Alternate tunings and vintage emulation

[OakBloodThree]

TL;DR: Here, have a SysEx file with a slightly randomized tuning.

Although I was initially looking at the alternate tuning capabilities of the Rev2 for use in microtonal composition and performance, it seems to have generated some interest here as a way to emulate the behavior of vintage synths.

I wanted to see how control voltages are derived from keyboard input in a vintage monophonic synth, so I took a look at the schematic for the Moog/Realistic MG-1¹. See attached clip from the service manual.

This generates 32 different voltages (one for each key), ideally all uniformly spaced from zero up.  The “Hi-End Trim” sets the voltage for the highest key.  (The nominal value shown of 2.58 volts is about 31÷12, or the number of semitones spanned by the keyboard, scaled to 1 volt per octave.)

In reality, since the resistors have a specified tolerance of ±1%, there’s potentially that much deviation in the width of a semitone between any two keys. The Hi-End Trim can only correct the overall span of the keyboard, and would need to be accurate to less than 1 mV to get the span tuned within 1¢.

This is very easy to model with a randomly-generated tuning table.

I have attached a SysEx² that will load into alternate tuning slot 4.  Scale steps were randomly generated in Google Sheets so that each semitone (across a 61-key range) is 100±1¢.  I didn’t rescale the overall to duplicate the “Hi-End Trim” adjustment. The total span ended up a bit flat-- about 5¢. That will emulate an older instrument that hasn’t had its High-End Trim readjusted in a while.
¹I chose this model mainly because my family had one-- I was in high school when Radio Shack sold them and Dad got one for us. It was also relatively inexpensive, so I expected to see some corners cut in the circuit design/engineering. Yes this synth had “polyphony” (which was more properly paraphony since there was one envelope and filter section) but that was a separate circuit from what I showed.
²If you're interested, here's the head and tail of the .scl file used to generate the SysEx:

Code: [Select]

! 61-Rando 1 pct.scl
!
Randomized +/- 1% over 61-key keyboard
 60
!
 100.07755
 199.45729
 298.68053
 399.62779
 498.88812
 598.46807
 698.61219
... 46 more in this pattern ...

Code: [Select]

5393.75303
 5494.70841
 5595.09756
 5695.59127
 5795.96337
 5895.48586
 5994.74663

[guyaguy]

Nice I'll have to try this. Thanks for sharing.

I'm no EE but I believe the way the poly side of the MG-1 works is called divide-down, which is how many combo organs worked too. Correct me if I'm wrong!

[philroyjenkins]

emulate the behavior of vintage synths.

Hey bud, cool stuff going on again. I have a request if possible because a lot of this scala/micro-tuning stuff is a bit over my head (and others' I'm sure.)

I can definitely hear the differences and even the stretched piano tuning you posted goes a long way to liven up the Rev2. There is almost this high end kind of static ping to the sound in the stock tuning that turns into a nice phasey movement with the microtunings. My favorite tuning I've tried from you is currently the 28ED5. Even playing two octaves of a note just sounds so much fuller in a way traditional detuning/slop haven't achieved.

Anyways I was hoping since you have a better understanding of this you could recommend or create a sort of "definitive" tuning to make the Rev2 sound more organic and imperfect as if it were maybe a physical instrument of some kind. To me the 28ED5 has been great for this, I just wonder what you would recommend with this specific goal in mind.

[OakBloodThree]

...I believe the way the poly side of the MG-1 works is called divide-down, which is how many combo organs worked too.

You are correct on both counts, and you just gave me an idea:  To make an alternate tuning that's based on the function of a "top octave generator", which not surprisingly generates the highest octave's notes.  Lower octaves are made by dividing the frequencies down by powers of two (which digital circuits have a thing for).  The top octave generator doesn't make true equal temperament, but approximates it with rational number divisions of a 2 MHz clock.  I have attached a SysEx emulating pitches generated by one of these.  It loads into slot 9.

...could recommend or create a sort of "definitive" tuning to make the Rev2 sound more organic and imperfect as if it were maybe a physical instrument of some kind.

No short answer, but I would say it depends on which physical instrument and what kind of music.

Maybe try one of the "well" temperaments?  I've atached a SysEx for a well temperament supposedly used by the Broadwood Piano Company in the late 19th Century, right around the time when instrument builders were transitioning to equal temperament.  It loads into slot 4.

I also made a variant that's stretched by the same amount as 19ED3. It loads into slot 8. The deviations due to stretching are relatively small compared to the ones due to the temperament, but they accumulate over the full range whereas the temperament returns to equal within the cycle of an octave.

FWIW, I do consider the Rev2 (and other digital-controlled analog synths) to be a "physical instrument."  The electrons bouncing around inside the oscillators are no less real than the air molecules bouncing around in an organ pipe.  Maybe a virtual synth is less physical as it's a calculated approximation of what a naturally-oscillating thing would do.

[philroyjenkins]

I would say it depends on which physical instrument and what kind of music.

I've always been a fan of Rhodes sounds. Not sure if there is any sort of insights here as far as the way they are tuned.

Somewhat on topic, I probably spend 70% of my time on the Rev2 using the filter as a keytracked sine wave. Its natural tendency to veer off at the edges might have been what keeps me using it.

Add in a little audio out mod and its just the right amount of slightly bittersweet dissonance.

[guyaguy]

...I believe the way the poly side of the MG-1 works is called divide-down, which is how many combo organs worked too.

You are correct on both counts, and you just gave me an idea:  To make an alternate tuning that's based on the function of a "top octave generator", which not surprisingly generates the highest octave's notes.  Lower octaves are made by dividing the frequencies down by powers of two (which digital circuits have a thing for).  The top octave generator doesn't make true equal temperament, but approximates it with rational number divisions of a 2 MHz clock.  I have attached a SysEx emulating pitches generated by one of these.  It loads into slot 9.

Will try that out too. Thanks!

[guyaguy]

I would say it depends on which physical instrument and what kind of music.

I've always been a fan of Rhodes sounds. Not sure if there is any sort of insights here as far as the way they are tuned.

Somewhat on topic, I probably spend 70% of my time on the Rev2 using the filter as a keytracked sine wave. Its natural tendency to veer off at the edges might have been what keeps me using it.

Add in a little audio out mod and its just the right amount of slightly bittersweet dissonance.

A perfectly tuned Rhodes wouldn’t have any variations from note to note because each key has its own independent tone bar. In practice these are usually not tuned perfectly so are slightly off. So something like OakBloodThree’s randomizer would get the job done.

[OakBloodThree]

I would say it depends on which physical instrument and what kind of music.

I've always been a fan of Rhodes sounds. Not sure if there is any sort of insights here as far as the way they are tuned.

A perfectly tuned Rhodes wouldn’t have any variations from note to note because each key has its own independent tone bar.

I agree that the tuning standard of the Rhodes was most likely Equal Temperament.  I'm not sure of the level of precision attainable in tuning it. I would bet that the more important aspect of the sound is the inharmonicity introduced by a relatively complex vibrating system:
Only the end with the screw is held fixed.  The hammer strikes the narrow tine at the bottom, which is also coupled magnetically to an electronic pickup coil.
Here's a paper on the subject: Non Linear Behaviour in Sound Production of the Rhodes Piano
This should give you some ideas on how to make a good approximation of the Rhodes piano on your Rev2 (or other analog synth.)