using Scala tuning map in Kontakt microtuning

[Zombie Queen]

I was pondering on some reasonably effortless way of applying Scala tuning maps in Kontakt... Tinkering with a script, where I could just type-in Scala formatted map. However, I'm not quite sure what am I doing there... Possibly, if you have some experience in micro-tuning, maybe you'd look-up, if it actually works, or I got it somehow wrong.

7 tone scale example:


linky (preset file, goes to Kontakt/presets/scripts):
http://www.fairlyconfusing.net/docs/tuna_fisz.rar

It requires Kontakt 5.5. There's still some typing-in / copy-paste to do to enter data, but once it's in, it can be saved in nka format.

[jancivil]

I suppose this is primarily about 'how does it work' for your edification; then you want 'does it work' tested. I'll take a look/listen this eve. I just use 'Microtuning.nkp' where you only have to see one octave (and I'll save as preset). Yours is like the format used to create a .tun implementation (ie., choice of 'tonic'/1:1 basis) of .scl iirc. What happens to the other five tones as to your 12-tone keyboard? I use scala fairly frequently, in Vienna Instruments Pro which only works with 12 to the octave; so I define dummy keys placeholders so, say '390' is a major 3rd keyboard-wise.

[Zombie Queen]

What happens to the other five tones as to your 12-tone keyboard?

They are being used to map following tones. Scala does it this way, when re-tuning with pitch-bend method, no? The idea is to enter tune map for following tones (next zero would indicate the end of scale), and let the computer, compute pitch change for each key over the midi range - the stuff, or right side is just preview of what is being computed, pitch map for standard tuning, pitch map for new tuning and computed difference. Does it make sense?

[jancivil]

So your work here shows you working down from C60, 0. You have a deviation @ A# where you describe the interval 160 cents down from the expectation @ A#. This is the scale's 40 cents sharp 'Ab' (reflects the 140 cents Db^), 840 per Scala. Then you describe a replacement for A @ minus 198 cents which is of course the scale's 'G', (correct typo:) 702 per Scala. So no, I do not want G and A mapped as G# and A#.

[Zombie Queen]

So no, I do not want G and A mapped as G# and A#.

Yes, I see, it would be better for sub 12 tone scales to map to nearest key and leave some keys unused (maybe color them on virtual keyboard as reminder). But when there are's more then 12 tones, the relevance to standard keys is something to forget, no? What would you do with something like this...

Code: Select all

Cassandra temperament (Erv Wilson), 13-limit, g=497.866
  0:          1/1               0.000000 unison, perfect prime
  1:         25.601 cents      25.600830
  2:         63.732 cents      63.732160
  3:         89.333 cents      89.332990
  4:        114.934 cents     114.933810
  5:        140.535 cents     140.534640
  6:        178.666 cents     178.665980
  7:        204.267 cents     204.266800
  8:        229.868 cents     229.867630
  9:        267.999 cents     267.998970
 10:        293.600 cents     293.599790
 11:        319.201 cents     319.200620
 12:        344.801 cents     344.801440
 13:        382.933 cents     382.932780
 14:        408.534 cents     408.533610
 15:        434.134 cents     434.134430
 16:        472.266 cents     472.265770
 17:        497.867 cents     497.866600
 18:        523.467 cents     523.467420
 19:        549.068 cents     549.068250
 20:        587.200 cents     587.199590
 21:        612.800 cents     612.800410
 22:        638.401 cents     638.401240
 23:        676.533 cents     676.532580
 24:        702.133 cents     702.133400
 25:        727.734 cents     727.734230
 26:        765.866 cents     765.865570
 27:        791.466 cents     791.466390
 28:        817.067 cents     817.067220
 29:        842.668 cents     842.668040
 30:        880.799 cents     880.799380
 31:        906.400 cents     906.400210
 32:        932.001 cents     932.001030
 33:        970.132 cents     970.132370
 34:        995.733 cents     995.733200
 35:       1021.334 cents    1021.334020
 36:       1046.935 cents    1046.934850
 37:       1085.066 cents    1085.066190
 38:       1110.667 cents    1110.667010
 39:       1136.268 cents    1136.267840
 40:       1174.399 cents    1174.399170
 41:          2/1            1200.000000 octave

Then, there's the matter of amount of pitch change being applied, which can get huge in such case, I guess I'd have to rework this in another way, so new key would be then sent to pitch-wise nearest note and re-tuned accordingly.

[Zombie Queen]

I made an attempt at automapping scales, with less than 12 tones, to nearest relevant key.



Unused keys can be ignored (tuning unchanged), blocked, or re-mapped to next key in scale. It only works (?) for scales <12. Linky updated.

