Forum Index Register Members List Events Mark Forums Read Help

 International Skeptics Forum Recent pieces of mine, with links

 Welcome to the International Skeptics Forum, where we discuss skepticism, critical thinking, the paranormal and science in a friendly but lively way. You are currently viewing the forum as a guest, which means you are missing out on discussing matters that are of interest to you. Please consider registering so you can gain full use of the forum features and interact with other Members. Registration is simple, fast and free! Click here to register today.
 21st December 2012, 03:41 AM #41 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by aggle-rithm Long ago I had the delusion that I could learn to be a piano tuner. The real trick, I soon found out, is to make every interval other than the octave just the right amount OUT of tune. Yeah, I tried it myself briefly and realized it was something better left to a pro. There's also the issue of how much you stretch the octaves in the bass.
 21st December 2012, 03:52 AM #42 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 I decided, in the last few days, that I needed to make some small changes to my tuning/scale. It's still 36 pitches in 13.793-cent increments (87EDO) which approximate 11-limit JI, with a 13/12 thrown in there. To change a Logic Pro "orchestra" with around 50 instruments in it to this tuning took around 5 hours of work -- tedious, but not prohibitive. This is the way the application Lil' Miss Scale Oven displays the scale, with some editing. It has way too much precision, I don't need anything more accurate than a tenth of a cent. // Scale Name: "rbhji87" // Repeat Ratio: 2:1 // Scale Pattern: 0., 110.345, 137.931, 151.724, 165.517, 179.31, 206.897, 234.483, 262.069, 303.448, 317.241, 344.828, 386.207, 413.793, 441.379, 496.552, 537.931, 551.724, 579.31, 620.69, 648.276, 662.069, 703.448, 758.621, 786.207, 813.793, 855.172, 882.759, 910.345, 937.931, 965.517, 993.103, 1020.69, 1034.483, 1048.276, 1089.655 // Scale Pattern Type: Cents Absolute // as cents difference: 110.345000, 27.586000, 13.793000, 13.793000, 13.793000, 27.587000, 27.586000, 27.586000, 41.379000, 13.793000, 27.587000, 41.379000, 27.586000, 27.586000, 55.173000, 41.379000, 13.793000, 27.586000, 41.380000, 27.586000, 13.793000, 41.379000, 55.173000, 27.586000, 27.586000, 41.379000, 27.587000, 27.586000, 27.586000, 27.586000, 27.586000, 27.587000, 13.793000, 13.793000, 41.379000, 110.345000 // Scale Pattern: 8, 2, 1, 1, 1, 2, 2, 2, 3, 1, 2, 3, 2, 2, 4, 3, 1, 2, 3, 2, 1, 3, 4, 2, 2, 3, 2, 2, 2, 2, 2, 2, 1, 1, 3, 8 // Scale Pattern: 0, 8, 10, 11, 12, 13, 15, 17, 19, 22, 23, 25, 28, 30, 32, 36, 39, 40, 42, 45, 47, 48, 51, 55, 57, 59, 62, 64, 66, 68, 70, 72, 74, 75, 76, 79, 87 Here are my premises: 1) use off-the-shelf gear, not specially-designed gear. 2) no activism -- this is not a social cause. 3) approximate JI, but a little beating, a little waver, is good. 4) number of notes is some multiple of 12 -- 36 is not to much, too few. Reason is that it will be easier to memorize a multiple of 12 pattern on a conventional keyboard layout. (This relates to 1) ). 5) 11-limit and as much 13-limit as you can get with 36 pitches. 6) moveable tuning-base -- (like the slide on a slide trombone). 7) equal step-sizes are not important at all, which is closely related to 8). 8) the 36-note set has many patterns, many subsets embedded in it, but it is not really a "thing" or a gestalt, in itself -- it's just the pitches you most often need to grab at once in one place for convenience. 9) I'm a scale-dweller rather than a scale-nomad: I practice a given scale/tuning for years and years. Others change scales more frequently. Because I practice something for a long time, complexity/difficulty is no objection. So a large number of pitches and unequal step-sizes are ok. 10) scale/tuning should be able to do everything that 12EDO can do and more. IMO, it turns out that Harry Partch was that rare critter: A better theorist than he was a composer. (Opposite of Hindemith or Schoenberg) Last edited by calebprime; 21st December 2012 at 04:04 AM.
 7th January 2013, 02:08 PM #43 asydhouse Master Poster     Join Date: Feb 2012 Location: Swansea in the UK Posts: 2,368 Hi Caleb, I just discovered this thread! Been listening to the first three pieces you posted in the OP.... enjoying the tonalities, if that's the word. I think this third piece is the most pleasing to me so far... the second one was a mite too eery for my comfort... I'm just pottering about getting ready to fly to India day after tomorrow, so it's not full concentration... hence my simple response referencing "comfort", which I realise is not the main consideration in these enterprises... I'll listen again in February after we get back!
 2nd February 2013, 10:53 AM #44 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by asydhouse Hi Caleb, I just discovered this thread! Been listening to the first three pieces you posted in the OP.... enjoying the tonalities, if that's the word. I think this third piece is the most pleasing to me so far... the second one was a mite too eery for my comfort... I'm just pottering about getting ready to fly to India day after tomorrow, so it's not full concentration... hence my simple response referencing "comfort", which I realise is not the main consideration in these enterprises... I'll listen again in February after we get back! Thanks for the listen and comment! Yeah, that piece is derivative of Bartok's Music for Strings, Percussion and Celeste (though he's far better) which sounds eerie to most normal people. But not too eerie for...children of the night.
 2nd February 2013, 11:14 AM #45 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Thought I'd post this here, rather than my old thread called 'another of cp's boring music theory threads' iirc. I still hope to write many microtonal pieces, but right now I want to write in good old 12-tone tuning, partly because it really is one of the "best" tunings if you like 5ths and not too many pitches! I thought I'd do another MOF-row piece. Today I found such a series, after quite a lot of hunting around, so I'm happy. (A rare thing) (transposition not important) C#,F,G#,D,Bb,F#,E,A,Eb,B,C,G This is a MOF series (Multiple Order Function). That's the term -- more or less -- for a series which has a version of itself embedded in itself. There are hundreds of possible complete MOF series, but relatively few that sound good to me. And there are many in which I wouldn't know how to exploit the MOF property in a way that sounds good. But this one works for me. It's derived by first defining a pattern of self-similarity (see user row), then seeing whether that pattern will support a complete permutation orbit with a chromatic scale (see grid row and orbit). This is an n-1 permutation, meaning that one note doesn't orbit, but stays fixed -- the A. Fortuitously, that pitch happens to land in a place that doesn't mess up the MOF property. Most n-1 perms have messed-up self-similarity (or MOF function). user row: C..D..Eb.F..G..Ab.Bb.B..C#.E..F#.A grid row:.Bb.B..C..C#.D..Eb.E..F..F#.G..Ab.A orbit:....Bb.C..Eb.Ab.F#.C#.F..B..D..G..E..A series:...C#.F..Ab.D..Bb.F#.E..A..Eb.B..C..G (sounds good!) intrvals: 4..3..6..8..8..10.5..6..8..1..7..6 ...................d.....f#....a..eb.b.....g .............f........bb....e...........c... ..........c#....g# 1) has tight embedding 2) good hexachords (014579 and 013589) 3) enough intervallic variety 4) good in relation to potential pedal-tones c#,e,g# (C# minor or 037) 5) tonal but not too much, no dead spots, nothing silly 6) Multiple Order Function is usable without it sounding like ca-ca. The B-C move in relation to C# is actually a sound I like a lot: b7 to 7 -- puts some tension on the tonic (C#) but doesn't wipe it out. Last edited by calebprime; 2nd February 2013 at 11:36 AM.
