While working on my microtonal equi-heptatonic anhemitonic balaphon patch, I ran into a problem for which I am hoping that someone has an elegant solution. Mine are wasting a lot of resources. Is there a way of isolating all the various pitch classes from the keyboard module? That is to say all the C's, all the C#'s, all the D's etc. Irrespective of octave. The note detector only detects specific note/octave combinations. But ... I want to be able to apply my tunings to all C's or pitches equivalently. I have found two methods. Both are brute force. Not particularly elegant, use way too many modules, and take up resources. Someone must have done this before. Is there a neat and simple method?

I've found a method. I'm using key quantizers. With the two top rows of knobs (on a NM key or rack) you can tune all C's C#'s,... etc. about a semitone up or down. I'm wondering if this is similar to one of your methods. The decoding part of my patch takes up about 43% of DSP power.

Not at home now, so can't send you the patch. However, I did not use the knobs preferring to use them for other purposes. One of my methods came in, if I remember, at about 40% resources, the other at about 45%. So ... roughly what you're getting. One used key quantizers, the other a whole bunch of note detector modules.

I did not finish deconstructing John Neumann's offering in order to see what I could build from his ideas, but I don't think it will come in at any less, really. This is a little annoying because although the balaphon is a very simple instrument to construct, the only way I can get close to the noise is to use sine and triangle oscillators in parallel -- which are expensive. I am loth to take any of them out just to get the tuning right. God help me when I get to the kora, that's all I can say.

I guess my way took too many modules?

Hi John.

As I said, I had completely forgotten about your patch and your method. Sorry.

Loaded it up and messed about with it. And, funnily enough, it uses pretty much the same method -- but totally inside the Nord -- that Tim Walters (in a private mail) offered to construct for me using Max. Since I am not on a Mac, Tim's kind offer was no help to me. In any case I want to do this totally inside the Nord so everyone can use it. Unfortunately ... your method is indeed also way too expensive. Using the germ of Tim's idea I came up with yet another method for isolating pitch classes -- but it requires 2 key quantizers, 2 control mixers, a gain control and an adjustable gain controller for each note one wishes to control!

Anyone know if it's yet too late to let Clavia know about this for the next update? If the Nord Modular is really to be culturally versatile, then it ought to be possible to use it to come up with alternate tunings a lot more easily than this. They just need to ship us a module that isolates pitch classes -- i.e. all C's etc. Trivial, surely.

I really don't see how I am going to create a realistic and tunable kora or balaphon for example without giving users the ability to microtune and remicrotune at will. Almost the whole point of a kora or donso ngoni, for example, is that a performer can quickly retune between songs and so come up with a sound more suited to what they want to play next. Really skilled players can even do this while still playing, and it's quite something to see/hear the whole character of a piece of music being changed in that way. Thus ... I would like to be able to give users of my patches the opportunity to do similar instant retunings simply by turning the NM's knobs. It's just another neat area of music-making that few Westerners ever really get into.

At present ... I can't see how to do this because all the microtonal tuning systems I am coming up with take so many resources that there really isn't enough room to build a decent instrument in the rest.

I haven't given up yet though. I am putting together all the various ideas that have come in and am trying to find a way to consolidate them into one stream-lined package .

I agree with you, Kofi, that being able to detect *any* "C" would be very useful. Go ahead and pass along to Clavia! :)

I was too lazy to have a look at the patches you considered sofar (for isolating pitch classes), but (instead) tried to find an independent aproach.

The example patch attached is worked out for 12 notes / octave. The number of notes / octave can be adjusted by setting the notes/oct module to some different value and adding or deleting to/from the keyboard splitter chain. The number of keyboard spliiters must be equal to the number you set into the the notes/oct module. The knobs in the coarse and fine sections can be positioned so that a usable tuning comes out - at least I hope so (didn't work it out, left that one as an exercise for you, laziness, again :-)

As you can see, the isolation part is not really the problem, it only takes three modules to do so. It's the recreation of a new pitch system that needs all the splitter modules. and mixers. To hear what I mean, try to connect the output of the quantizer module to the frequency control input of the oscillator (the synth) - it collects all the c's, c#'s etc. Maybe also try the output of the notes/oct module on that oscillator input, it jumps on octaves.

