear and write the notation, and only use the MIDI to check the notes as a “Plan B” option.
Professional arranger/conductor Tim Davies has a video on his debreved.com site called “Extreme Australian Orchestration”
Extreme Australian Orchestrating - YouTube
And before the end of the first minute of the video, he starts discussing cleaning up and importing MIDI files into a notation package.
That’s what LUA scripts will be for…and eventually Dorico may well include some user interfaces to build simple scripts on the fly (like the Logic Editors in CuBase).
In CuBase we can use simple logic editors to automate things. I.E. Select a range of notes and quanatize just those note-on events to a strict set of rules. Select some more and quanatize just the note off events to some set of rules. Etc… One can even build scripts that’ll do the selecting for you…
I.E. select anything within 30 ticks of beat A, with a duration no shorter than B, or longer than C in bars 20 - 40. Then you could look at the score and quickly deselect things with the mouse that you don’t want to alter, then run another bit of logic that does something to all those selected notes, etc.
More complex scripts could do things like use a groove template to predict where the ‘beat’ really lay in a selected range of bars, and quanatize it to straight beats for the sake of scoring. In Cubase I actually cheat sometimes with SMF files that have a flowing tempo, and just go through and mark the beats with some midi notes of an extreme range (I.E. note 120 for beat one, 121 for beat 2, 122 for beat 3, etc. Next I use that as a reference to lock onto all notes in a track/stave close to it and shift it to where I want it for scoring, then strip (or mute until I’m done with them) my beat marker note events. It’s a heck of a lot faster than trying to do everything manually with note by note editing, or transcribing the whole thing on paper first then entering the whole piece in one note at the time.
Point is, at some point we’ll be able to automate a whole bunch more stuff in Dorico than is obvious in this early version 1.x series of releases. Great things are coming…Unless of course the developers decide all this is ‘obsolete’, and ‘no one will use it’, and thus decide, ‘not to make it possible’. I think they’ll eventually get around to putting in support for an option that importing GM SMF files might attempt to set up the initial Layout for us and plug in GM instruments based on the PC events. It’s not a deal breaker for me, but I just get discouraged any time I see an individual get cocky and declare things as ‘obsolete since its inception’ without considering that their own sister company sells thousands of SMF generating devices on a daily basis…even in 2017, and millions of people use them at least on a weekly basis, for ‘professional work’ in class rooms and churches the world over.
It kind of goes back to my initial post on this thread about Linden Labs attitude that the ‘entire MIDI protocol’ is ‘obsolete’. They could be selling sims to music and dance studios, and content makers could be doing all kinds of creative things synced to their own music, using mere kilobytes of server-side storage space and band-width…but since a small group of loud mouths decided MIDI is ‘obsolete’ all you can get are 10 second chunks of wav files that can’t even be synced to anything. There are reasons people went to all the trouble to create various standards…no one person gets to decide they are suddenly ‘obsolete’.
I convert live performances into notation all the time here in Cubase for church musicians who record organ and piano performances directly to floppy or flash media on their X thousand dollar Yamaha console, and want them transcribed into scores they can study and manipulate. Of course it’s a nightmare if you go through and try to do it one note at a time…but scripts/plugins/logic editors/etc, help make fairly quick work of it. Presently, I’m most often expected to convert all my CuBase work into Finale or Sibelius, but I’m looking forward to a day when Digital Instrument dealers might push Dorico as part of the deal for that sweet 6k digital piano (or acoustic model with MIDI player mechanics) or MIDI organ console or digital organ that’s about to go into a class room or church.
Dorico is young yet…but I’m confident it’ll get all this stuff and more. The potential market out there for people who make SMF files on their digital instruments is huge. I’ve no problem with them having bigger priorities than play back right now. It is a very smart and competent team, as long as they don’t decide too many things are totally obsolete without thinking it through…the right kind of SMF importing and exporting could actually help tap and build a very valid, ethical, and under-valued market of ‘professional musicians’.
Different folk, different tools. I learned how to do this sort of stuff with two standard issue ears, a pencil and some MS paper before personal computers even existed. From your posts on music technology here, I suspect that you didn’t!
You would be wrong. I transcribe from ear to paper rather frequently. When I started my formal music education there was KCS Omega, and Hybrid Arts. I could not afford any Atari ST. My university music department did not have a single personal computer, and they would not even let us have email accounts on the campus main-frames (only engineering and business students could get that). By the time I graduated, Cu-base, Cakewalk, Notator, Bars and Pipes, and a few others existed. I was not able to afford a PC (A used Atari ST) until sometime in the mid 90s…after I’d graduated. I could not afford a multi-timbral ‘synth/sampler’ to hook up to it until even later, and it was more than 2k USED, and rather limited in polyphony at that.
I still do marching drill on a homemade light board because PyWare is an inefficient disaster for me. The thing is…even in my small town of only 3,000 people…there are more than a dozen churches, with keyboardists who practice, improvise, and record right on their console in SMF. They hand me piles of flash drives and floppy disks and get scores back in their favorite format(s) that they can analyze and do whatever they want with from there.
I can do things for them for a few bucks an hour and it’s very affordable to them. In contrast, if I tried to do this by ear to paper for them it would take me much longer…they would be rather put off by how much it’d cost them to get me to do a few minutes of score for them. Thanks to tools like CuBase and Dorico, I can deliver months of practices, performances, improvisation sessions, etc. fairly easily and quickly…it’s more affordable to them, so they keep sending me work. They get better, I get better, etc…
My apologies - I don’t mean to denigrate anyone still using GM files. My original point which got lost was in response to the OP’s assertion: “99,99% of midifile are in GM standard”. That may be true as a proportion of all the midi files on the Internet, but I don’t think it’s the case that 99% of Dorico users are using GM.
Anyhow, we realise that GM does have a role in particular workflows so we hope to support those aspects in time.
YESSS !!! LUA is very good choice to write plugin for Dorico.
I looks forward to the next release of Dorico which will include the access to all the objects of the score for writing good plugin such as the one that you propose : a logic editor cubase like. I wrote myself some plugin for Sibelius …
Rob Here it’s a topic about musical software, How can you write this ???
By what right can you suspect Brian (and I guess that’s also the case as for me …) not to be able to write some score without computer ?
It’s 
…
Thank you very match Paul !! With such a team, Dorico will be the best software of musical notation, I am sure !!!