 5th February 2013, 06:00 AM #46 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... ...another MOF-row piece. ... (transposition not important) C#,F,G#,D,Bb,F#,E,A,Eb,B,C,G This is a MOF series (Multiple Order Function). That's the term -- more or less -- for a series which has a version of itself embedded in itself. ... It's derived by first defining a pattern of self-similarity (see user row), then seeing whether that pattern will support a complete permutation orbit with a chromatic scale (see grid row and orbit). This is an n-1 permutation, meaning that one note doesn't orbit, but stays fixed -- the A. Fortuitously, that pitch happens to land in a place that doesn't mess up the MOF property. Most n-1 perms have messed-up self-similarity (or MOF function). user row: C..D..Eb.F..G..Ab.Bb.B..C#.E..F#.A grid row:.Bb.B..C..C#.D..Eb.E..F..F#.G..Ab.A orbit:....Bb.C..Eb.Ab.F#.C#.F..B..D..G..E..A series:...C#.F..Ab.D..Bb.F#.E..A..Eb.B..C..G (sounds good!) intrvals: 4..3..6..8..8..10.5..6..8..1..7..6 ...................d.....f#....a..eb.b.....g .............f........bb....e...........c... ..........c#....g# 1) has tight embedding 2) good hexachords (014579 and 013589) 3) enough intervallic variety 4) good in relation to potential pedal-tones c#,e,g# (C# minor or 037) 5) tonal but not too much, no dead spots, nothing silly 6) Multiple Order Function is usable without it sounding like ca-ca. The B-C move in relation to C# is actually a sound I like a lot: b7 to 7 -- puts some tension on the tonic (C#) but doesn't wipe it out. attempt at being more clear: Purpose of this kind of series: It's one way that the harmonies and lines can be controlled and derived from the same material. Unity, in short. Vaguely fractal idea: The same patterns occur at different speeds. Problem with this method: It only determines a pattern of self-similarity, it doesn't determine the other qualities of the series it makes -- whether it ends up sounding good is chance*. So you have to make hundreds of these before you find one you like, if you're choosy. Better explanation of this method of making MOF series: I'll use numbers instead of pitch-classes. grid row:.10.11.0..1..2..3..4..5..6..7..8.(9) user row: 0..2..3..5..7..8..10.11.1..4..6..9 (10 goes to 0, 0 goes to 3, 3 goes to 8, 8 goes to 6, etc. 9 stays fixed here) orbit:....10.0..3..8..6..1..5..11.2..7..4.(9) one-line permutation series:...1..5..8..2..10.6..4..9..3..11.0..7 swap values, positions of one-line perm intrvals: 4..3..6..8..8..10.5..6..8..1..7..6 differences ...................2.....6.....9..3..11....7 (series + 1, in two cycles of series) .............5........10....4...........0... ..........1.....8. This is Mod 12 arithmetic -- everything is thought of like a clock... C=0, Db and C#=1, D=2, D# and Eb=3, E=4, etc. Start with the pattern of embedding. Is there some alignment of that series of numbers with some version of 0,1,2,3,4, etc. that produces a complete orbit? Or second-best, one where 11 elements orbit, and one stays fixed? Having come up with a one-line permutation, think of that series as a set of paired order-numbers (0 to 11) and values, like so: (10.0..3..8..6..1..5..11.2..7..4.)(9) one-line permutation is 0,10 1,0 2,3 3,8 4,6 5,1 pairs of order-numbers and values Then, to get the final result, you swap these pairs. Instead of 10 in position 0, you have 0 in position 10 Instead of 0 in position 1, you have 1 in position 0 Instead of 3 in position 2, you have 2 in position 3 etc. to get 1..5..8..2..10.6..4..9..3..11.0..7 converting these numbers to pitch-classes, you get your series. Now, does it sound good? Could it be the source of a piece? Probably not. Rinse and repeat. *Well, not chance really, but not foreseeable or "forehearable" if you will Last edited by calebprime; 5th February 2013 at 06:12 AM.
 20th February 2013, 07:07 PM #47 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... I thought I'd do another MOF-row piece. ... (transposition not important) C#,F,G#,D,Bb,F#,E,A,Eb,B,C,G ... https://www.box.com/s/litnhwpcyk8swpk46vo7 What's on paper seems good to me, but this sketch is ponderous. Only took about three hours from paper to this 9+ minute realization. So I'm calling it Sea Bass with Frickin' Lasers. It will take a lot more work to make something with more subtlety than this. One problem is that I suck on piano. Another is that this was just slapped together. In that light, it came out pretty good. Trying to exactly make the Multiple-Order embedding sound good turned out to be harder than it looked on paper. This is only an .aif file, so it's 95+ megs, so it may not play directly, you may need to download it.
 22nd February 2013, 06:05 AM #49 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Looking for another kernel to base a piece on. The disadvantage of the kind of series I just used is that the self-similarity is irregular, unlike another kind of series, the indices of power-residue series. The indices of power residues have perfect, regular self-similarity. There are only two basic sources for such series in the land of conventional 12-tone tuning, because there are only two primes in the vicinity of 12: 11 and 13. The series derived from these primes will have one less element, so 10 and 12 numbers, respectively. mod 13 generates the Mallalieu series, so let's try mod 11. 2^n mod 11: 1,2,4,8,5,10,9,7,3,6 indices: 0,1,8,2,4,9,7,3,6,5 eta: If we Google this series, we see that it's well-known in number-theory, etc: http://www.google.com/search?q=%220%....1.L_kl1vTeSrE eta2: also, in the OEIS: http://oeis.org/search?q=0%2C1%2C8%2...lish&go=Search 6^n mod 11 indices (0 to 9): 9,8,1,7,5,0,2,6,3,4 is simply inversion of above 7^n mod 11 indices: 9,2,3,5,1,6,0,8,7,4 [eta4: removed error] 8^n mod 11 is only the inversion of above. However, the resulting tables of number series can be mapped to any set of pitches. It can be thought of as a source of contours rather than a more literal 0=C, 1=C#, 2=D, mapping. Here's such a table: 0, 1, 8, 2, 4, 9, 7, 3, 6, 5 1, 2, 9, 3, 5, 0, 8, 4, 7, 6 2, 3, 0, 4, 6, 1, 9, 5, 8, 7 3, 4, 1, 5, 7, 2, 0, 6, 9, 8 4, 5, 2, 6, 8, 3, 1, 7, 0, 9 5, 6, 3, 7, 9, 4, 2, 8, 1, 0 6, 7, 4, 8, 0, 5, 3, 9, 2, 1 7, 8, 5, 9, 1, 6, 4, 0, 3, 2 8, 9, 6, 0, 2, 7, 5, 1, 4, 3 9, 0, 7, 1, 3, 8, 6, 2, 5, 4 Here's the 7^n mod 11 table 9, 2, 3, 5, 1, 6, 0, 8, 7, 4 2, 5, 6, 8, 4, 9, 3, 1, 0, 7 5, 8, 9, 1, 7, 2, 6, 4, 3, 0 8, 1, 2, 4, 0, 5, 9, 7, 6, 3 1, 4, 5, 7, 3, 8, 2, 0, 9, 6 4, 7, 8, 0, 6, 1, 5, 3, 2, 9 7, 0, 1, 3, 9, 4, 8, 6, 5, 2 0, 3, 4, 6, 2, 7, 1, 9, 8, 5 3, 6, 7, 9, 5, 0, 4, 2, 1, 8 6, 9, 0, 2, 8, 3, 7, 5, 4, 1 When the series are mapped to more-or-less familiar scales, you can easily generate polyrhythmic textures moving at 1:2:3:4:5 speeds -- or however complicated you want -- with the overall sound being a pleasing "chord-scale" rather than the more gnarly 12-notes-at-once sound. And these scales can be tuned to more-or-less conventional tunings. 10 pitches per octave (10EDO) would be one obvious choice. (And I like the sound of 10EDO, somewhat.) Or the "scale" can be any arbitrary set of pitches ranked from low to high. Last edited by calebprime; 22nd February 2013 at 06:42 AM.
 23rd February 2013, 12:04 AM #50 Orphia Nay Penguilicious Spodmaster.Tagger     Join Date: May 2005 Location: Ponylandistan Presidential Palace (above the Spods' stables). Posts: 36,315 Originally Posted by calebprime https://www.box.com/s/litnhwpcyk8swpk46vo7 What's on paper seems good to me, but this sketch is ponderous. Only took about three hours from paper to this 9+ minute realization. So I'm calling it Sea Bass with Frickin' Lasers. It will take a lot more work to make something with more subtlety than this. One problem is that I suck on piano. Another is that this was just slapped together. In that light, it came out pretty good. Trying to exactly make the Multiple-Order embedding sound good turned out to be harder than it looked on paper. This is only an .aif file, so it's 95+ megs, so it may not play directly, you may need to download it. That was rather lovely, caleb. Subtle, but not too subtle. Love the title! __________________ Challenge your thoughts. Don't believe everything you think.