Anyway, it's just an idea that I thought you could use maybe. Seems to be a bit cheaper, dsp wise, than what I saw mentioned before. But there is a hidden cost, the zero page usage of the patch is 47 % (press CTRL+P to see it). You might want to change the setting of the rightmost knob in the coarse+fine module, it might be easier that way to achieve the tuning you want. Maybe try to use note detectors instead of splitter to get the dsp load a bit lower, but it would be a less general solution then, as you'ld get stuck into an octave based system then, I think.

Hi Jan,

Well ... you do seem to have accomplished it with your patch! I really am very grateful.

However ... as you say, there is a hidden cost ... I will just have to seehow this affects things in detail, although I have thrown in a few oscillators and such like with no problems so far.

Actually, you know ... I had also been having a great deal of trouble in building a djembe. I now think it was because I simply could not get the subtle combinations of the tuning and filtering right Seems to me that I might be able to use your isolation method to attack the djembe a bit more fruitfully.

Anyway, I have been playing around with your patch for about half-an-hour now trying this and that, and I hope to work on my balaphon patch later. I will incorporate your isolation method into it so that it will hopefully have the effect I want -- for people to be able to explore some nice and variable tunings.

If all goes well I hope to be able to submit it tomorrow, otherwise it will have to keep until the end of the week again.

Ehm, discovered a little problem, the note quantizer has a bug. It quantifies negative values (below middle E) different from positive ones. For the note classification it makes no difference, for some reason I do not quite understand. But octave detection went wrong and a fix was needed. While fixing I changed a few other things as well. The mixer modules from the previous patch might be adequate to tune the thing, but my ears are not & so I need to calculate instead of do things by listening alone. Linear controls are easier to be calculated in this case, so I decided to change the audio mixers for control mixers - cycle wise a little cheaper as well, but again there is a drawback - more zero page memory used :-(

Well it would be easy to put the audio mixers back in, but i think the difference to be not big enough for that - only about 3 % or so, and this works much easier, for me. Tuning can be accurate now to 1/8 th of half a note, guess that will do. See iso4.pch and tune it with the leftmost knobs on the control mixers, they are labeled from c# to b. C can't be tuned (to save resources), but the patch can be changed easily to allow for a tunable c as well.

As a bonus I'll add two sick patches, but I'm note quite sure why - they are sick. Maybe it's just me being a bit strange after listening through most of the 200 nm songs on mp3. Yes, I finally managed to get my proxy settings ok so that mp3 would work from my audio pc. What a day - some of you really surprised me, but I'll have to go back there to be able to tell who and what and how. Glad it finaly works though.

> Jan Punter said: Guess it's about time I made the effort to understand this side of the NM a little better. Anyway, it's just an idea that I thought you could use maybe. Seems to be a bit cheaper, dsp wise, than what I saw mentioned before. But there is a hidden cost, the zero page usage of the patch is 47 % (press CTRL+P to see it)

My approach to the Modular is more intuitive than anything else -- but this has its disadvantages. Since it now seems to matter a bit to what I'mtrying to do, guess I'd better understand this 'zero page' stuff a bit. So ... please could someone let me know if my understanding is correct. The machine code for the modules is held in the programme memory of the DSP -- which is divided into convenient little blocks called 'pages'. Correct?

The first page of memory is called 'the zero page' and holds indexes in the form of tables of values. The sequencer modules for example, have tables of 16 values (one for each position) which need an index of some kind on the zero page -- meaning that there is a maximum (is it 28?) to the number of -- for example -- sequencer modules there can be in a patch. Thus, and as with all other modules, the number of sequencer modules you can incorporate in a patch can on the one hand be set by the number of indexes available to them on the zero page. The DSP percentage on the other hand -- and we can see this in the patch data -- gives an indication of the number of calculations that the NM can do in one unit of time which is (I think) 1/96000 seconds. Is this correct? Therefore, this sets -- does it not -- a second and somewhat independent limit on the size and complexity of the patches we can construct.

That is to say, the size and complexity of a patch can be limited EITHER by the fact that we run out of indexes in the zero page memory (and even though the modular can strictly speaking perform all the calculations needed within its unit of time (whatever that may be if I have not given the right figure)); OR we can run out resources of a patch because the Nord Modular cannot complete the calculations demanded by the modules and their connections and so forth in time (and even though the zero page memory might not yet be full so that technically speaking a few more modules can be attached).

So ... even though to all intents and purposes constant modules do not take up any resources and involve no use of DSP time (because they are mathematical statements and not expressions that need to be evaluated) the reality is that it is still possible to put in a constant module that is the straw that breaks the camel's back' in that it fills the zero page even every microsecond of calculation time the Nord possesses is still available. At that maximum number (is it 112?) the NM returns a memory error simply because the zero page is full and not so much because it can't keep up.

