A question about Diva's sound

[Odd Fella]

I have never played, let alone analyzed a Minimoog, which the default Diva module selection seems to be based on. I noticed that there are dissonant noises when playing very high notes. Especially when you detune two oscillators (7 or 5 st apart) and play two notes at the same time. The noise itself is lower than the played notes, though. I thought, maybe aliasing?

I opened a frequency analyzer and noticed that even on a C9 there is quite some frequency content below 200Hz or so. Did the original Minimoog also have that DC content no matter what the note?

I have a little plugin (Doffset) that removes DC and when I apply it to Diva, the signal below 200Hz is gone from the frequency analyzer. But the dissonant noise is still there. So it is probably something else.

On other synths I don't hear it when I do the same things, nor do I see the DC frequencies on the analyzer.

So, is it a bug or a feature? The quality setting has no influence by the way.
And it only happens with the Ladder filter and to a lesser degree with the Bite filter. The others sound clean.

[Urs]

Harmonic distortion does not only create harmonics above the fundamental frequency, but it can also create harmonics lower than that. E.g. if you tune your oscillators a 5th apart, you also get a harmonic and octave (or two?) below the lowest oscillator frequency IIRC. Or so (need more coffee and verify empirically).

Anyhow, I do remember being surprised by what seemed to be aliasing in a Minimoog, but it turned out to be normal.

Still, what you see could be aliasing (what's that C9's fundamental frequency?), surely, I'd need presets, MIDI, audiofile and screenshot of analyser to figure that out, but not sure if I have the time to right now even if you provided it...

[EvilDragon]

C9? Are a lot of bats going to listen to your music?

[noiseresearch]

If I'm not mistaken C9 is ≈8.300Hz. It's the first overtone of the highest note on a 88 keys piano (given the 88 keys spans a range of ≈ 27Hz to ≈ 4.150 Hz).

[Odd Fella]

Harmonic distortion does not only create harmonics above the fundamental frequency, but it can also create harmonics lower than that. E.g. if you tune your oscillators a 5th apart, you also get a harmonic and octave (or two?) below the lowest oscillator frequency IIRC. Or so (need more coffee and verify empirically).

Anyhow, I do remember being surprised by what seemed to be aliasing in a Minimoog, but it turned out to be normal.

Still, what you see could be aliasing (what's that C9's fundamental frequency?), surely, I'd need presets, MIDI, audiofile and screenshot of analyser to figure that out, but not sure if I have the time to right now even if you provided it...

OK, I have attached a little audio highlighting the noise I was talking about, basically the init patch with slight changes. I think it is mostly due to the fine detuning, judging from my experiments. But the fine detuning is exactly the same for the second part of the audio, which is clean. It is the same patch, only the filter model is different.

The first screenshot shows the c9, one osc only for the ladder filter, just in order to show the low-frequency content, which is always there regardless of the note. But on lower octaves it simply does not stand out, obviously.

The second screenshot is a comparison between the ladder and the cascade filter, otherwise identical settings (2 oscillators, the second one tuned +7 st plus a little fine detuning; the note played is an f#6). The ladder filter seems to have many more high frequency peaks. Not sure what that means...

[Urs]

The frequency peaks are due to the asymmetric distortion, and thus even harmonics being added. That's also the reason for the higher low frequency contribution.

I guess, the only thing that might be aliasing are those little spikes that start around -90dB.

[Odd Fella]

The frequency peaks are due to the asymmetric distortion, and thus even harmonics being added. That's also the reason for the higher low frequency contribution.

I guess, the only thing that might be aliasing are those little spikes that start around -90dB.

Would the Minimoog sound like in the audio if it were polyphonic?

Is the distortion you mentioned deliberate or a flaw of the hardware that you emulated?

[EvilDragon]

This does not sound like aliasing.

[Urs]

Is the distortion you mentioned deliberate or a flaw of the hardware that you emulated?

It's the hardware. Not a flaw, but a property of physics. I think Monark and The Legend behave similar, and so should probably any decent emulation.

[zerocrossing]

The frequency peaks are due to the asymmetric distortion, and thus even harmonics being added. That's also the reason for the higher low frequency contribution.

I guess, the only thing that might be aliasing are those little spikes that start around -90dB.

Would the Minimoog sound like in the audio if it were polyphonic?

Is the distortion you mentioned deliberate or a flaw of the hardware that you emulated?

This is a comparison of Nina (VCO poly) Diva and an ATC-X (analog mono with Model D clone oscillators) all playing C6 at the same oscillator frequency. I left Diva on "Great" because putting it on "Devine" didn't make a difference. Having a polyphonic setup really doesn't make a difference. Well, it shouldn't. For what it's worth, Repro has no such artifacts.

NinaDivaATCSawsC62.png

[Benedict]

I am thinking you are describing Intermodulation Distortion which used to be pretty common on 70s Rock guitars etc (and John Lord's B3) but increasingly fell out of favor after the 90s when people didn't want the fun it brings (lest they have a feeling :-O).

While there is a similarity to Aliasing, this is perfectly normal in drive, esp once it gets heavy and has several notes at once. It will be more noticeable in higher notes so C9+C#9 will deffo become pretty jarring with new 'ring modulation-like' artifacts thrown all over the Clarity spectrum (as opposed to hiding in the murk - or being High Passed back).

Exactly the same happens with Subtraktor or anything put through Scream4 in Reason and it starts with that lump your pic shows in the subs.



BTW, hearing the thing would help waaaay more than pictures or let us hear along with the pictures

[Urs]

Intermodulation Distortion

Ah, that's the term

Replace "Harmonic Distortion" in what I wrote with "Intermodulation Distortion".

(my head is elsewhere... working on MSEGs atm)

[Odd Fella]

I downloaded Legend just because of that and it does not have that low-frequency content according to the analyzer.
Oddly, that frequency range also briefly increases a bit both when I hit the note and when I release it, but not while I keep it pressed.

Also, I do not hear that distortion or grinding noise when playing two or three chords on higher octaves.

I noticed another difference. Legend seems to use a kind of pickup mode for the filter cutoff when playing the same key repeatedly. Diva does not seem to do that.

[Odd Fella]

BTW, hearing the thing would help waaaay more than pictures or let us hear along with the pictures

I did provide a short audio above...

[Odd Fella]

Just for the sake of it, the two analyzer screenshots for Legend: on the left the f#6 with two oscillators, and on the right the c9 that shows the lack of the low-frequencies.