Why analog waveforms are trembling?

Why analog waveforms are trembling and how to implement it in Reaktor?

 

Tremolo or vibrato? For tremolo modulate a vca slightly with an lfo. For vibrato modulate pitch.

 

Hello Brett! No, I'm talking about that kind of voltage instability what makes waves trembling/moving. It brings some life to sound! And actually thats exactly what makes analog sound more pleasant to me. In opposite I also compared it with static lifeless sample-based wavetable synthesis in Serum using Minimoog and Voyager waveforms (single-cycle). Such waveforms sound everytime same and lifeless. Look at oscilloscope with Monark (or Serum) osc then with real Voyager. Voyager waveform is unstable and constantly drifting.

 

Use pulse wave oscillator with slowly modulation of pulse width or (and) two oscillators which are a little bit detuned.
ciao herw

 

PWM is a big deal but not in that case. But thank you I think this problem should be solved on technical level, but not creative by using PWM because you can notice that drifting using any waveforms in analog osc. I don't know much about unison and voicing in that case.. it seems like Voyager is monophonic too.

Thanks Brett,
Probably, modulating pitch and amplitude by noise would help.

 

This is actually considered a problem with old synths you know. What you need is to look at slow waveform pitch modulation and other types of modulation of the waveform that adds movement. You should use your ears and not the oscilloscope if you want to learn what kinds of modulations you prefer. In modern synths you add the life to sounds in your programming.

 

Yeah! That works! Even 0.01 makes sense How to implement this into new Monark-based "Voyager osc"?

 

Maybe Minimoog and Voyager oscs were designed to be slightly modulated by noise to get this iconic "phat" and "full" "analogue (Moog-style) sound"?
Maybe this is question of sound design more than question of mystic analog=)

 

why analog oscs spectrum is limited at around 17 kHz with a steep cut?
and why monark has a steep limit at around 22.5 kHz?

 

This is waveform of Voyager square wave


This is waveform of Monark square wave with slightly modulated pitch by noise


Close, but so different yet. So as you can see Voyager is bit noiser than Monark and less sharp.
So I have 2 questions:
1. How to achive this harmonic content (this high frequency wobbles/noise)? The Voyager's waveform has nonlinear edges, but Monark's has flat linear edges. Or maybe it's just amp noise?
2. And how to make Monark less sharper/brighter (I should use FIR filter or some other way?)?

 

That scope trace from the voyager doesn't look right to me - a lot of noise from somewhere. I've seen plenty of scope traces from analog oscillators, and none of them look that noisy. It's interesting that the obvious noise in that pic is high frequency noise, it doesn't indicate how much frequency variation there is.

Another point to note is that there have been comments from owners of both Minimoog and Voyager that the Voyager can sound duller at the top end (maybe something to do with not being able to fully open the filter), so the fact that the Monark is brighter/sharper is a good sign.

It's also important to remember that Monark and the Monark blocks already have 'analog' pitch drift/variation built in, so it shouldn't be necessary to add more. Another interesting aspect is that Monark seems to have carefully modelled per note tuning variation, I suppose this is to simulate tuning idiosyncrasies of Minimoog's analog voltage to pitch keyboard. I suspect that this it a big part of the character - more so than the more common random drift that everyone seems to think is important.

Personally, I don't think that adding random noise to pitch or amplitude is going to make anything sound better or more authentic. I've spent a lot of time on this, and invariably what you get if you add noise to the pitch of a weak sounding digital oscillator is a noisy weak sounding digital oscillator.

A really good analog synth can sound amazing, and there will be some noise. That doesn't mean that it's the noise that makes it sound amazing.

 

It's not a bad technique, and some folks swear by it - there are different opinions on the subject

Brainstorming wise, I've wondered for a while how much parasitic distortion/compression there is in some of these old analog circuits. The power supply just like the rest has to be as simple and cheap as possible, so while sensitive stuff like pitch circuitry is designed to gracefully handle small changes, do other sections like mixer and amp interact via parasitic feedback through power and ground lines? Does this also effect the oscillator to some degree? maybe negative feedback prevents any obvious fluctuations, but leaves a subtle modulation that is dynamic and definitely not random...
Has anyone analysed the power supply on an active minimoog? (anyone who isn't gagged by a NDA that is ). How well isolated are the power supply taps for various sections from each other - particularly while driving the thing hard ?

 

Apart from the noise I see here trembling / wobbling back and forth slightly

Adding noise source for pitch modulation osc makes that rembling / wobbling back and forth (it seems for me so). Monark seems to be not weak sounding by default=)

I don't know why but adding noise source for pitch modulation osc makes sense

Yes, Minimoog is brighter than Voyager but they both have a steep limit at around 17 kHz (but Monark at around 22.5 kHz)

What exactly makes them sound amazing? And what I should do for that?

 

So why not just use it?

That is the $64000 question.

 

Yes, it 'makes sense', then when you try it, it sounds nasty until you turn the amplitude of the noise down, then it starts to sound better. Then you tweak the level by turning it up until it just starts sounding nasty, then back it off a little. Then you A/B it with a version that has no noise modulation, and they sound exactly the same - as soon as you can hear the effect of the noise, it sounds worse!
To use noise, you need to be clever - maybe bandlimiting it?. Or using noise to control a slower 'drift' modulator etc. But Monark oscs already have that built in. AFAIR, the NI engineers modelled it on an actual Moog until they felt it was just right. So adding more probably won't make it better - more likely worse, or no different.

 

on noise modulation - it actually doesn't make that much sense in my opinion.

i mean, i get the urge and of course the same thought has occurred to me as well. but the same thought has occurred to at least 3 of us on this thread, if it were that simple it would be just a known fact at this point.

yes, there is fluctuation in electrical components that seems to add a little bit of 'life', but i'm not seeing any reason for the proposal that said fluctuations in any way resemble digital white (or pink, or whatever) noise. in fact, i very much doubt that they are randomly fluttering all over the place at 44.1khz.

 

So I can't do anything to make it sound better and more authentic?

I would like to analyse analog/digital difference and get better conception of it.

 

Yes, me too. I guess exist some kind of mathematical relationships

 

I didn't say that at all. However the answer you are looking for is not in the public domain, so you will have to do the work yourself. Lots of people have tried with varying success. Maybe you will solve it ?

Well get analysing then .

I guess that if you carefully analyse the partials for the waveform you are modelling, then you can start recreating that using a piecewise oscillator algorithm using polybleps or similar with maybe some filtering. Then feed that into some sort of waveshaping process to create the appropriate saturation/compression, you will be very close. You could then add the nasty stuff like pitch instability and noise, but I doubt that would 'improve' the sound. For a really accurate model, you could look at how the waveform changes with pitch and incorporate that into your model.

 

I fully understand your objective, but I think you should realize that getting to know a synth, soft or hardware and learn how to make the sounds you like in actual tracks on it will teach you quite a lot about the possibilities and similarities with your pick that possibly extend any waveform analysis. A first step would be not to view the analogue synths as the benchmarks and just do programming experiments.

I am not 100% satisfied with Monark in the higher registers. I feel the Cakewalk Z3TA+ 2 sounds a lot better for those types of sounds, but these 2 are not directly comparable otherwise. I like Monark for noisy basses but I'm thinking of dropping it out instead of going for it. Your objective does interest me still. Analsysis can be great as long as you don't overdo it.