Are there more than 128 notes in MIDI 2.0?

[tristanbay]

Sorry if this is the wrong place to ask this, but if it is, could you please direct me to the right place?

So I was wondering about the number of note IDs that MIDI 2.0 can understand. I tried finding this info out for myself by going to the MIDI website and poking around a little, but I didn't find anything that answered my question.

For MIDI 1.0, I believe the notes that you could play went from ID 0 to ID 127, with ID 69 being note A5. What I hope with MIDI 2.0 is that they've expanded this to something like ID 4095 or even 65535 (so MIDI note IDs are represented with something like 12 bits or 16 bits, respectively, instead of 7).

Currently for most musicians, this may be an unimportant feature, but for people like me who want to make microtonal/xenharmonic music, this could come in very handy for playing scales with a high note density, such as 53 or 313 equally-spaced pitches per octave, and changing key/scale during a song within that system of tuning, like how there are key changes in regular pop songs. Of course, these note IDs will still be associated with western note names like C5 and F#3, but some audio plugins and hardware have the ability to map them to different tunings through things like Scala files.

There's also MIDI Tuning standard, which I don't know how I would use.

So are there more note IDs in MIDI 2.0? If not, what are some good workarounds to consider?

[nix808]

as far as i understand

midi 2 supports microtonal pitch shifts

so if u want to use a micro-tonal scale, it isn't necessary to use the pitch bend wheel

[anomandaris1]

This is not a problem even for midi 1.0 since you can for example detune stuff by let's say 1 cent and proclaim it as a new note.
The problem is that you are not an octopus, so playing in granular system is a very dubious task.
I can imagine keyboards with up to 31 notes per octave at some point becoming popular, but 53... really - no. In non-meantone systems like for example 41, 46 and 53 notes per octave the number of possible pentads (generalized added 9th chords) is already insane (when we consider the possible microtonal alterations), I cannot imagine someone being able to freely improvise such "jazz" music.

[nix808]

@anomandaris1

afaik we must use pitch bend wheel to detune by 1 cent.
--as u know pitch bend depth is implemented by the VSTi also

yes- any microtonal scale must still be divided by 12 on a keyboard

Jan may find this interesting, their music is not equal temperament if it wants to be

[BertKoor]

Looks like in midi2.0 there are 8 bits so 256 possible note values:

(from https://www.soundonsound.com/music-busi ... ng-midi-20 )

[EvilDragon]

But it says 7 bits semitones, so that's 128 notes. There's a reserved bit there, currently unused. Who knows if it ever will be.

[tristanbay]

But it says 7 bits semitones, so that's 128 notes. There's a reserved bit there, currently unused. Who knows if it ever will be.

What should I do about this, then? Develop an entire alternative to MIDI!? Or should I just figure out how to use MIDI Tuning Standard and switch between tuning systems via automations in my DAW? And don't a lot a DAWs (e.g. FL Studio which is what I use) not even natively support MTS yet?

[lfm]

Not sure why you pick key based instrument to that kind of playing.
That would be with pitch wheel combination then.

Or pick a synth with microtuning like Korg Prologue if that might offer something like that.

I would do slide work on a guitar, or steel guitar or a sitar and move frets where ever you want.
I remember seing Jeff Beck mimicing this asian style throat singing on guitar with tremolo and fingers.

If JB can do that on guitar, I bet it is possible with pitch wheel and keys too.

Fretless instruments like violin and fretless bass would also offer any scale.

I also came to think of an experiement I did translating a bass line to midi and used Waves Tune - that plugin inserted a lot of pitch wheel in between since a string slightly alter pitch as you attack string. So in this way you could use another instrument that does not have key limitations and later generate midi that would play any library you have but with the pitches you played.

Bunch of ideas....

[tapper mike]

You still have 16 channels to deal with in midi 1.0 so you can have up to 106 notes per octave.
https://www.youtube.com/watch?v=ZMRUm_CoW-It=352

Jump to the 5:50 minute mark. Both controllers are using midi 1.0

[BertKoor]

But it says 7 bits semitones, so that's 128 notes. There's a reserved bit there, currently unused. Who knows if it ever will be.

Duuh, I only counted the bits. That's only for the raw note number. The text below says 7 bits for the semitone (as per midi 1) and another 9 bits for the fraction, which is what the 2-byte Attribute Data contains when Attribute Type = 0x03. So no bit is reserved or unused.

Me thinks the semitones should map one on one to the 12ET semitones (note 60 being central C) and the (keyboard) controller should convert a pressed key to semitone+fraction combination.

So basically you are directly targeting a specific frequency using a logarithmical scale. In effect midi2 moves the microtonal problem from the receivers (having to prep the synth with a patch to let it understand your scale) to the creators of the notes. Immediately unambiguously show the intent, reproducable on any device. I think that's a step forward. For a single scale you get to pick 256 out of 65536 possible notes instead of 128, in a resolution of about 0.2 cents. Whether such a scale is fixed, I dunno. That probably depends on your controller:

For microtonal composers. you can carry detailed pitch information in every individual note and event. When you did microtonal work before, you’d have to set up individual scales within a synth - often using Scala tuning files. Now you can have a Note On event each time you hit middle C that actually sends out a different microtonal pitch - and that’s carried directly in the datastream, it’s not a configuration thing.