 24th February 2013, 10:41 AM #52 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Still looking for the next little kernel. It occurs to me that if I'm going to be mapping number series to scales with 7 elements -- such as most of the familiar scales -- I should be looking for self-similar sequences of either 7 elements or 14 elements. If my goal is the standard "every-other" permutation pattern, there turn out to be no perfect patterns of 7 elements. But there are a lot of near-perfect patterns of 14 elements, so I'll be looking at these over the next few days. The actual series -- the results -- are here called "x-swapped". The series, from 0 to 13, will be mapped onto 2 octaves of a 7-note scale. Here are what seem to be the only four candidates, each with its quirks. user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 1 2 3 4 5 6 7 8 9 10 11 12 13 0 orbit: 1 0 13 11 7 12 9 3 4 6 10 5 8 2 X-swapped: 1 0 13 7 8 11 9 4 12 6 10 3 5 2 intervals: 13 13 8 1 3 12 9 8 8 4 7 2 11 13 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 3 4 5 6 7 8 9 10 11 12 13 0 1 2 orbit: 3 0 9 12 5 4 2 13 7 8 10 1 11 6 X-swapped: 1 11 6 0 5 4 13 8 9 2 10 12 3 7 intervals: 10 9 8 5 13 9 9 1 7 8 2 5 4 8 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 11 12 13 0 1 2 3 4 5 6 7 8 9 10 orbit: 11 0 6 5 3 12 2 10 13 4 1 8 9 7 X-swapped: 1 10 6 4 9 3 2 13 11 12 7 0 5 8 intervals: 9 10 12 5 8 13 11 12 1 9 7 5 3 7 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 13 0 1 2 3 4 5 6 7 8 9 10 11 12 orbit: 13 0 2 6 1 4 10 9 7 3 8 5 12 11 X-swapped: 1 4 2 9 5 11 3 8 10 7 6 13 12 0 intervals: 3 12 7 10 6 6 5 2 11 13 7 13 2 1 Suppose we take the last one, and map this to A major for the moment, just to see how it works. 0=A1, 1=B1, 2=C#2, 3=D2...etc. to 6=G#2, 7=A2, 8=B2, 9=C#3...13=G#3 || 14 would equal A3 We get the sequence of pitches B1,E2,C#2,C#3,F#2,E3,D2,B2,D3,A2,G#2,G#3,F#3,A1 (repeat) and embedded in this, starting on the third note, is the sequence at "every other note" C#2,F#2,D2,D3,G#2,F#3,E2,C#3,E3,B2,A2,(G#3,A1(or adjusted) Which is identical to the first sequence, up a diatonic step -- except for the last two notes. This process of embedding can be carried out as far as one would like, with any scales of 7 notes, not just major scales. I like music where the scales change, because of my jazzoid background. I'm going to try this for my next piece, unless I see some fatal flaw or get a better idea in the meantime. Last edited by calebprime; 24th February 2013 at 10:55 AM. Reason: fix error
 25th February 2013, 03:36 AM #53 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime Still looking for the next little kernel. It occurs to me that if I'm going to be mapping number series to scales with 7 elements -- such as most of the familiar scales -- I should be looking for self-similar sequences of either 7 elements or 14 elements. If my goal is the standard "every-other" permutation pattern, there turn out to be no perfect patterns of 7 elements. But there are a lot of near-perfect patterns of 14 elements, so I'll be looking at these over the next few days. The actual series -- the results -- are here called "x-swapped". The series, from 0 to 13, will be mapped onto 2 octaves of a 7-note scale. Here are what seem to be the only four candidates, each with its quirks. user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 1 2 3 4 5 6 7 8 9 10 11 12 13 0 orbit: 1 0 13 11 7 12 9 3 4 6 10 5 8 2 X-swapped: 1 0 13 7 8 11 9 4 12 6 10 3 5 2 intervals: 13 13 8 1 3 12 9 8 8 4 7 2 11 13 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 3 4 5 6 7 8 9 10 11 12 13 0 1 2 orbit: 3 0 9 12 5 4 2 13 7 8 10 1 11 6 X-swapped: 1 11 6 0 5 4 13 8 9 2 10 12 3 7 intervals: 10 9 8 5 13 9 9 1 7 8 2 5 4 8 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 11 12 13 0 1 2 3 4 5 6 7 8 9 10 orbit: 11 0 6 5 3 12 2 10 13 4 1 8 9 7 X-swapped: 1 10 6 4 9 3 2 13 11 12 7 0 5 8 intervals: 9 10 12 5 8 13 11 12 1 9 7 5 3 7 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 13 0 1 2 3 4 5 6 7 8 9 10 11 12 orbit: 13 0 2 6 1 4 10 9 7 3 8 5 12 11 X-swapped: 1 4 2 9 5 11 3 8 10 7 6 13 12 0 intervals: 3 12 7 10 6 6 5 2 11 13 7 13 2 1 Suppose we take the last one, and map this to A major for the moment, just to see how it works. 0=A1, 1=B1, 2=C#2, 3=D2...etc. to 6=G#2, 7=A2, 8=B2, 9=C#3...13=G#3 || 14 would equal A3 We get the sequence of pitches B1,E2,C#2,C#3,F#2,E3,D2,B2,D3,A2,G#2,G#3,F#3,A1 (repeat) and embedded in this, starting on the third note, is the sequence at "every other note" C#2,F#2,D2,D3,G#2,F#3,E2,C#3,E3,B2,A2,(G#3,A1(or adjusted) Which is identical to the first sequence, up a diatonic step -- except for the last two notes. This process of embedding can be carried out as far as one would like, with any scales of 7 notes, not just major scales. I like music where the scales change, because of my jazzoid background. I'm going to try this for my next piece, unless I see some fatal flaw or get a better idea in the meantime. Heh. Not a fatal flaw exactly. When I took out two pitches and mapped this number series to a good-old chromatic scale, I came up with: C,Eb,Db,Ab,E,Bb,D,G,A,Gb,F,B First thing I noticed was that this is a rather good 12-tone MOF. Maybe too good. Next thing I did was write down the intervals: 3,10,7,8,6,4,5,2,9,11,6,1. Uh oh. Good MOF, and is an all-interval series. Next thing I did was look it up in my data-base of all-interval series. Yep. It's the Mallalieu series. There's nothing wrong with it, it's the only perfect series there is. But I've spent years looking for alternatives to it, for series that have many of its properties but lack perfection in some way. Here's the entry in the my list. Mallalieu is #44. 044 014295B38A76014295B38A76 5*/0 Mallalieu series C,Db,E,D,A,F,B,Eb,Ab,Bb,G,F#,C,Db,E,D,A,F,B,Eb,Ab, Bb,G,F# 1,3,10,7,8,6,4,5,2,9,11,6,1,3,10,7,8,6,4,5,2,9,11,6 Slonimsky's #1302 11,9,2,5,4,6,8,7,10,3,1,6,11,9,2,5,4,6,8,7,10,3,1, 6 7,9,10,1,8,6,4,11,2,3,5,6,7,9,10,1,8,6,4,11,2,3,5, 6 5,3,2,11,4,6,8,1,10,9,7,6,5,3,2,11,4,6,8,1,10,9,7, 6, 19^n mod37 also Slonimsky's #1308 Maybe I should just write a piece based on Mallalieu series, or maybe I should keep looking. Thing is, I like the way it sounds starting on the second note.
 25th February 2013, 06:32 AM #54 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime Heh. Not a fatal flaw exactly. When I took out two pitches and mapped this number series to a good-old chromatic scale, I came up with: C,Eb,Db,Ab,E,Bb,D,G,A,Gb,F,B First thing I noticed was that this is a rather good 12-tone MOF. Maybe too good. Next thing I did was write down the intervals: 3,10,7,8,6,4,5,2,9,11,6,1. Uh oh. Good MOF, and is an all-interval series. Next thing I did was look it up in my data-base of all-interval series. Yep. It's the Mallalieu series. There's nothing wrong with it, it's the only perfect series there is. But I've spent years looking for alternatives to it, for series that have many of its properties but lack perfection in some way. Here's the entry in the my list. Mallalieu is #44. 044 014295B38A76014295B38A76 5*/0 Mallalieu series C,Db,E,D,A,F,B,Eb,Ab,Bb,G,F#,C,Db,E,D,A,F,B,Eb,Ab, Bb,G,F# 1,3,10,7,8,6,4,5,2,9,11,6,1,3,10,7,8,6,4,5,2,9,11,6 Slonimsky's #1302 11,9,2,5,4,6,8,7,10,3,1,6,11,9,2,5,4,6,8,7,10,3,1, 6 7,9,10,1,8,6,4,11,2,3,5,6,7,9,10,1,8,6,4,11,2,3,5, 6 5,3,2,11,4,6,8,1,10,9,7,6,5,3,2,11,4,6,8,1,10,9,7, 6, 19^n mod37 also Slonimsky's #1308 Maybe I should just write a piece based on Mallalieu series, or maybe I should keep looking. Thing is, I like the way it sounds starting on the second note. How about applying your 7+7 mapping idea to pentatonic scales, and using the mod11 generator to make a 10-note pattern, apply that to a 5+5 mapping? There's a lot of wiggle-room with pentatonic scales. You've got yer Balinese, Javanese, yer hemitonic & anhemitonic, yer shredder-pentatonic, and yer bro-step pentatonic. Everyone likes pentatonics! Here's the 7^n mod 11 table. map any row to a 2-octave pentatonic, or 10EDO tuning, back to that idea? 9, 2, 3, 5, 1, 6, 0, 8, 7, 4 2, 5, 6, 8, 4, 9, 3, 1, 0, 7 5, 8, 9, 1, 7, 2, 6, 4, 3, 0 8, 1, 2, 4, 0, 5, 9, 7, 6, 3 1, 4, 5, 7, 3, 8, 2, 0, 9, 6 4, 7, 8, 0, 6, 1, 5, 3, 2, 9 7, 0, 1, 3, 9, 4, 8, 6, 5, 2 0, 3, 4, 6, 2, 7, 1, 9, 8, 5 3, 6, 7, 9, 5, 0, 4, 2, 1, 8 6, 9, 0, 2, 8, 3, 7, 5, 4, 1 Last edited by calebprime; 25th February 2013 at 06:56 AM.