So ... all modules use a percentage of the zero page memory; and some contribute more to the DSP load than others. Therefore, it is perfectly possible to reach the physical limits, through the zero page, of the DSP quite some time before one reaches the calculation time limits of those same DSP's.

In other words, Jan Punter's isolation patch although extremely clever in that it has saved me a considerable quantity of DSP calculation time still requires me to be somewhat careful in what I am doing in that I may still very easily reach the limits of what I can in regard to the indexes remaining available to me through the zero page memory.

I would be very grateful if someone would let me know if I am in the right kind of ball park here with what I think I know ... and if I am in error and someone could correct it for me then I would be greatly obliged. I have my own strange way of knowing things, but it also consists of asking stupid questions when I am confused about what are probably extremely simple things to others -- so please be patient with me.

> Friday's Child wrote: In other words, Jan Punter's isolation patch although extremely clever in that it has saved me a considerable quantity of DSP calculation time still requires me to be somewhat careful in > what I am doing in that I may still very easily reach the limits of what I can in regard to the indexes remaining available to me through the zero page memory.

Your conclusion is correct. Your derrivation along a (surprisingly) technical line of reasoning may be correct as well, but that's for others to decide. I think it's best to simply look at zero memory and dsp cycles (and Prog, X, Y and Dyn mem as well, in theory) as resources that one can run out of - leave the details to the ones who create the synth.

I made a list resource usage for every class for every module type, but while patching you could press CTRL+P every now and then to get a feel for resource usage other then dsp cycles.

> Jan Punter said: Tuning can be accurate now to 1/8 th of half a note, guess that will do. Well ... it won't QUITE do!! Although if this is all that can be done then I will have to accept it.

But ... first of all ... thank you very much for that amended patch of yours. Obviously, the email I sent in last night crossed over with the one that you were busy sending to me. As for whether it will do or not ... sorry for the long explanation, but I need to get across the idea of what I want.

The whole issue of tuning arises because of the distinction that I made earlier between 'frequency' and 'pitch'. Frequency is Hz or cycles per second, whereas pitch is a question of pitch perception involving human ears and brain.

As with all animals, human beings need to organize auditory sensations into a coherent whole. A part of this is pitch perception. Since this involves neural impulses, habit plays a big part. The 'pure harmonic series' is important up to a point, but over and above that, the 'agreements' that human cultures make about what notes are 'acceptable' plays just as important a part. 'Acceptable' is not always rational.

If you take a given note, its octave is universally accepted as being another 'acceptable' note because the two notes have a high degree of self-similarity. The 5th is also -- almost universally -- an agreed-upon acceptable note because it, too, seems to have a high degree of consonance and similarity. One way to think of it is: what chance would a despot have of saying to a group of human beings that they cannot play the fifth. It has such a high degree of consonance and is so easy to find that it would be kind of difficult to ban. Beyond the fifth, however, cultures have differed widely in what they are prepared to regard as 'acceptable' for making music. At the turn of the century classical composers regarded the major 7th as way too dissonant, whereas now you hear it everywhere in romantic rock and pop ballads.

Notes sound different from each other for many different reasons, all of which affect pitch perception. For example, when a sound originates to our right, it reaches the right ear before the left. And ... when it comes from behind it hits the physical structure of the ear first before reaching the ear canal. All of these cues are organized by the brain which then establishes the perceived pitch out of the fact that the pitches of all those original sounds are microscopically different on account of the different lengths of time they have taken to arrive. A given set pitch is kind of cobbled together out of those auditory cues, and the brain comes up with one pitch even though the stimuli originally had microscopically different frequencies.

The 'musical scale' that a culture eventually settles on to make music with is nothing more nor less than a pre-determined agreement regarding which of the "noises" available from the universe of sounds are to be used for making music. This requires a method for making choices. It also involves settling the issue of what human beings can actually hear, and what they think they can hear.

But ... once the selection has been established, then the musicians within that culture can create melodies (and harmonies) by providing their listeners with selections of pitches from that culturally determined selection of what are regarded as acceptable possibilities. Western music has, essentially, the diatonic scale -- which lives in the slightly larger set of possible choices called the chromatic scale.

The intervals in the diatonic scale are not equidistant from each other. This is a standard problem in all scales -- except that the Western chromatic scale did its best to open up a few more possibilities for music (in the shape of playing in keys -- which is a whole other can of worms) by establishing an equal distance between the notes in the chromatic scale. It certainly opens up many possibilities, but it also has its own problems an limitations. This is why other cultures have not always adopted it so readily.