[jancivil]

So no, I do not want G and A mapped as G# and A#.

Yes, I see, it would be better for sub 12 tone scales to map to nearest key and leave some keys unused (maybe color them on virtual keyboard as reminder).

Well, it's like this: musically there are 2 clear character tones, via the deviation +40 cents. With C for 'tonic' we have the 140-cent 'Db' to C and 'Ab' to G, the real point of this intonation. I'm going w. 'Db' and 'Ab' for my conception. For me on keys, placing these as 'D' and 'A' wouldn't in
itself bother me a lot.

But when there are's more then 12 tones, the relevance to standard keys is something to forget, no?


What would you do with something like this...

Code: Select all

Cassandra temperament (Erv Wilson), 13-limit, g=497.866
  0:          1/1               0.000000 unison, perfect prime
 [...]
 37:       1085.066 cents    1085.066190
 38:       1110.667 cents    1110.667010
 39:       1136.268 cents    1136.267840
 40:       1174.399 cents    1174.399170
 41:          2/1            1200.000000 octave

Then, there's the matter of amount of pitch change being applied.

Well, 40 cents is 4/5 of a quarter tone and I would call >50 cents deviation the next letter name. This 40-to-the-octave intonation does not appear to reflect tonal concepts and I don't know what its inception is about, but if one needs letter names one needs strategize that reasoning (re: 'Dastgah Chahargah', musicians in the tradition have access to systems with signs of 'half-flat' etc.). At this density I wouldn't tend to talk much of note names; I have used tunings of more than 12-to-the-octave, the extreme case was "Overtone Series" and note names here are of no real use to me; that one is in Absynth and as such 'C' is '0' per se. (I suppose MIDI note 0 as fundamental but so much octave button +/- I don't think I cared.) So there is a fabric to this field and its first principle is pyramidal distribution of spacing notes, ever closer intervals as the thing ascends; all of my choices must be 'by ear'. (There may be materials consisting of more than 12-to-the-octave with some basis in tonal harmony but largely a tendency away from tonal meanings will be likely IMO, yeah.

[woggle]

Hi zombie queen - I am not clear on what you are trying to achieve. It looks like you want to automap various scales to the keyboard in Kontakt- is that it? Trying to find an optimal way to map divisions of 4,5,12 whatever to the octave and also non-octave, non repeating scales. ?

If so are you looking for playability or memorability or both?

I would think there are much better places than KVR for asking playability/memorability questions.

For me a less than 12 to the octave based scale is better mapped to the octave starting on C probably and finishing 'early' I don't care about having any names for the individual notes.

but there will plentyof people with practical experience on places like https://www.facebook.com/groups/xenharmonic2/
and hopefully this is not you but whoever does this is probably chock full of knowledge http://xen-arts.net/

[Zombie Queen]

It looks like you want to automap various scales to the keyboard in Kontakt- is that it?

Yes, I do automap. I was looking for a convenient way, to try out different scales in Kontakt, not that I am particularly familiar with the subject, but driven by usual curiosity, what would it sound like? Thought I would try to script it myself. Since, as I said, I'm not particularly familiar with the subject, I thought, it would be fruitful to post.

For me a less than 12 to the octave based scale is better mapped to the octave starting on C probably and finishing 'early' I don't care about having any names for the individual notes.

I thought so initially, but then it is really easier, when you see octaves on keyboard. Besides problem with smaller/bigger number of notes par octave in my initial method is that pitch change grows significantly as you move away from reference key. I need to work on this. Now there are two modes, "stupid mode" where octave ends "early/late" and is repeated on "offset" and "sub 12" where there are unused keys (only for less than 12 tone tunings). I will try to code a "smart mode", which would find the nearest existing pitch to re-tune, that is reduce pitch change for notes further away from reference note.

[jancivil]

I think it best to keep in mind the utility of this type of idea v. another type. A 40-note octave is not the same praxis as a maqam-theory scale or the parent scale of a Raga family. So, with more than 12 defined pitches there are two paradigms: 1. there is basically 12 with inflectional alternatives (eg., Danielou's '22-Sruti" conception of Hindustani where 10 of these are adjustments of the syntonic comma, ie., 5/4 & 81/64 are both 'E' to a C tonic) and 2. there are basically more than 12 independent & identifiable tones you may or may not require names for. It doesn't make sense to me to have a (5, 6 or) 7-note scale and no attention paid to note names. Or a statement 'best to base on C', particularly where C is not relevant to the use case.*

I have two vi's I use for intonation deviations fr. 12tET: Kontakt (I use the preset Microtuner which allows for choice of 'tonic' or 1:1); and VI Pro which sees .scl in its 'Scales' directory if formatted for 12-note octave (and intonations will base upon any of the usual 12). The factory library for VI Pro contains such as *"Hijaz on D" which ('on D') is not anything per the interface since the interface indicates to the user +/- deviations from 12tET founded in "C = 0" (and selection of another tonic changes nothing of its intervals).