 25th February 2013, 12:49 PM #55 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Here's a perfect 10-note self-similar pattern, all-interval mod10. I came up with it in a weird way, as I'll show. It doesn't seem to be Mallalieu minus a couple pitches, nor does it seem to be the same as the series you get from power-residue index series. That I've discovered a genuinely new thing seems very unlikely, but I don't recognize this. As Robert DiDomenica used to say, "You'll be disappointed." Seems too good to be true: 7, 6, 9, 5, 3, 8, 0, 4, 1, 2 6, 5, 8, 4, 2, 7, 9, 3, 0, 1 5, 4, 7, 3, 1, 6, 8, 2, 9, 0 4, 3, 6, 2, 0, 5, 7, 1, 8, 9 3, 2, 5, 1, 9, 4, 6, 0, 7, 8 2, 1, 4, 0, 8, 3, 5, 9, 6, 7 1, 0, 3, 9, 7, 2, 4, 8, 5, 6 0, 9, 2, 8, 6, 1, 3, 7, 4, 5 9, 8, 1, 7, 5, 0, 2, 6, 3, 4 8, 7, 0, 6, 4, 9, 1, 5, 2, 3 and its inversion 2, 3, 0, 4, 6, 1, 9, 5, 8, 7 2, 3, 0, 4, 6, 1, 9, 5, 8, 7 3, 4, 1, 5, 7, 2, 0, 6, 9, 8 4, 5, 2, 6, 8, 3, 1, 7, 0, 9 5, 6, 3, 7, 9, 4, 2, 8, 1, 0 6, 7, 4, 8, 0, 5, 3, 9, 2, 1 7, 8, 5, 9, 1, 6, 4, 0, 3, 2 8, 9, 6, 0, 2, 7, 5, 1, 4, 3 9, 0, 7, 1, 3, 8, 6, 2, 5, 4 0, 1, 8, 2, 4, 9, 7, 3, 6, 5 1, 2, 9, 3, 5, 0, 8, 4, 7, 6 I got this result by entering these series into my permutation algorithm-thingy, and looking for complete orbits only: grid row: 0,7,3,6,2,9,5,1,8,4, user row: 8,6,4,2,0,9,7,5,3,1 and it found: user row: 8 6 4 2 0 9 7 5 3 1 grid row: 6 2 9 5 1 8 4 0 7 3 orbit: 6 8 9 4 7 3 1 0 5 2 X-swapped: 7 6 9 5 3 8 0 4 1 2 intervals: 9 3 6 8 5 2 4 7 1 5 You're too good to be true Can't take my eyes off you. I love you baby, and if it's quite alright... Can I generalize this at all -- to at least a rule of thumb, and find more? It's probably merely some transformation of a power-residue series in mod10 arithmetic. Last edited by calebprime; 25th February 2013 at 12:52 PM.
 25th February 2013, 12:59 PM #56 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... Seems too good to be true: 7, 6, 9, 5, 3, 8, 0, 4, 1, 2 6, 5, 8, 4, 2, 7, 9, 3, 0, 1 5, 4, 7, 3, 1, 6, 8, 2, 9, 0 4, 3, 6, 2, 0, 5, 7, 1, 8, 9 3, 2, 5, 1, 9, 4, 6, 0, 7, 8 2, 1, 4, 0, 8, 3, 5, 9, 6, 7 1, 0, 3, 9, 7, 2, 4, 8, 5, 6 0, 9, 2, 8, 6, 1, 3, 7, 4, 5 9, 8, 1, 7, 5, 0, 2, 6, 3, 4 8, 7, 0, 6, 4, 9, 1, 5, 2, 3 ... It's probably merely some transformation of a power-residue series in mod10 arithmetic. yep, a quick Google shows it's just a log series, or power-res index series. http://www.galaxyng.com/adrian_atana.../cript/c12.pdf De aici rezult˘a imediat tabelul logaritmilor ˆın baza 6: β 1 2 3 4 5 6 7 8 9 10 log6β 0 9 2 8 6 1 3 7 4 5
 25th February 2013, 02:41 PM #57 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Here's another perfect one that I don't recognize yet: 0, 7, 6, 4, 8, 3, 9, 1, 2, 5 7, 4, 3, 1, 5, 0, 6, 8, 9, 2 4, 1, 0, 8, 2, 7, 3, 5, 6, 9 1, 8, 7, 5, 9, 4, 0, 2, 3, 6 8, 5, 4, 2, 6, 1, 7, 9, 0, 3 5, 2, 1, 9, 3, 8, 4, 6, 7, 0 2, 9, 8, 6, 0, 5, 1, 3, 4, 7 9, 6, 5, 3, 7, 2, 8, 0, 1, 4 6, 3, 2, 0, 4, 9, 5, 7, 8, 1 3, 0, 9, 7, 1, 6, 2, 4, 5, 8 eta: oh, that's just 8 ^n mod 11 indices , but no one seems to have mentioned it so that it shows up in Google. Last edited by calebprime; 25th February 2013 at 02:44 PM.
 26th February 2013, 03:23 AM #58 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 I have another app written by 69Dodge that finds series with self-similarity within some error tolerance. That's a way of getting around the fact that there are no new perfect series under the sun. Maybe if one of the series below is mapped to a 6+6 arrangement -- 2 octaves of a 6-note scale -- there will be self-similarity that is close enough for Cal, and with the necessary euphony. The numbers after the dash on the right mean that each of these has 10 positions where the self-sim is exactly 2, the maximum error of any position is +-1, and the total sum of errors is 2. 0 1 7 3 6 9 10 5 4 8 2 11 - 10 1 2 0 2 7 3 6 9 10 5 4 8 1 11 - 10 1 2 most of these are basically mallalieu intervals x2, +-1 0 2 7 4 5 9 10 6 3 8 1 11 - 10 1 2 0 2 7 4 6 9 10 5 3 8 1 11 - 10 1 2 0 2 8 4 5 9 10 6 3 7 1 11 - 10 1 2 0 2 8 4 5 10 9 6 3 7 1 11 - 10 1 2 0 2 8 4 7 10 11 6 5 9 3 1 - 10 1 2 0 2 8 4 7 11 10 6 5 9 3 1 - 10 1 2 0 2 9 4 6 11 10 7 5 8 3 1 - 10 1 2 0 2 9 4 7 11 10 6 5 8 3 1 - 10 1 2 0 2 9 5 6 11 10 7 4 8 3 1 - 10 1 2 0 3 9 5 6 11 10 7 4 8 2 1 - 10 1 2 These can be considered either as standard 12-tone rows or as the result of their mapping to a 6-note, 2-octave scale, or both. The series below, with the search parameters above them, are interesting because the perms are as free of unpleasant or dorky contiguous harmonic relations as the listed series. So the permutatated series sounds as good as the listed one. These are self-similar at 4 semitones, or a major third. M) modulus: 12 P) permutation: 1,3,5,7,9,11,0,2,4,6,8,10 C) center interval of self-similarity: 4 O) minimum number of occurrences of center: 5 D) maximum deviation from center: 1 S) maximum sum of all deviations: 7 F) fixed-position notes: 0:0 E) excluded cells within range: 0,1,2:3 0,3,6,9:4 R) excluded intervals within range: A) filter original solution: on B) filter permuted solution: on X) permutation depth: 3 N) number of solutions to print: 1000 W) wrap: on 0 3 4 7 1 9 8 11 10 5 6 2 - 6 1 6 0 3 4 7 2 9 8 11 10 5 6 1 - 6 1 6 0 4 2 7 10 6 8 11 5 3 1 9 - 6 1 6 0 4 2 8 10 6 7 11 5 3 1 9 - 6 1 6 0 4 3 8 9 6 7 11 5 2 1 10 - 6 1 6 0 4 3 8 10 6 7 11 5 2 1 9 - 6 1 6 0 4 5 8 2 10 9 1 11 6 7 3 - 6 1 6 0 4 5 8 3 10 9 1 11 6 7 2 - 6 1 6 0 4 6 8 2 10 9 1 11 5 7 3 - 6 1 6 0 4 6 8 2 11 9 1 10 5 7 3 - 6 1 6 0 5 4 9 10 7 8 1 6 3 2 11 - 6 1 6 0 5 4 9 11 7 8 1 6 3 2 10 - 6 1 6 found 12 solutions. Last edited by calebprime; 26th February 2013 at 03:42 AM.