There are 1200 cents in an octave, and 100 cents per semitone. Pretty much all intervals in Western music are, therefore, multiples of 100 cents. Which is frankly tedious. Other cultures have intervals of quite other numbers.

The ear of the average human adult can distinguish intervals as small as 14 or 16 cents. That is to say, once a system of music or a piece gets going, then the human brain is organized in such a way that the human ear tends to accept intervals 7 or 8 cents this way or that way as close enough to make all those be pretty much 'the same note'. This range is ... about 1/8th of a semitone!!

So ... using the Nord, Jan Punter has kindly come up with a method for tuning the Nord's oscillators so that 'notes' can be varied by ... 1/8th of a semitone. This is right on the limits of human pitch perception.

Unfortunately ... this is not quite good enough for me because -- if you look at the diatonic scale then even when you look at a tone, what this means is that to allow for complete flexibility and freedom, one rather wants to increase the tuning possibilities for a whole tone to ... double the amount, which is some 28 to 32 cents. Since the Western conception of C# is missing, as far as a non-Western musician is concerned any note in that kind of area is free to be assigned to either the C or the D. One day I may want to make a not close to the Western C# a part of my C, but another day I may want to make it count as a part of my D.

And then ... when one moves onto some of the scales that African musicians use, the scope for variability is greater than that -- as, for example, between the 'third' and the 'fifth' where you have got anything up to just short of a major third as an area in which a musician can play and tune an instrument and create a specific 'feeling' for a song or an instrument because there is a much larger amount of room for manoeuvre.

So ... having got so far, I would now like to go one step further. When a balaphon or kora player, or even an African singer, knows that they are going to play with others in a group, they all tend to have an agreement that once they have agreed upon a tonic note, whatever that might be, they will all also accept the third, but mess around with the rest. If one of the instruments cannot adapt its tuning (say like the balaphon) then that player is given the opportunity to play and let ring out the notes that they have on their instrument so everybody knows what they are. After that, the fun pretty much begins.

In the equiheptatonic anhemitonic scales I mentioned the other day, for example, the intervals are only approximately equal. Musicians can pretty much set up their instruments how they like. However, given the need to play with others the fourths and fifths are usually tuned very close to what one would expect given the harmonic series -- i.e. reasonably 'pure'. So -- when they play together in the familiar 'call and response' kind of way, there is a reasonable expectation of what is to come. Harmonies are created by DURATIONS. They are not created, so much, by musicians trying to play together in a key-centred kind of way. Westerners like to play in keys. Other cultures like to play in scales.

Thus -- when it gets the handover point, a musician or singer extends their note so that it then locks in with and plays along with -- harmonically, melodically, or both -- the first note played by the next person. Thus the variation in those handover points is usually kept quite restricted -- while it is on the other hand opened up a lot by for example the third, which people really vary quite a lot. Therefore, overlaps or thirds and fourths and thirds and fifths are the recognized handover structures. One musician can either give or else pick up the fourth or the fifth, which is a kind of accepted and little varying standard note. The other musician will either receive or donate the third that goes with that fifth or fourth, but varied according to how they want. This is done so that simultaneous FOURTHS never in fact occur. This is because nobody wants tones or semitones that clash. Thus if one musician hears the other one playing the fifth (or fourth) they will harmonize with some other note; and if they hear some other note then they know that the fifth is still open for them to play. But they do not play them at the same time. In practice, tunings systems are crude, and it is very hard for one kora or xylophone maker to know exactly which C another one has picked. Although in practice instruments do vary very slightly in their tunings of the fifth, everyone likes to pretend that they are actually the same. If they were played together then the dissonances would become too obvious that this balaphon is a little different from that balaphon. Those are not the notes that people are interested in changing. It is a bit like polyrhythms where everyone has an agreed upon beat ... and everyone agrees that everyone should start playing a given rhythmic cycle one beat before it 'actually' 'begins'. Thus everybody knows that two balaphons are unlikely to agree exactly on their pitches of octaves and fifths... but if we avoid playing them together for those notes then nobody really cares.

Thus .. African musicians make great use of the fallibilities of human pitch perception to give the illusion of constant and unvarying fourths, fifths and octaves by means of durations. One musician extends a note a bit while another uses it as a cue and a filler. As said above, this is one of the things taken care of by polyrhythmic textures. Everybody knows the main beat, the timeline, and the accented beats, and everyone has a pretty good idea what to play when.

