Alternate tunings on the Rev2

Alternate tunings on the Rev2
« on: March 20, 2019, 01:05:52 PM »
As I find time, I hope to post more about the Rev2-specific considerations when using alternate tunings1, but since stretched tuning came up on creativespiral's Voice Component Modeling thread, I thought I'd quickly write up something more general. (As I got going, it ended up not so quick.)

Part 1: Stretched Tuning
Although "piano" stretched tuning was mentioned, there's no single standard for it2, and few good reasons to apply it to a synth3.  Instead, I have attached a sysex for the slightly and uniformly stretched scale known as 19ED3.  It stretches every semitone by about 0.1¢, so that twelfths (a span of 19 semitones, e.g. between C2 and G3) are tuned "perfect". The frequency of the 3rd harmonic of a tone corresponds to a tone perfect 12th above the fundamental, so we're aligning those. In standard 12-Tone Equal Temperament (12TET/12EDO/12ED2), octaves are perfect, but twelfths are about 2¢ flat of perfect.4

The sysex will load 19ED3 tuning into slot 6 on your Rev 2.  Enable it through Global parameters.

To really get the [relatively subtle] difference, you should compare the sound of various intervals using a very simple polyphonic patch (e.g. the "init") with the different tunings. Complex patches will likely obscure the differences.

Some Rev2 patches will need to be adjusted to work as intended with alternate tunings; others may never really work, depending on the particular tuning and patch. (Some aspects of this were a surprise to me, so more on that later.)

1Briefly: Alternate tunings are a "nice to have" feature for me, but if I had been expecting the Rev2 to be a xenharmonic workhorse, I would be posting over in the Rev2 breakup thread.

2The reason piano tuning is stretched in practice is to account for the inharmonicity of a piano's sound, caused by the stiffness of the strings and other real-world mechanical influences. Octaves are stretched to sound pure, despite the inharmonicity. The stretching is not uniform across the instrument's range, and every piano is different. Larger pianos like grands generally require less stretching to sound good, and smaller ones like consoles require more. Applying "piano-style" stretching is not really useful unless you really need your synth to be exactly in tune with one particular piano, across a wide range of its keyboard.  (There is a thing called a "Railsback curve" that is an average of observed tunings found in the wild, but it's at best a starting point for tuning a piano that gets polished by ear.  Tuning your synth to this might put you more in tune with a given piano, or it might be worse.)

3Synth waveforms are almost completely harmonic, except when we intentionally make them inharmonic with nonlinear functions like ring modulation (or the AUDIO MOD parameter of the Rev2). There may also be some tiny inadverdent inharmonic components introduced by things like intermodulation distortion in analog circuits.  If we're going to use equal nearly-octave-repeating tuning on a synth, it makes more sense to bring some harmonics in line as in 19ED3.

4As every equal-step tuning system is a compromise, getting perfect 12ths results in imperfect octaves.  Major thirds will be a little worse than they already are on standard 12-EDO, while fifths will be a little better.

shiihs

  • **
  • 103
  • phasing in and out of reality
Re: Alternate tunings on the Rev2
« Reply #1 on: March 22, 2019, 04:30:06 AM »

1Briefly: Alternate tunings are a "nice to have" feature for me, but if I had been expecting the Rev2 to be a xenharmonic workhorse, I would be posting over in the Rev2 breakup thread.

Ok, now you've got me curious. What's so awful about the tuning support of rev2? I've been having tons of fun (as well as some productive use) with it. You can literally assign any frequency to any key. Is it really unexpected that a patch where - say - osc1 is tuned to a different note than osc2 sounds wrong if you change the underlying tuning tables?
--
gear: prophet rev2 16 voice, kawai NV10, casio wk-7600, Roland Integra-7, supercollider, ardour

links:

https://www.youtube.com/stefaanhimpe
https://soundcloud.com/stefaanhimpe
https://technogems.blogspot.com
https://a-touch-of-music.blogspot.com/

Re: Alternate tunings on the Rev2
« Reply #2 on: March 22, 2019, 05:37:01 AM »
Ok, now you've got me curious. What's so awful about the tuning support of rev2?
That will be in Part 2, but let's just say that how awful depends, mostly on what particular type of tuning you're using.
Quote
Is it really unexpected that a patch where - say - osc1 is tuned to a different note than osc2 sounds wrong if you change the underlying tuning tables?
Unexpected that the DCO behavior was programmed to be so different from VCO behavior.

Re: Alternate tunings on the Rev2
« Reply #3 on: March 22, 2019, 09:06:53 AM »
Cool stuff man! I use a piano pack for Ableton that has a stretched option. I've wanted to get that vibe going on the Rev2 but haven't had the time to figure out scala. Thanks for the link!

Re: Alternate tunings on the Rev2
« Reply #4 on: March 22, 2019, 11:44:03 AM »
Cool stuff man! I use a piano pack for Ableton that has a stretched option. I've wanted to get that vibe going on the Rev2 but haven't had the time to figure out scala. Thanks for the link!

Attached an A/B example! Oh man, I'm going to leave my Rev2 on this tuning for a while, I dig it!

Re: Alternate tunings on the Rev2
« Reply #5 on: March 22, 2019, 12:45:26 PM »
Attached an A/B example! Oh man, I'm going to leave my Rev2 on this tuning for a while, I dig it!
Glad you like it!

Now, let's go in the opposite direction and squeeze the scales. I have attached two more sysex files:

  • 28ED5 (loads into slot #17): This compresses the 5/1 ratio into 28 equal steps, so that an interval of two octaves plus a major 3rd (e.g. C2 - E4) is "just".  (Each step is about 99.5¢.)
  • 34ED7 (loads into slot #16): Same idea, but more extreme, compresses the 7/1 ratio into 34 equal steps.  If you play an interval of two octaves plus a minor 7th (e.g. C2 - B♭4) it will be a harmonic 7th. (Each step is about 99.1¢.)
Both of these make the octave flat and 5th even more flat than it is in 12EDO.  They can sound really nice in widely spaced 7th chords with the 5th omitted, like G2+F4+B4 to C2+E4+B♭4 to F2+E♭4+A4 and so on.

Re: Alternate tunings on the Rev2
« Reply #6 on: March 30, 2019, 10:14:36 AM »
Thanks OakBloodThree. You made a great synthesizer even better!
28ED5 makes my Rev2 fuller and more organic sounding without exaggerating this effect.
Just can't decide if I should have Osc Slope at 0 on all patches from now on or not.

Naboo

Re: Alternate tunings on the Rev2
« Reply #7 on: February 23, 2020, 11:32:09 AM »
Thanks for sharing these!  I’m going to try them in the OB6 and AS-1 too.  Is there a way to edit the sysex file to change the slot Number it’s loaded into?

666

Re: Alternate tunings on the Rev2
« Reply #8 on: April 22, 2020, 04:33:51 AM »
Thanks for sharing these!  I’m going to try them in the OB6 and AS-1 too.  Is there a way to edit the sysex file to change the slot Number it’s loaded into?

have you managed to change the slot number?

666

Re: Alternate tunings on the Rev2
« Reply #9 on: April 22, 2020, 12:04:14 PM »
@OakBloodThree i tried it on my ob6 and p6. it's subtle but it does make a big difference. more silk, less stiffness. so worth it. thanks! this could be the cure to the most criticised thing about dsi synths and all modern analog synths for that matter.