 28th February 2013, 06:54 AM #59 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 A new tuning for the next piece. 31 Pitches per 2/1 (octave). Basically a modified Just Intonation (JI) approach with a complete circle of near-5ths. The tuning base can be moved so that 1/1 can be any frequency. Modulation -- or moving the tonal center -- is achieved by common-tones between the old tonal center and the new one that one gets to by pitch-bending the whole structure. There are so many criteria for this tuning that I won't even attempt to list them all -- unless someone wants to know. This took a lot of thought, the help of the Lil' Miss Scale Oven (now sadly no more) and some things I learned from the microtonal folks. However, it's neither a Partch tuning nor one of the new tunings or temperaments that people such as Gene Ward Smith and Graham Breed are developing. This will be for sampled instruments in Logic retuned via Lil Miss Scale Oven. It's in Scala format, so anyone who uses that can just plug it in. (Technology is sometimes a wonderful thing!) The numbers are in cents (2/1 or octave divided into 1200 parts) There might be a flaw, but I haven't found any so far. It might be possible to smooth some rough edges by slightly adjusting the whole thing*, but I don't know of any easy way to do that with my software, and I can't think of any pitches I don't need. *eta: such as quantizing the pitches to 140EDO, or something??? !Caleb 31comp#1.scl ! First try at <36 with complete circle of 5ths, 4ths, lots of 13, 701.62 or 111.86 generator 31 ! ! 0.00 91.9 ! m2 small pyth 111.86 ! 16/15 138.573 ! 13/12 182.404 ! 10/9 203.24 ! 9/8 231.174 ! 8/7 266.871 ! 7/6 ! 305.12 ! pyth min third 315.1 ! 6/5 t 386.52 ! 5/4 t 406.48 ! 81/64 t 416.98 ! 14/11 434.414 ! 9/7 498.38 ! 4/3 ! 551.32 ! 11/8, no sym 590.28 ! pyth 7/5 ! 609.72 ! pyth 10/7 ! 701.62 ! 3/2 765.586 ! 14/9 783.02 ! 11/7 and small pyth min 6th 813.48 ! 8/5 t 840.528 ! 13/8 885.42 ! 5/3 904.86 ! pyth maj 6th ! 932.794 ! 12/7 t 968.826 ! 7/4 996.09 ! 16/9 1049.36 ! 11/6 1088.14 ! 15/8 t 1108.1 ! maj 7th, large pyth ! 1200.00 **eta2: Here are the above pitches quantized to 140EDO, so they make more consistent step-sizes. Not sure if there are any audible advantages, have to try it: 0., 94.286, 111.429, 137.143, 180., 205.714, 231.429, 265.714, 300., 317.143, 385.714, 402.857, 420., 437.143, 497.143, 548.571, 591.429, 608.571, 702.857, 762.857, 780., 814.286, 840., 882.857, 900., 934.286, 968.571, 994.286, 1045.714, 1088.571, 1105.714 and in "relative srutis": 11, 2, 3, 5, 3, 3, 4, 4, 2, 8, 2, 2, 2, 7, 6, 5, 2, 11, 7, 2, 4, 3, 5, 2, 4, 4, 3, 6, 5, 2, 11 nah, 140edo version is too screwed up to fix easily. Last edited by calebprime; 28th February 2013 at 07:24 AM.
 28th February 2013, 01:37 PM #60 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Added on more pitch to tuning: 9/5 -- which I couldn't seem to live without -- bringing the total to 32. Quantizing to 140EDO seems, actually, to be ok, but I'm not sure if it sounds better yet. (I arrived at it by adding 53EDO and 87EDO, two fairly well-known tunings.) 32 pitches per 2/1 is different enough from my previous scale of 36 pitches per 2/1 that it was worth the effort, and different enough that when I practice it, the patterns won't get too confused with the old scale patterns -- always a problem if you try to practice with more than one tuning system. It's like learning a language. At some point it becomes second nature.
 1st March 2013, 04:26 AM #61 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Nah, not happy with it. Too little /11 and /13 and 13/ available. There's one radical solution that I've never tried. Could be awesome. A tuning that's designed purely for keyboard and computer realization. It would have far fewer pitches in the lowest register, then more in higher registers per 2/1. (Like the overtone series, but not the same as the overtone series.) The problem is that a full 13-limit system has too many pitches to fit into an 88-note keyboard -- you end up with less than two full octaves. The opportunity, however, arises from the fact that you might not need 11/8 or 13/8 in the lowest registers -- for me, they are pitches that only sound good in a higher register. So there's waste in most tunings. That's an opportunity. The other opportunity is provided by the fact of the moveable tuning-base. So all pitches are available in theory. What's in the scale on the keyboard is what you need to grab quickly and at the same time. This would be the most efficient use of keyboard space, but also the ugliest and hardest-to-memorize pattern -- a full 88-note scale with no repeating patterns at all. (The practical solution is simply to label every key by writing on a strip of tape right above the keys.) A true monstrosity, and well outside the kind of temperaments that the highly competent folks in the microtonal community are designing. Scale to follow. To be called Pyramid Scheme, or something.
 1st March 2013, 07:55 AM #62 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Here's the basic idea. Three tiers. I present: The Güber Mensch, or Cal's Inverted Pyramid bass -- 16-tone subset mid -- 29-tone subset treble -- 42-tone set plus last 2/1 = 43 ---------------------------- =88 keys Tones are derived from 13-limit JI, with a 19/16 stuck in there, and then tempered to something like 111edo, which gives a very good approximation of everything. Sadly, there are trade-offs even with an eff'd up scheme like this. Bass 1/1, 16/15, 12/11, 9/8, 8/7, 6/5, 16/13, 5/4, 4/3, 10/7, 16/11, 3/2, 8/5, 5/3, 16/9, 15/8 Mid 1/1, 16/15, 12/11, 10/9, 9/8, 8/7, 7/6, 6/5, 16/13, 5/4, 9/7, 4/3, 18/13, 7/5, 10/7, 16/11, 3/2, 20/13, 14/9, 8/5, 18/11, 5/3, 12/7, 7/4, 16/9, 9/5, 20/11, 24/13, 15/8 Treble 1/1, 16/15, 14/13, 13/12, 12/11, 11/10, 10/9, 9/8, 8/7, 7/6, 19/16, 6/5, 11/9, 16/13, 5/4, 81/64, 14/11, 4/3, 11/8, 18/13, 7/5, 45/32, 10/7, 16/11, 3/2, 20/13, 14/9, 11/7, 8/5, 13/8, 18/11, 5/3, 27/16, 12/7, 7/4, 16/9, 9/5, 20/11, 11/6, 24/13, 13/7, 15/8, 2/1 Last edited by calebprime; 1st March 2013 at 07:58 AM.
 1st March 2013, 08:45 AM #63 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 GuberMensch Tuning in Scala Format -- 88keys !GuberMensch Tuning ! 3-tier tuning from 13L-JI with 16tones in bass, 29 in mid, and 42 tones in treble ! Look Upon my Peas You Mighty, and Eat! Goodness, How Delicious! 88 ! ! 1/1 16/15 12/11 9/8 8/7 6/5 16/13 5/4 4/3 10/7 16/11 3/2 8/5 5/3 16/9 15/8 ! 2/1 32/15 24/11 20/9 9/4 16/7 7/3 12/5 32/13 5/2 18/7 8/3 36/13 14/5 20/7 32/11 6/2 40/13 28/9 16/5 36/11 10/3 24/7 7/2 32/9 18/5 40/11 48/13 15/4 ! 4/1 64/15 56/13 52/12 48/11 22/5 40/9 9/2 32/7 14/3 19/4 24/5 44/9 64/13 5/1 81/16 56/11 16/3 11/2 72/13 28/5 45/8 40/7 64/11 6/1 80/13 56/9 44/7 32/5 52/8 72/11 20/3 27/4 48/7 7/1 64/9 36/5 80/11 22/3 96/13 52/7 15/2 8/1 3600.00 Last edited by calebprime; 1st March 2013 at 08:57 AM. Reason: 11/2 !???
