CV to computer

[b]<<"In your face, Fred Mundell! " >> - GordonC [/b]

LOL! ;-)

[b]YOU were absolutely right![/b] I had not, at that time, realised the simplicity of even the most complex Theremin waveforms.. My expierience was with trying to derive frequencies from things like guitar strings.. Also, even with Theremin waveforms, it does get more difficult if one takes an amplitude modulated signal... I had not realised what a huge difference having (a) Simple waveform (b) constant amplitude, would make... The fact that one could use the pre-VCA signal for determining frequency had not really dawned on me.. and this fact makes a huge difference - it even allows one to have quite severe filtering on the signal if one has a more complex waveform.. Doing this on an amplitude modulated signal is much more complex, because when the signal drops below reliable detection levels (even though the unfiltered audio is still loud enough to be heard) all sorts of nasty things happen.

[i]<< "I think you missed my point. Not "different pitches have different timbres" but "while the pitch is changing the waveform gets distorted." - GordonC >> [/i]

Not sure I absolutely get what you are saying here..

With (almost) every method of changing timbre as a function of pitch (morphing / VCF / PWM) the statement "while the pitch is changing the waveform gets distorted." is, I think, true - This happens to every cycle of the audio (certainly with analogue VC modules where there is no stepping).

One thing bothers me a bit - I am sure some mathematician would answer this and have a laugh - Take a pure sine wave, and sweep its frequency in an analogue way, so that the frequency change is occurring (albeit fractionally) during each cycle.. Lets just look at a crude example.. first 1/4 cycle has a period of 1ms, 2nd 1/4 cycle has period = 1.001ms, 3rd = 1.002ms, 4th (final) = 1.003ms... This sine wave is, by definition, being distorted by merit of its frequency changing - the shape of a complete cycle is not a pure sine(!?)..

Am I correct in believing that the degree of distortion 'imposed' upon any waveform as a function of frequency change, is proportional to the rate of that change?
Also.. Taking the sine example.. this is (when frequency is constant) without any higher frequency harmonics.. But with a complex waveform, this effect will act on each harmonic, to a lesser degree the higher the harmonic.. one would? (I wonder) get maximum distortion on the fundamental (which will produce resulting harmonics while pitch is changing) and the effect will be most noticable at lower frequency (frequency change vs waveform period)..

All without any extra processing.... But probably quite miniscule - perhaps even unnoticable... impossible to observe at audio frequencies with my equipment.. and I dont know if I can hear it.. Changing pitch does sound 'brighter' - but is this due to some other neurological function rather than hearing extra harmonics?

[i](these thoughs were inspired by some experiments I did recently.. I bought two wonderful old HP signal generators with 1Hz resolution, from 1Hz to 20MHz, extremely pure sine, and precise sweeping - I am using these to calibrate my Theremin circuits.. I can feed these generators directly into the heterodyning stage.. Anyway - I ran a (fast) sweep checking the waveforms on my analyser, and noticed a small distortion.. My analyser is not brilliant, signal below about -75db gets lost.. so I was surprised to see any distortion. When I stopped the sweep, there was no distortion (that I could see), when I slowed down the sweep, the distortion dropped out of view... which got me thinking 'bout all this again)[/i]

Fred. Sorry. Couldn't resist the "told you so." Of course, to your credit you went from there to devising a method for theremin pitch determination that is faster than zero-crossing!

Joe. Yup, the processor isn't the bottleneck. Nonetheless, the fastest way is not to determine the pitch at all! Hard Sync offers some interesting possibilities. Oddly enough there does not seem to be a lot of info about it on the 'net - unlike the other stalwart of 1980's electronica, FM Synthesis, but it's worth looking at.

(Quick rundown on Hard Sync. You have a fixed frequency oscillator that can be restarted via a trigger. Every time the input frequency is triggered the oscillator is restarted.

Features - the timbre as well as the pitch changes in relation to the input frequency. There is a distinct and sudden change around the point where the input frequency equals the fixed frequency dividing the pitch field into two distinct areas, like a split keyboard on a synth. Which is nice.

The amplitude decreases as the input frequency increases above the equal zone as smaller and smaller slices of the fixed waveform are generated. Not so good.

Hardware analogue devices such as the FreqBox Moogerfooger, the Syndustries Zeroscillator and the Abstract Data Synthex augment this by making the fixed oscillator voltage controlled, and treating a secondary audio input as its control voltage, giving rudimentary FM Synthesis.

I expect this is available in most software synths too, but have only actually noticed it in the the rather wonderful Jasuto for iPhone/iPod Touch (which is like a reactable in your pocket) and also implemented as a VST plugin for Mac and PC. Advert over. I should be on a commission!

Fred: [i]This sine wave is, by definition, being distorted by merit of its frequency changing - the shape of a complete cycle is not a pure sine[/i]

Yes, that's what I'm saying. You can see it happen (edit - playing catchup on concurrent postings - you [i]have[/i] seen it happen!) - record a steady low note, then zip up to a high note as fast as you can, and look at the waveform on the recording. As you might expect, it changes in shape continuously, particularly during acceleration and deceleration. To recreate that in the digital domain would need some pretty fast sampling.

It is a small effect, but I can hear it when the change is fast enough.

"Fred. Sorry. Couldn't resist the "told you so."

;-) No problem.. I enjoy it! ... Makes me feel less bad - because I am quite prone to subtly (lead brick thinly covered in lemon) imply "told you so."

I should really go back over my older posting - a lot of thing I have said, I can now see were if not wrong, I was damn close to wrong.. I ought to add some updates.. but dont have the time right now.

That would be obsessive. Lordy - if I did that with my postings it would take forever - learning never ends.

" To recreate that in the digital domain would need some pretty fast sampling." - GordonC

You are right! ... I think analogue, so missed what you were saying.. Independent cycles are the fixed sample shape..

Never dawned on me! That is HORRIBLE! - Its no wonder that digital synths always sound different to analogue ones.. Particularly when pitch bend and portamento is used.

I love the Mini-Moog portamento.. Never found anything which sounds close.. Always wondered what it was that prevented this sound from REALLY being created on a PC virtual synth.

Thanks Gordon - one less mystery in this chaotic world! ;-)

Gordon: Yeah, I know about hard syncing oscillators. (And soft syncing too!) I'm an old analog dog who used to run a studio with a Moog System 35 in it. (And my own Oberheim 4-Voice SEM modular.)

Casio made a whole digital waveshaping system out of a similar idea, the "Phase Distortion Modulation" in their CZ and VZ series synths.

I think we discussed this here before - frequency tracking has a "quantum limit" that you can't get past. You have to have 'x' number of waves be counted to determine a frequency - three is probably the absolute low limit to detect a periodic waveform. At low frequencies it causes a perceptible lag.

Maybe determining a "restart" point via syncing to a fixed "carrier" wave is the solution. Nice idea!

Cool. I'm just learning about this stuff. [i]Soft[/i] syncing?

The way I remember soft-syncing explained is that instead of resetting the slave osc to a certain phase (zero) with every cycle of the master osc, it only resets if the phase of the slave osc is within a certain pre-defined range of phase difference with the master. As a result, only *some* of the waves of the slave are reset as they cycle. At least, that's what the soft/hard sync switch on the Moog Oscillator Driver was supposed to do. (And of course, how periodically they reset varies with the pitches.) Sonically it's deeper, less bright and "brittle" as full hard syncing.