So ... the 'Jan Punter' system I now have does give me a variable tuning system that still allows me to build a reasonable instrument without breaking the resource limits of the NM, but it is still respects the Western chromatic scale way too much for my tastes. I want to be able to make the intervals between notes grow and shrink, and for some of them to be tunable within the degree of accuracy that 1/8th of a semitone that I now have ... but I would like to be able to assign other notes a much greater tuning range.

So ... next issue ... how to sometimes 'double up'. For example, if I make a 7-note balaphon that just uses c, d, e, f, g, a and b, there is no c#. But I want to be able to give say 20% of the tuning possibilities of that missing C# to the C; while giving say 50% to the D which can then be tuned sometimes very flat, and sometimes quite sharp. Much the same goes for the D#. It can give say 20% of itself to the D, and the remaining 80% to the E. This E can then get a further amount of tuning possibilities from the F, which being the fourth is not going to change much. So ... let's say that 80% of the tuning possibilities of the F are also given to the E. It then has 100% belonging to itself; 80% given to it by the F; and a further 80% given to it by the missing D#. That way, the instruments and singers have a lot of possibilities opened up to them.

Sorry for the long explanations -- but what I am looking for next is this kind of variable tuning assignments -- the ability to kind of morph how much tuning possibility each note has so that it is somewhat greater than 1/8th of a semitone because, as I have explained, this is kind of on the edge of human pitch perception, and the human ear tends to feel that if one sometimes plays a little bit that way and a little bit that way, then there is a central note intended which the player or singer is really aiming for ... and I want enough variations in the tunings so that it is very clear to the ear that this is not at all the case, but that something else is going on.

Phew.

Hope I have not bored you all rigid. But that's what I'm looking for. I am again away from home, but I hope to submit my balaphon patch with the best I can do for variable tunings this weekend.

That's a long explanation indeed, and I'm not quite sure whether I understand it all (actually I'm quite sure I do not ...). But let's see, here is an iso5 patch in which I try to make sense of what you wrote about distributing notes. I left out out all the # notes, so please play the white keys only, in this patch the black keys make no sense. But the good news is - 16 note polyphony on a full modular (as the zero page usage dropped below 50 %).

I tried to distribute the notes according to what you wrote, but ended up tuning them in a way that seemed pleasant to my ears - so probably you'll want to retune it.

For comparison I added the synth part a second time - tuned 'normal'. Select by pushing buttons on the iso5 module. Maybe the ease of tuning is not quite what you expect it to be, but it should be manageable I think by assigning physical knobs to the virtual tuning knobs. Anyway I can't come up with a better way.

When you'd want finer tuning capabilities the patch could be modified a bit. As thing are now the tuning knobs need 96 positions for an octave, and this can easily be increased to 128 steps by adding an extra control mixer after the sample and hold module (hold calc. note). Feed the signal through it and set the knob to 95 (or 96 maybe). But I thought 96 was nice because it can be divided by 12 (so a perfect chromatic tuning can be achieved).

Also added an auto version.

Hi Jan,

> That's a long explanation indeed

I am afraid I am often long-winded!! Don't know if you have noticed. But ... it's because I try to be as clear as mathematicians are, except I never learned the conciseness of those equations and formulae.

> I'm not quite sure whether I understand it all (actually I'm quite sure I do not...)

Judging by what you provided, you understood it quite well enough!!

> so please play the white keys only

You're talking to an African here Jan, so please be very careful what you say about this kind of thing. (THAT WAS JUST A LITTLE JOKE, GUYS OK!!)

> the black keys make no sense.

Like I said, be very careful what you say sometimes.

> But the good news is - 16 note polyphony on a full modular (as the zero page usage dropped below 50 %).

That is indeed good news.

> I tried to distribute the notes according to what you wrote, but ended up tuning them in a way that seemed pleasant to my ears - so probably you'll want to retune it.

Well ... unfortunately I am not at home again so I can't actually hear the wretched thing.

> For comparison I added the synth part a second time - tuned 'normal'. Select by pushing buttons on the iso5 module.

Good idea.

> Maybe the ease of tuning is not quite what you expect it to be, but it should be manageable I think by assigning physical knobs to the virtual tuning knobs.

I'll see.

> Anyway I can't come up with a better way.

Well ... if you can't then I don't think I will either. Your way is way better than anything I could think of.