 1st March 2013, 09:02 AM #64 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 GuberMensch Tuning quantized to 111EDO, Scala Format ! GuberMensch 16+29+42 111EDO 88 ! 108.10800 151.35100 205.40500 227.02700 313.51400 356.75700 389.18900 497.29700 616.21600 648.64900 702.70300 810.81100 886.48600 994.59500 1091.89200 1200.00000 1308.10800 1351.35100 1383.78400 1405.40500 1427.02700 1470.27000 1513.51400 1556.75700 1589.18900 1632.43200 1697.29700 1762.16200 1783.78400 1816.21600 1848.64900 1902.70300 1945.94600 1967.56800 2010.81100 2054.05400 2086.48600 2129.73000 2172.97300 2194.59500 2216.21600 2237.83800 2259.45900 2291.89200 2400.00000 2508.10800 2529.73000 2540.54100 2551.35100 2562.16200 2583.78400 2605.40500 2627.02700 2670.27000 2702.70300 2713.51400 2745.94600 2756.75700 2789.18900 2810.81100 2821.62200 2897.29700 2951.35100 2962.16200 2983.78400 2994.59500 3016.21600 3048.64900 3102.70300 3145.94600 3167.56800 3178.37800 3210.81100 3243.24300 3254.05400 3286.48600 3308.10800 3329.73000 3372.97300 3394.59500 3416.21600 3437.83800 3448.64900 3459.45900 3470.27000 3491.89200 3600.00000 3600.00000
 1st March 2013, 12:52 PM #65 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 I like this idea, and it's very workable. I'm testing it with a Pianoteq piano sound. Mostly, I'm very happy with what I'm hearing. It turns out to be not hard to play if the ratios are labelled on the keyboard (assuming you already know JI theory, which I do.) Unfortunately, I need a little more in the bass and mid, in order to get "5ths", (or 3/2 ratios) down there -- which means robbing a few pitches off the top. If only there were mass-produced midi controllers with velocity, etc. with more than 88 keys! None of this would be necessary! While it's tempting to just declare the process done, that would be premature closure. I'm going to be married to this scale for years if the concept works out, so it makes sense to put up with the slight aggravation of making a few adjustments, based on listening to it. I need a 12/7 in the bass, and maybe another pitch or two, or at most 3 more. I can get rid of 45/32 in the treble, for starters, maybe a few others. It's a tedious process from here on. No tuning is perfect. There are always trade-offs. Partch stopped at the 11-limit. It's not that the 13th partial doesn't sound good, or /13 ratios -- they sound great on piano if voiced correctly. It's that there's a proliferation of pitches with a 13-limit, and the whole system takes a while to tweak. Just labelling the keys takes an hour or so. Gubermensch 2.0 soon. This time, it's personal.
 2nd March 2013, 12:40 AM #66 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Here's the version I'm going with for now. Had to rob some /13 and /11 ratios in the top, add a bunch of 3/2 ratios to the bass. So the difference between the number of pitches in the bass and the number in the treble is now less pronounced. To my eye, this is one ugly scale. To my ear, it's all the good harmonic relationships that will fit in 88 keys. If certain pitches are needed but unavailable, the whole tuning bass can be moved using a table converting cents to pitch-bend info in Logic. I've tested this with simple sine waves in the high register (to be able to hear even very slow beating) and the accuracy is perfect if you use the right number. In this system, the pitch-bend range of the instruments is set to +-12 semitones. It wouldn't work for improv or real-world instruments, but it's the simplest, most accurate system for composition using Logic. (Notation is simply conventional notation with ratios written above the pitches, and the issue of pretty notation doesn't interest me in any case.) A few days of testing, and I marry this scale, probably in some cheap chapel in Las Vegas, with a preacher dressed like Elvis. It's strange how the desire for elegance on paper, and the desire to get with the program (i.e., use one of the tunings people in the tuning community have named) has prevented me from arriving at this solution. She ain't a looker -- this tuning -- but she do the job. bass: 1/1, 16/15, 12/11, 9/8, 8/7, 32/27, 6/5, 16/13, 5/4, 4/3, 10/7, 16/11, 3/2, 8/5, 5/3, 27/16, 12/7, 16/9, 24/13, 15/8, mid: 2/1, 32/15, 24/11, 20/9, 9/4, 16/7, 7/3, 64/27, 12/5, 32/13, 5/2, 18/7, 8/3, 36/13, 14/5, 20/7, 32/11, 3/1, 40/13, 28/9, 16/5, 36/11, 10/3, 27/8, 24/7, 7/2, 32/9, 18/5, 40/11, 48/13, 15/4, treble: 4/1, 64/15, 56/13, 13/3, 48/11, 22/5, 40/9, 9/2, 32/7, 14/3, 128/27, 24/5, 44/9, 5/1, 81/16, 56/11, 16/3, 11/2, 28/5, 40/7, 6/1, 80/13, 44/7, 32/5, 13/2, 72/11, 20/3, 27/4, 48/7, 7/1, 64/9, 36/5, 80/11, 22/3, 96/13, 15/2, 8/1 Last edited by calebprime; 2nd March 2013 at 12:51 AM.
 3rd March 2013, 03:02 AM #67 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 This might be final version. I'm cautiously optimistic. I made some small tweaks, but decided to preserve the 81/16 to 14/11 difference, also substituted 16/13 for 11/9 in the top register. Here's a version quantized to 111edo (111=3x37) This is a working Scala file that I'm trying out with PianoTeq: ! GuberMensch 2.8 111edo 87 ! 108.10800 151.35100 205.40500 227.02700 291.89200 313.51400 356.75700 389.18900 497.29700 616.21600 648.64900 702.70300 810.81100 886.48600 908.10800 929.73000 994.59500 1059.45900 1091.89200 1200.00000 1308.10800 1351.35100 1383.78400 1405.40500 1427.02700 1470.27000 1491.89200 1513.51400 1556.75700 1589.18900 1632.43200 1697.29700 1762.16200 1783.78400 1816.21600 1848.64900 1902.70300 1945.94600 2010.81100 2054.05400 2086.48600 2108.10800 2129.73000 2172.97300 2194.59500 2216.21600 2237.83800 2248.64900 2259.45900 2291.89200 2400.00000 2508.10800 2529.73000 2540.54100 2551.35100 2562.16200 2583.78400 2605.40500 2627.02700 2670.27000 2691.89200 2713.51400 2756.75700 2789.18900 2810.81100 2821.62200 2897.29700 2951.35100 2983.78400 3016.21600 3102.70300 3145.94600 3178.37800 3210.81100 3243.24300 3254.05400 3286.48600 3308.10800 3329.73000 3372.97300 3394.59500 3416.21600 3437.83800 3448.64900 3459.45900 3491.89200 3600.00000
 3rd March 2013, 06:24 AM #68 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 And here, finally, is a version that at least has no more weirdness than the basic idea requires. A few more fixes, and quantized to 87EDO. Viable Scala file format: ! GuberMensch 3.0 87#4 9/7 87 ! 110.34500 151.72400 206.89700 234.48300 289.65500 317.24100 358.62100 386.20700 496.55200 620.69000 648.27600 703.44800 813.79300 882.75900 910.34500 937.93100 993.10300 1062.06900 1089.65500 1200.00000 1310.34500 1351.72400 1379.31000 1406.89700 1434.48300 1462.06900 1489.65500 1517.24100 1558.62100 1586.20700 1641.37900 1696.55200 1765.51700 1779.31000 1820.69000 1848.27600 1903.44800 1944.82800 2013.79300 2055.17200 2082.75900 2110.34500 2137.93100 2165.51700 2193.10300 2220.69000 2234.48300 2248.27600 2262.06900 2289.65500 2400.00000 2510.34500 2524.13800 2537.93100 2551.72400 2565.51700 2579.31000 2606.89700 2634.48300 2662.06900 2689.65500 2717.24100 2758.62100 2786.20700 2813.79300 2835.08 2896.55200 2951.72400 2963.38 2979.31000 3020.69000 3103.44800 3144.82800 3213.79300 3241.37900 3255.17200 3282.75900 3310.34500 3337.93100 3365.51700 3393.10300 3420.69000 3434.48300 3448.27600 3462.06900 3489.65500 3600.00000
 4th March 2013, 04:06 AM #69 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Had to tweak one note of the above tuning, to have a 16/11 up in the high register. Done. Time to start making some music. I'm now married to an 87edo approximation of a sort of 13-limit Just Intonation scale, with the world's most horrible keyboard layout. There seems to be no other way. Time to start making some frickin' music. Enough frickin' Fokker bloody blocks. Enough commas. http://en.wikipedia.org/wiki/Fokker_periodicity_blocks
 6th March 2013, 04:22 AM #70 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 One more post about the tuning/scale to show how I've nearly come full circle. I think this is the framework I'm going with. Two keyboards with 88 keys, 4 octaves to hand, total, with an upper 4/1 on both keyboards. 43 pitches of 87 edo. Same number of pitches as Harry Partch, but I squeeze in the 13-limit by leaving out some pitches he included. A pair of "minor seconds" or major 7ths -- one at 15/8, the other at 82.75 cents to make a decent 4th (4/3) with 7/5, or 5th (3/2) with 10/7. Each key labeled with the closest ratio, and I'm going to add labels with number of steps ("srutis") in 87 edo, until the two numbers become synonymous in my mind. Heavy reliance on a pair of sustain pedals at my feet. Pitch-bend set to +-12, and reference to a conversion chart (1 full octave of pitch bend = 8192 steps in Logic). I think I've gotten rid of all the eccentricities except the uneven fingering patterns. There always has to be a compromise, some extra difficulty. steps of 87edo difference: 6, 2, 1, 2, 1, 1, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 1, 3, 4, 1, 1, 3, 1, 1, 4, 3, 1, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2, 1, 1, 1, 2, 2, 6 steps of 87edo absolute: 0, 6, 8, 9, 11, 12, 13, 15, 17, 19, 21, 23, 25, 26, 28, 30, 32, 33, 36, 40, 41, 42, 45, 46, 47, 51, 54, 55, 57, 59, 61, 62, 64, 66, 68, 70, 72, 74, 75, 76, 77, 79, 81, 87 Scala file: ! 43 tones of 87 #3 43 ! 82.75862 110.34483 124.13793 151.72414 165.51724 179.31034 206.89655 234.48276 262.06897 289.65517 317.24138 344.82759 358.62069 386.20690 413.79310 441.37931 455.17241 496.55172 551.72414 565.51724 579.31034 620.68966 634.48276 648.27586 703.44828 744.82759 758.62069 786.20690 813.79310 841.37931 855.17241 882.75862 910.34483 937.93103 965.51724 993.10345 1020.68966 1034.48276 1048.27586 1062.06897 1089.65517 1117.24138 1200.00000 Last edited by calebprime; 6th March 2013 at 04:26 AM.