[Tj Shredder]

This is not a problem even for midi 1.0 since you can for example detune stuff by let's say 1 cent and proclaim it as a new note.
The problem is that you are not an octopus, so playing in granular system is a very dubious task.
I can imagine keyboards with up to 31 notes per octave at some point becoming popular, but 53... really - no. In non-meantone systems like for example 41, 46 and 53 notes per octave the number of possible pentads (generalized added 9th chords) is already insane (when we consider the possible microtonal alterations), I cannot imagine someone being able to freely improvise such "jazz" music.

Oh Aaron Hunt, the inventor of the Tonal Plexus can…
https://hpi.zentral.zone/tonalplexus
I was involved in the only ever written scored piece for that instrument. It was a predecessor for the Tonal Plexus and a nightmare to setup. It simply used 128 notes times 16 Midi channels to send the information. Nowadays I would simply implement MPE to deal with that, but Aaron would insist on his own mappings. We needed some custom made software to make it work, and finally I wrote a Max patch which made it way easier in the end.
Yes, that is all still Midi 1.0. Don’t get confused with the 7-bit part in Midi 2.0. As a note carries more attributes than that. So it would be much easier to assign those 2048 keys to independent pitches… One column of the Tonal Plexus would share the same 7-bit number, but would have different cent values…
But beware don’t think of getting a Tonal Plexus, its not a useful playable instrument. No velocity, cheap buttons which easily fall off and in this case handmade seemed to be a disadvantage…
I didn’t like it…

[BertKoor]

The answer is a clear yes: you get 256 notes. With your scale of 53 notes per octave, that is 4.8 octaves. Most acoustical instruments (if not all except for the piano) have less range.

The confusion arises from the presence of two note numbers in a single note-on message with attribute data x030 present. It's just the second 7-bit one combined with 9-bit fine pitch which is used by the synth, giving a range from 8Hz up to 12kHz in a resolution of 0.2 cents. You can leave the first 7-bit note number zero if note numbers are not a part of how your particular musical system works. With attribute data x030 present, the first note number serves merely as decoration.

[tristanbay]

The answer is a clear yes: you get 256 notes. With your scale of 53 notes per octave, that is 4.8 octaves. Most acoustical instruments (if not all except for the piano) have less range.

The confusion arises from the presence of two note numbers in a single note-on message with attribute data x030 present. It's just the second 7-bit one combined with 9-bit fine pitch which is used by the synth, giving a range from 8Hz up to 12kHz in a resolution of 0.2 cents. You can leave the first 7-bit note number zero if note numbers are not a part of how your particular musical system works. With attribute data x030 present, the first note number serves merely as decoration.

But didn't the chart you posted earlier have a reserved bit like EvilDragon mentioned? Wouldn't that make it so that there are only 128 notes, or 2.4 octaves for a 53edo scale? And what's attribute data x030?

[tristanbay]

You still have 16 channels to deal with in midi 1.0 so you can have up to 106 notes per octave.
https://www.youtube.com/watch?v=ZMRUm_CoW-It=352

Jump to the 5:50 minute mark. Both controllers are using midi 1.0

That looks and sounds pretty neat. But where does the number 106 come from? 128 notes is about 10.5 octaves in 12edo tuning. If you multiply 128 by 16, you get a total of 2048 notes. If you divide that by 10.5 octaves, you get about 195. Or similarly, if you multiply the 12 notes per octave by 16 channels, you get 192 notes per octave. So really, you should be getting 190-200 notes per octave with the same number of octaves that you were getting with only one 128-note MIDI channel in 12edo tuning. So I don't know where 106 comes from.

But I'd like to also discuss a couple problems with the controllers that she's using. The first problem is that they're using MIDI, which was originally designed (and still is intended) for one instrument/controller per channel. You obviously don't *have* to follow this, but because this solution of using multiple MIDI channels seems like kind of a dirty and/or non-seamless hack, it may be worrying that compatibility issues are going to arise, including doing things such as controlling VST instruments through a DAW.

The second issue is that these controllers don't have an open platform that they're running on. What I mean by this is that these devices are running closed-off, proprietary systems, which keeps them to some extent from being extensible and universal. If I were to make a controller like this, I'd want the firmware and software made for it to be free and open, and work on more platforms than just Mac and Windows. How cool would it be if it could connect not only to your Linux machine, but also to your Android or iPhone? And if it was running a free and open system, it would be easier for people to make their controllers do exactly as they please.

I would one day like to make a controller that overcomes this second issue and possibly even the first. I'm just not that experienced in programming, product design, software design, marketing, and whatever else I'd need to do to get this product out on the market.

[BertKoor]

\But didn't the chart you posted earlier have a reserved bit like EvilDragon mentioned?

I don't know what his source is. Afaik note numbers may go all the way up to 256. The big question though remains how your DAWs piano roll will interpret such data. Best case it shows a double-sized piano keyboard, worst case you cannot see the notes beyond 127.
Ah, I see! It's the "r" in the first bit marking that bit as "reserved". You can be a bad boy and use it nevertheless (with no guarantees ofcourse)

But the discussion was about Midi2, not how DAWs handle unconventional scales.

And what's attribute data x030?

Fine micro pitch ofcourse, the last 16bits that effectively make the note number obsolete.

Did I say attribute data? I meant the attribute type (with corresponding data)