> When you'd want finer tuning capabilities the patch could be modified a bit. As thing are now the tuning knobs need 96 positions for an octave, and this can easily be increased to > 128 steps by adding an extra control mixer after the sample and hold module (hold calc. note). Feed the signal through it and set the knob to 95 (or 96 maybe). But I thought 96 was > nice because it can be divided by 12 (so a perfect chromatic tuning can be achieved).

I'll let you know how they work out in practice ... and ... expect a balaphon sounding instrument named in your honour to arrive in your mailbox shortly after I get home and finish the thing.

>Friday Child wrote: so please play the white keys only ... You're talking to an African here Jan, so please be very careful what you say about this kind of thing.

> Jan wrote: here is an iso5 patch in which I try to make sense of what you wrote about distributing notes. I left out all the # notes, so please play the white keys only...

I've been following the tuning thread with interest, and this latest set of patches (iso5.pch etc.) look great. Very clever. I was about to suggest omitting modules associated with accidentals which are not in the desired scale to reduce DSP, but I see others realized it as well. However, Clavia please note - lots of other synths have had this type of capability for many years (Yamaha, Korg, and Ensoniq for example). They often provide several factory preset scales and give the user the option to create additional scales and store them independent of the patches, and then you select which scale you want to use within the individual patch.

In the NM, these cold be provided in the form of a keyboard processing module. I'm picturing something that looks like a key quantize module, with twelve knobs or sliders for the twelve keys. Input the note data from the Keyboard Out module, move the knobs/sliders or select a preset scale from a drop down menu (like on the drum module), and viola! Perhaps something like this could be on a NM update TO DO list?

David Peck wrote: In the NM, these could be provided in the form of a keyboard processing module. I'm picturing something that looks like a key quantize module, with twelve knobs or sliders for the > twelve keys. Input the note data from the Keyboard Out module, move the knobs/sliders or select a preset scale from a drop down menu (like on the drum module), and viola! > Perhaps something like this could be on a NM update TO DO list?

I was thinking along the same line, and indeed such a module would be a very nice and welcome addition. Especially as this would detach the scaling from the actual keyboard, so it can be used for noodles as well.

There's a but though - I think- it would be limited to scales with (maximaly) 12 notes / octave, and I was wondering if maybe a construction could be realized that is chainable, more or less the way that sequencers can be chained. Or maybe a knob to scale the whole module and another one (or two, for fine control) to set an offset value.

I'm picturing something that looks like What Emu does for this sorta thing: preset scales, plus user scales

I reviewed this question of how other synths provide a method for custom tunings, and greater-than-12-tone-octaves, and my personal opinion is that the method used on the old Ensoniq VFX (& related synths) is really nice. It allows you to retune any of the 127 midi notes individually, and it also has a set of INTERPOLATE and EXTRAPOLATE controls to make things much easier.

First, you select the number of tones per octave. For example, you press E4 and G5 to establish a 16 tone scale. You then press the INTERPOLATE button to divide the 16 tones into equal spaces. (You can skip these steps if you want a scale of 12 or less tones).

Then, you individually fine tune the 16 tones as needed.

Then, you press the EXTRAPOLATE button to replicate your custom 16 tone scale above and below the E4 to G5 range.

It's a comprehensive and user-friendly system. But it doesn't seem like the kind of thing that could be done entirely in a module. Perhaps the module would just have a scale selection window with up/down arrows like the timbre select on a timbral osc, and you would actually create, name, and save scales in a new type of tool accessed from the tool menu.

CLAVIA - if you have any plans to implement a custom scale feature, please DO NOT make it a global synth setting. It should be patched into a patch via a module, so you can use different scales at the same time (even within the same patch).

> David Peck wrote: In the NM, these could be provided in the form of a keyboard processing module. I'm picturing something that looks like a key quantize module, with twelve knobs or sliders for the twelve keys. ...

There's a but though - I think- it would be limited to scales with (maximaly) 12 notes / octave, and I was wondering if maybe a construction could be realized that is chainable...

Sounds great, feel honoured, thanks !

I'll certainly try a few things with this, later on this evening. I already did some noodles with this tuning stuff, but during the process of recording they got so heavily deconstructed that I dare not those.

The balaphon patch I just submitted got me thinking. Since the kind of things that are 'relevant' to it -- i.e. being able to vary the tunings -- are not the kind of thing most people do, perhaps a few words on the idea behind it might not go amiss.