 7th March 2013, 07:31 AM #71 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... I think this is the framework I'm going with. Two keyboards with 88 keys, 4 octaves to hand, total, with an upper 4/1 on both keyboards. 43 pitches of 87 edo. Same number of pitches as Harry Partch, but I squeeze in the 13-limit by leaving out some pitches he included. ... I think I've gotten rid of all the eccentricities except the uneven fingering patterns. There always has to be a compromise, some extra difficulty. steps of 87edo difference: 6, 2, 1, 2, 1, 1, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 1, 3, 4, 1, 1, 3, 1, 1, 4, 3, 1, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2, 1, 1, 1, 2, 2, 6 steps of 87edo absolute: 0, 6, 8, 9, 11, 12, 13, 15, 17, 19, 21, 23, 25, 26, 28, 30, 32, 33, 36, 40, 41, 42, 45, 46, 47, 51, 54, 55, 57, 59, 61, 62, 64, 66, 68, 70, 72, 74, 75, 76, 77, 79, 81, 87 ... 1) basic intuition: low-number frequency-ratios can sound good. 12-tone tuning approximates only multiples of 3-ratios. 2) second intuition: harmony that moves, that changes, sounds good. Tuvan throat-singing sounds great for about 15 seconds, then you start to smell the Yak-dung, get bored. But it sure is harmonic. 3) Working out: 87 divisions of the 2/1 ratio is the lowest-numbered EDO to approximate the overtone series to prime number 13 . 4) 87 pitches too much for one human playing a keyboard to handle. 5) choose most important of those 87. 6) 43 tones are do-able, barely. 7) Harry Partch already did this, but stopped at prime 11. 8) The 43-tone "Genesis" scale of Partch is designed to have nearly consistent interval sizes. 9) The "GuberMensch Supreem" of Calebprime has inconsistent interval sizes, but gets the 13 ratios. 10) How do you master such a scale? Practice. Here are the fingering patterns to play, essentially, 8 transpositions of an 8-tone overtone series. The series is: 8,9,10,11,12,13,14,15, (16) This inconsistency is enough to make the GuberMensch lose any chance of winning any design awards in the microtonal community. We keep our poor GuberMensch child chained to the radiator in the basement, feeding it only leftover scraps and water from the same radiator. Eight different fingering patterns to play the same goddam thing! /8: 7,7,5,6,5,5,6,2 (sum=43) /9: 7,6,6,6,5,4,5,4 /10: 8,6,6,5,6,4,4,4 /11: 7,7,5,4,6,5,4,5 /12: 7,7,7,4,4,5,5,4 /13: 7,6,7,7,3,3,6,4 /14: 8,6,6,6,6,3,3,5 /15: 9,7,7,4,5,7,2,2 Who would win? GuberMensch or Genesis? Last edited by calebprime; 7th March 2013 at 07:36 AM.
 10th March 2013, 09:26 AM #72 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... Eight different fingering patterns to play the same goddam thing! /8: 7,7,5,6,5,5,6,2 (sum=43) /9: 7,6,6,6,5,4,5,4 /10: 8,6,6,5,6,4,4,4 /11: 7,7,5,4,6,5,4,5 /12: 7,7,7,4,4,5,5,4 /13: 7,6,7,7,3,3,6,4 /14: 8,6,6,6,6,3,3,5 /15: 9,7,7,4,5,7,2,2 Who would win? GuberMensch or The 43-tone Harry Partch Genesis scale? It turns out that the Genesis scale has inconsistent-enough fingering on a standard keyboard that you'd have to learn it the same way I'm having to learn my own 43-note scale. A scale with perfectly consistent fingering means that you can always find the right pitch (given good relative pitch) by ear and the right pattern of key-relationships. But Partch's 43-note scale (tuning, whatever) doesn't really do that. So, with both scales, you have to label and color-code the keys. I tried to upload a photo of my keyboard, but it will have to wait 'til I can cut the size of the .JPG file down. With 8 different colors or styles of marking, it's easy to distinguish at least 8 different basic overtone-series patterns by eye. (Memorizing all the other myriad scales will come after that first task.) Here, for anyone whose software synth can read a Scala file, is the Harry Partch Genesis scale, taken from some website (not confirmed accurate, but it looks good.) ! Partch Genesis scale 43 ! 21.506 53.273 84.47 111.731 150.637 165.004 182.404 203.91 231.174 266.871 294.135 315.641 347.408 386.313 417.51 435.08 470.78 498.045 519.551 551.318 582.512 617.488 648.68 680.449 701.955 729.219 764.916 782.492 813.687 852.592 884.359 905.865 933.129 968.826 996.091 1017.596 1034.996 1049.363 1088.269 1115.533 1146.727 1178.494 1200.00000 Last edited by calebprime; 10th March 2013 at 09:30 AM.
 11th March 2013, 04:55 PM #73 CplFerro Graduate Poster   Join Date: Jul 2005 Posts: 1,962 Dear calebprime, Thanks for posting your recent musical work. I have listened to the fugazzi one twice now, and it occurs to me to ask you about it. I'm not trained in music and have no idea why people composed what they composed. I'm afraid I'm mystified by complex instrumental music. What do Mozart's pieces mean, for instance? So, I'm wondering what your piece means, or is it a Rorschach inkblot in which we see whatever we bring to it? Cpl Ferro
 12th March 2013, 12:35 PM #75 CplFerro Graduate Poster   Join Date: Jul 2005 Posts: 1,962 Thanks for that explanation, calebprime, and thanks also for publishing your works here; I'm enjoying your Fugazzi piece. If you would, though, another question, pardon me if I'm being thick: You talked about objective technical aspects of music, and how listeners have a subjective reaction to the products of this, but I'm wondering what your intention was with it? Is your goal to achieve technical excellence, a kind of puzzle-solving, or is there an idea you are trying to express? E.g. Beethoven's pastoral expresses the glory of nature. Cpl Ferro
 12th March 2013, 12:46 PM #76 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Heh, funny, I imagined you might ask me that. I'll try to answer more tomorrow. Briefly, though, something on the beautiful/sad side of things. But it might be that my ear is attracted to some harmonies I find beautiful, which have a certain amount of tension in them. Then, can I make four voices play those lines without it turning to crap? That is, keeping the beauty of those harmonies going, while increasing the complexity, not letting the piece devolve into aimlessness? That's part of the challenge of fugue. But here the rhythm is like a cloud or like an amoeba, so entropy or aimlessness will occur if the pitches aren't right, maybe even faster than if the piece had a strong rhythm to carry it along. The process might go: Like harmonies. Fit into mold of fugue. Admire Bartok. Be myself. Then, listening, like everyone, I imagine scenarios. I'd say my inspiration is first harmonies, then textures, rhythms, forms, melodies. Beethoven's Pastoral was perhaps his most programmatic piece. Most are less obviously about something. eta: Oh, another thing: I'm a lot like Ynot trying to make a self-powered helicopter fly. It's the challenge of making some long-cherished technical idea work. But not to show off. More to make the idea sing to you, give you something back for your pains. Last edited by calebprime; 12th March 2013 at 12:51 PM.