BUT it indicates 'Hijaz on D', this is a thing. I mean that chances are the person wanting eg., 'Hijaz on D' wants to see a scale with 7 conseq. letter names from D.

[Zombie Queen]

1. there is basically 12 with inflectional alternatives

I observed that many of these systems are variations of twelve tone with further "subdivisions", like this...



I'm not sure there would be convenient way to map it to standard keyboard, perhaps using mod wheel to switch through alternatives... maybe not.

[woggle]

There is not going to be a perfect system (what would that even be?) But you can have a convention that will be easier to remember and more playable. I can't see any way to get an easy consistent mapping between spatial layout and frequency divisions across the huge number of possibilities. But you can get a linear ordering.
It's a bit like a more complex London Underground map problem and I think needs a similar solution to Harry Beck's
http://homepage.ntlworld.com/clivebills ... /tube.html

the <12 subdivisions are easy using the gap method but >12 to the octave is more difficult. You could apply the <12 solution and just span octaves and leave unused keys - I think that is easy to remember but might lack playability. Alternatively for a more compact representation you can identify visual cues on the keyboard and go with them. the black/white key groupings are pretty obvious here. The problem is you end up with different fingering per repetition of the scale which works against memorability.

I don't see the problems jancivil has with starting on C - or anywhere consistent - as it is trivial to transpose either physically or mentally and you're breaking the keyboard convention anyway. Someone who is using the scale regularly will learn the new system - the next key is the next note and that note is ... whatever - and someone who is using the scale casually probably won't under any mapping learn the 'whatever' bit. But they will learn the next key is the next note, and that is consistently the case for all scales (taking in to account the gaps)

[Zombie Queen]

>12 to the octave is more difficult. You could apply the <12 solution and just span octaves and leave unused keys

An option I did, is to map to white keys only. I find it quite easier, when you break apart from standard octave division.
I hit a brick-wall with the idea for better tuning (less pitch change), with some dense scales, mostly it will happen with >12, multiple tones get mapped to the same key (with different re-tuning), and key release will release all of them, for which I don't see a universal workaround (fixing it would require adding groups, that is rebuilding existing instrument...). Anyway, it's sufficient for exploratory use.

There is not going to be a perfect system (what would that even be?)

Get five of those...

Glue them to the table, one after another, color code progressing divisions... just like on this old drawing.

[jancivil]

I don't see the problems jancivil has with starting on C - or anywhere consistent - as it is trivial to transpose

Yeah, seeing as I don't have that problem or one even vaguely resembling it and was responding to *this:

a less than 12 to the octave based scale is better mapped to the octave starting on C probably and finishing 'early' I don't care about having any names for the individual notes.

I think 'read the posts first' will be fair advice. "Transposition is trivial" (all of this really; by the time you said this there was plenty of context to work with), being as I am calling the 'tonic' 1:1, is stunningly obvious and so utterly basic.

(*: I have no idea what "finishing early" is supposed to be. If the problem is that we must make a 12-note .scl file of a <12 scale behind the limitation of eg., Vienna Instruments Pro, well I would create placeholders such as duplicate tones or I would create a gamut of 12 (the model being that the less-than-12-member definition derives from a 12 (or more) member system. Hijaz there (per the VI Pro visual) is not a 12-note scale, but someone wrote 5 additional definitions.)

[woggle]

I don't see the problems jancivil has with starting on C - or anywhere consistent - as it is trivial to transpose

Yeah, seeing as I don't have that problem or one even vaguely resembling it and was responding to *this:

a less than 12 to the octave based scale is better mapped to the octave starting on C probably and finishing 'early' I don't care about having any names for the individual notes.

I think 'read the posts first' will be fair advice. "Transposition is trivial" (all of this really; by the time you said this there was plenty of context to work with), being as I am calling the 'tonic' 1:1, is stunningly obvious and so utterly basic.

(*: I have no idea what "finishing early" is supposed to be. If the problem is that we must make a 12-note .scl file of a <12 scale behind the limitation of eg., Vienna Instruments Pro, well I would create placeholders such as duplicate tones or I would create a gamut of 12 (the model being that the less-than-12-member definition derives from a 12 (or more) member system. Hijaz there (per the VI Pro visual) is not a 12-note scale, but someone wrote 5 additional definitions.)

Not really my problem.