It's a bit depressing, actually, to see people stumble on some exotic new instrument, and then hear them set about treating that instrument the same as every other instrument they're used to.

May not seem relevant ... but ... a long time ago Pope Gregory put a strait jacket around music by laying down a set of rules about what music was and what it wasn't, and about how things should be played. In one fell swoop he set the cause of music back by about 500 years. (OK ... I'm exaggerating. But not about the 500 years bit. Closer to a thousand, actually).

A big problem with Gregorian chant is that it is kind of horizontal -- single lines flowing along. As the 'Middle Ages' came to an end, choir masters and composers had begun to experiment with the style a little. They would take different lines of a chant, and then repeat them; sing them out of order; put different lines together to see how they sounded; and things like that. They gradually realized that some notes and parts sounded 'good', 'harmonious', when played together. So, gradually, composers began to experiment with this idea of different lines of music that could be played simultaneously, and that flowed along together, and that somehow supported each other even though they were not the same.

One of the people famous for conducting those kinds of experimentations was one Vincenze Galilei -- who had a far more famous son called Galileo Galilei. HE is so famous that he is immediately identifiable only by his first name. Kind of like Madonna.

Anyway, Vincenze was a member of a group of people called the Camerata. What they were doing was so revolutionary, so outrageous and so new that they were breaking all known rules of music, meaning that they had to do their stuff in secret. Since they met behind closed doors or 'in camera' to discover and make music they eventually got their name of Camerata.

What the people of the Camerata eventually discovered was 'the key centre'. (Well ... there were interludes and stuff like 'figured bass' and so forth, but I'm not writing an encyclopaedia article here, just giving a general indication). That is to say, by talking and thinking about KEYS, the musicians of the Camerata came up with a powerful new organizing structure and concept for making and writing music. There were two other reasonably simultaneous discoveries that established this new way of making music as something so powerful that we still have it today. One was -- a new way of writing music down along with the discovery of printing. The other was a unique and powerful instrument for exploiting the key centre: the clavichord, harpsichord, clavinet, fortepiano, pianoforte or whatever other name was given to those keyboardy type instruments.

If you think about what those people achieved, you will see what a thing they came up with. If you can read music and can play a keyboard, then you can play and convey anything in Western music simply by interpreting those notes and then waggling your fingers over a keyboard. And ... key centres ... or music played in specific keys harmonized in given ways ... is pretty much all anyone knows now. At least ... in the Western world. (Please don't get me started on all that post-tonality stuff).

The Balaphon patch I just submitted is an instrument that belongs to a way of making music that is not in the least concerned with key centres. It is not concerned with playing along with other instruments in a harmonically governed kind of way.

As I explained in another post (search the archives), most African instruments are played in a polyrhythmic setting, and along with other instruments of variable tunings and tuning systems. Therefore, African instruments do not always sound 'good' when trying to play the same kind of thing with each other. Their notes, overtones and harmonies do not always gel well. This is because of the lack of a decent tuning fork and the love for variable tunings. In Ancient China, for example, whenever a new emperor ascended the throne, one of his (or her) first duties was to review the Empire's note and tuning system. Therefore, when African musicians play with each other they tend to play in an agreed polythythmic kind of way, and with an agreed upon method of passing and handing notes back and forth. They do not play many things simultaneously.

A variably tunable instrument is actually a very nice way to explore some new musical textures. However, in order to do this it is best not to 'swamp' them with other Western-style instruments because those other instruments will assuredly be of the equal-tempered variety. Therefore, the opportunity to really HEAR and get to know the possibilities of different tunings and systems is completely lost. If anything, fire up a sequencer module and input a shortish 5 or 10 note riff. Then leave a gap so that you can do a 'call and response' thing with it. It plays and then you respond. Really listen to the tunings. It plays, you respond.

In the Jan's Balaphon patch I just submitted I explained that I tuned the F to C 'happy' and the C to F 'sad'. Try play around with that. Sequence a phrase creating a effect, and then either enhance that effect with another choice of tunings and notes, or else counter it. In other words create a matching or complementary tuning.

A tuning system is just as valid a way of trying to establish a mood, or make a statement, as is a chord sequence or rhythmic pattern. Shame not to get to know about them. That's all I'm saying. Why get stuck in a rut. Patch up your own instrument. Then turn them tuning knobs. Pope Gregory died CENTURIES ago. And ... key centres have been around way too long. Time to move on. Tune your ears.

(PS Very sorry indeed for any typos here ... but I am WAY too tired to go back and check anything).