 14th March 2013, 03:47 AM #77 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime Still looking for the next little kernel. It occurs to me that if I'm going to be mapping number series to scales with 7 elements -- such as most of the familiar scales -- I should be looking for self-similar sequences of either 7 elements or 14 elements. If my goal is the standard "every-other" permutation pattern, there turn out to be no perfect patterns of 7 elements. But there are a lot of near-perfect patterns of 14 elements, so I'll be looking at these over the next few days. The actual series -- the results -- are here called "x-swapped". The series, from 0 to 13, will be mapped onto 2 octaves of a 7-note scale. Here are what seem to be the only four candidates, each with its quirks. user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 1 2 3 4 5 6 7 8 9 10 11 12 13 0 orbit: 1 0 13 11 7 12 9 3 4 6 10 5 8 2 X-swapped: 1 0 13 7 8 11 9 4 12 6 10 3 5 2 intervals: 13 13 8 1 3 12 9 8 8 4 7 2 11 13 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 3 4 5 6 7 8 9 10 11 12 13 0 1 2 orbit: 3 0 9 12 5 4 2 13 7 8 10 1 11 6 X-swapped: 1 11 6 0 5 4 13 8 9 2 10 12 3 7 intervals: 10 9 8 5 13 9 9 1 7 8 2 5 4 8 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 11 12 13 0 1 2 3 4 5 6 7 8 9 10 orbit: 11 0 6 5 3 12 2 10 13 4 1 8 9 7 X-swapped: 1 10 6 4 9 3 2 13 11 12 7 0 5 8 intervals: 9 10 12 5 8 13 11 12 1 9 7 5 3 7 user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 13 0 1 2 3 4 5 6 7 8 9 10 11 12 orbit: 13 0 2 6 1 4 10 9 7 3 8 5 12 11 X-swapped: 1 4 2 9 5 11 3 8 10 7 6 13 12 0 intervals: 3 12 7 10 6 6 5 2 11 13 7 13 2 1 Suppose we take the last one, and map this to A major for the moment, just to see how it works. 0=A1, 1=B1, 2=C#2, 3=D2...etc. to 6=G#2, 7=A2, 8=B2, 9=C#3...13=G#3 || 14 would equal A3 We get the sequence of pitches B1,E2,C#2,C#3,F#2,E3,D2,B2,D3,A2,G#2,G#3,F#3,A1 (repeat) and embedded in this, starting on the third note, is the sequence at "every other note" C#2,F#2,D2,D3,G#2,F#3,E2,C#3,E3,B2,A2,(G#3,A1(or adjusted) Which is identical to the first sequence, up a diatonic step -- except for the last two notes. This process of embedding can be carried out as far as one would like, with any scales of 7 notes, not just major scales. I like music where the scales change, because of my jazzoid background. I'm going to try this for my next piece, unless I see some fatal flaw or get a better idea in the meantime. The 43-note scale is ok to play on two keyboards, but trying to compose a self-similar piece with it is a nightmare, because the range is so narrow. To do the beginnings of a piano sketch, I already had to have 6 tracks of piano, one in each octave! Lines were being divided between tracks. Madness. So I'm back to a slightly-modified version of my 11-limit JI scale with 36 pitches per octave. The orchestra is all ready to go. I'm back to the 14-tone 7+7 mapping idea as sketched out above. But I'm confused. I entered what I thought were exactly the same search parameters, and came up with a different answer, or so it seems in the bleary light of day. Would you like to find (1) orbits of size n, (2) orbits of size n - 1, or (3) both? 3 grid row: 0,1,2,3,4,5,6,7,8,9,10,11,12,13 user row: 1,3,5,7,9,11,13,0,2,4,6,8,10,12 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 4 5 6 7 8 9 10 11 12 13 0 1 2 3 orbit: 4 1 8 9 11 0 6 5 3 12 2 10 13 7 X-swapped: 5 1 10 8 0 7 6 13 2 3 11 4 9 12 intervals: 10 9 12 6 7 13 7 3 1 8 7 5 3 7 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 6 7 8 9 10 11 12 13 0 1 2 3 4 5 orbit: 6 1 4 10 9 7 3 8 5 12 13 0 2 11 X-swapped: 11 1 12 6 2 8 0 5 7 4 3 13 9 10 intervals: 4 11 8 10 6 6 5 2 11 13 10 10 1 1 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 8 9 10 11 12 13 0 1 2 3 4 5 6 7 orbit: 8 1 0 13 11 7 12 9 3 4 6 10 5 2 X-swapped: 2 1 13 8 9 12 10 5 0 7 11 4 6 3 intervals: 13 12 9 1 3 12 9 9 7 4 7 2 11 13 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 10 11 12 13 0 1 2 3 4 5 6 7 8 9 orbit: 10 1 11 3 0 9 12 5 4 2 13 7 8 6 X-swapped: 4 1 9 3 8 7 13 11 12 5 0 2 6 10 intervals: 11 8 8 5 13 6 12 1 7 9 2 4 4 8 The hi-lighted one seems ok to base a piece on. Is this really different from what I posted above/before, and why? More coffee is needed.
 14th March 2013, 05:32 AM #78 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Originally Posted by calebprime ... But I'm confused. I entered what I thought were exactly the same search parameters, and came up with a different answer... Would you like to find (1) orbits of size n, (2) orbits of size n - 1, or (3) both? 3 grid row: 0,1,2,3,4,5,6,7,8,9,10,11,12,13 user row: 1,3,5,7,9,11,13,0,2,4,6,8,10,12 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 4 5 6 7 8 9 10 11 12 13 0 1 2 3 orbit: 4 1 8 9 11 0 6 5 3 12 2 10 13 7 X-swapped: 5 1 10 8 0 7 6 13 2 3 11 4 9 12 intervals: 10 9 12 6 7 13 7 3 1 8 7 5 3 7 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 6 7 8 9 10 11 12 13 0 1 2 3 4 5 orbit: 6 1 4 10 9 7 3 8 5 12 13 0 2 11 X-swapped: 11 1 12 6 2 8 0 5 7 4 3 13 9 10 intervals: 4 11 8 10 6 6 5 2 11 13 10 10 1 1 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 8 9 10 11 12 13 0 1 2 3 4 5 6 7 orbit: 8 1 0 13 11 7 12 9 3 4 6 10 5 2 X-swapped: 2 1 13 8 9 12 10 5 0 7 11 4 6 3 intervals: 13 12 9 1 3 12 9 9 7 4 7 2 11 13 user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 10 11 12 13 0 1 2 3 4 5 6 7 8 9 orbit: 10 1 11 3 0 9 12 5 4 2 13 7 8 6 X-swapped: 4 1 9 3 8 7 13 11 12 5 0 2 6 10 intervals: 11 8 8 5 13 6 12 1 7 9 2 4 4 8 The hi-lighted one seems ok to base a piece on. Is this really different from what I posted above/before, and why? More coffee is needed. The user row you entered was trivially different, and should have produced the same results. Hence* the confusion. 0,2,4,6,8,10,12,1,3,5,7,9,11,13 is only a rotation of 1,3,5,7,9,11,13,0,2,4,6,8,12. So, different paramater entered but it should produce the same results. I have to understand why this is. Otherwise I'm just chasing my tail. Let's look at an example: These two should produce the same results. user row: 1 3 5 7 9 11 13 0 2 4 6 8 10 12 grid row: 8 9 10 11 12 13 0 1 2 3 4 5 6 7 orbit: 8 1 0 13 11 7 12 9 3 4 6 10 5 2 X-swapped: 2 1 13 8 9 12 10 5 0 7 11 4 6 3 intervals: 13 12 9 1 3 12 9 9 7 4 7 2 11 13 and user row: 0 2 4 6 8 10 12 1 3 5 7 9 11 13 grid row: 1 2 3 4 5 6 7 8 9 10 11 12 13 0 orbit: 1 0 13 11 7 12 9 3 4 6 10 5 8 2 X-swapped: 1 0 13 7 8 11 9 4 12 6 10 3 5 2 intervals: 13 13 8 1 3 12 9 8 8 4 7 2 11 13 with the eat-space forum formatting, these can't be aligned easily, but they should produce the same results, and they're not. Software bug or interesting real result that is an opportunity? Answer: When we designed this, we decided -- in cases of n-1 orbits, where one set of numbers matches and the rest orbit -- to stick the matching number at the end, as you would in the standard way of notating a one-line permutation. So depending on where the orbit starts, different results will be produced in n-1 cases, because it's as if one note can move around in the series. This is an opportunity, but I still don't know whether it's an opportunity because I'm not that hep, or a real mathematical principle. It means, I think, that in n-1 cases with 14 elements, there will be 14 slightly different variants of each basic orbit. (?) It's an opportunity to move the non-matching part of the self-similarity to the end of the series, where it's least audible. (?) But, to my surprise, each result is different, not merely a rotation of the other results. (?!) * A good word. Last edited by calebprime; 14th March 2013 at 05:51 AM.
 14th March 2013, 06:23 AM #79 aggle-rithm Ardent Formulist     Join Date: Jun 2005 Location: Austin, TX Posts: 15,334 Originally Posted by CplFerro Dear calebprime, Thanks for posting your recent musical work. I have listened to the fugazzi one twice now, and it occurs to me to ask you about it. I'm not trained in music and have no idea why people composed what they composed. I'm afraid I'm mystified by complex instrumental music. What do Mozart's pieces mean, for instance? Cpl Ferro It just so happens that I've written a blog post on this subject recently: http://www.johnmreese.net/music/ I mainly use this blog to bounce ideas around in my head. Kind of like caleb does with his threads. __________________ To understand recursion, you must first understand recursion. Woo's razor: Never attribute to stupidity that which can be adequately explained by aliens.
 14th March 2013, 08:22 AM #80 calebprime moleman     Join Date: Jul 2006 Posts: 12,274 Wow, I scanned that quickly, and it's a very serious attempt to make sense of things. I'll read it again. Maybe I can ask you some intelligent questions about it. But when it comes to reasoning on that level of abstraction, I don't do that well. More on that later, after I've worked out the mechanics for today.

International Skeptics Forum