> Friday's Child wrote: The balaphon patch I just submitted got me thinking. Since the kind of things that are 'relevant' to it -- i.e. being able to vary the tunings -- are not the kind of thing most people do, > perhaps a few words on the idea behind it might not go amiss.

I have a tune in my head, and on tapel, from a radio programme on world music. I can whistle it, sing it, don't really need the tape anymore, but it never works on a keyboard or even on my guitar - strange.

On your balaphone however it suddenly lives (the B is not quite right as it is, but that's a nice entry for experimentation). This sort of is a miracle for me, a new way to look at things. I did know about other tuning systems, but never realized it could be that important for my own musical perception. Or sound perception, I'm not very sure about what music really is - maybe just a chase for the lost chord, or monotov, glitch. The arts pages in my newspaper today told me that techno is out, it's bossa nova now ... could have known it, of course.

Is there any musical theory about the subject of tuning systems, preferably on the web, preferably in a format suitable for people with a background in physics rather than musicology (or what's it named) ? Maybe CMJ - I'll try that one for myself .. later. This is going to cost me years, if done well ....

There is hope though, after listening to a strange tuning for a while the western way sounds funny, suddenly ... so aparently it is possible to discriminate tuning systems into classes - that's a good starting point for discovery.

Anyway, that's what my thoughts were thinking.

Or more concrete, to the point, I don't really understand what you mean when you speak about distribution of notes over, contributing some percentage of themselves to, neighbour notes.

There's a little bit of stuff on Wendy Carlos' site about alternate tunings.

Some in an interview here:

Some "Alternate Tuning Guides" (?) here:

There's also an interesting piece here which has some theory as related to rabbit reproduction!?! Actually, it's not as weird as it sounds:

Interesting topic on a number of levels. "Beauty in the Beast" is a MUST listen if you're interested in alternate tunings, and is a masterwork of synthesis as well. Should be in any serious electronic musicians library.

I remember how much my wife HATED this CD the first time she heard it. She still hates it. Though she's Chinese, she's American born, and has perfect pitch. She can't get past the fact that it's "out of tune" to her ears regardless of how hard she tries no ignore that fact. Her ears are tuned to the European 12 tone system, and can't seem to get past that.

I love anything that makes noise myself, regardless of the pitch, and thought "Beauty in the Beast" to be an amazing and beautiful work. Guess my ears are more accepting than hers.

Further thinking on this:

I used to have a Yamaha TX802, which allowed complete access to the tuning tables. There were a number of preset tables, and then you could create your own. I use a Yamaha KX88 as a master controller, and you could set the TX802 up to have a one octave range across the entire keyboard if you wanted to...88 notes per octave.

I spent a couple of months exploring ideas using various tunings on the instrument, playing around with odd scales, but haven't done much in this area since. Hadn't really thought about trying it on the Nord.

I notice that the balaphon patch is only tunable "per octave"....

Is there any musical theory about the subject of tuning systems, preferably on the web, preferably in a format suitable for people with a background in physics rather than musicology (or what's it named) ? Maybe CMJ - I'll try that one for myself .. later. This is going to cost me years, if done well ...

A great place to start is the Just Intonation Network:

There are a zillion links to various tuning sites as well.

The tuning "bible" is Harry Partch's GENESIS OF A MUSIC, an amazing book by an amazing guy. It's quite technical, though.

I'll second the recommendation for BEAUTY IN THE BEAST and Carlos' site.

Thank you Les and Tim,

I'll have a look into this tonight.

To complete things, here are tuning systems for 5, 7 and 12 notes / octave. I set them equally tempered, but that can be changed, of course.

I've also been working on a scale with 19 notes / octave - but can't get it right. Give it up for now, added the patch, but please note that it's a failure really.

Les Mizzell remarked that the scales presented so far are all octave based - this is true, but it could be easily changed by adding an amplifier module after the main note output (on the S&H module) so the octave can be re-scaled at will. This however will also require you to re-tune all other tuning knobs.

A good thing to know is that the scaler modules (the ones that follow the discriminator chain) have a sensitivity of an octave / 96 steps, or a semi tone / 8 steps, or 12.5 cents / step.

I typed some info into the floating notes for easy reference (see tools/notes floater).

In the Balaphone patch made by Kofi I noticed him using the gate output of the keyboard module to trigger the envelope generators. Normally this will not be a problem, but in general it would be a better idea to use the the delayed gate from the module marked trig&hold (as the note output will only be valid after the rising edge of that module's output occurred).