I've been fiddling with a VST synth (Charlatan) and noticed a preset sounded slightly distorted. It's two oscillators each have their own volume control, and there's a master volume control.

It seemed to me that the oscillator volumes were too high which was causing some sort of distortion. I wanted to see it could be fixed, so I decided to match levels as best I could:

1. (Audio) Lowering only Master Volume.
   
2. (Audio) Lowering only the Oscillator volumes.

So if you can hear it, the second version sounds 'clearer' while the first has some fuzz.

But I'm wondering if some mastering plugin can show objectively if the signal is distorted, without having to rely on hearing which can be subjective. Or is this basically impossible?

 

'Total Harmonic Distortion' - THD. Some tools (like Plugindoctor) can measure THD.

MathAudio's THD Meter can do it for free, setting it up is a bit tedious though (https://mathaudio.com/thd-meter.htm)

 

Yes. Use a spectrum analyzer or sonogram. Your DAW should already provide these plugins by default.

 

ATME, PD can't run synths.

I doubt that an analyzer or a spectrometer would help. I dl'ed both files, noticed an audible difference but neither of these tools helped me to see that one file has more audible distortion. The visual difference could aswell be caused by a higher RMS or something else.

 

Last edited: Jul 24, 2022

It would be tough to see the effects in a spectrum analyzer. Perhaps some high frequency boost from the distortion but that can also show from an EQ boost or exciter (air band etc).

Best way I can see the distortion is through a high res spectrogram zoomed in:



Red arrow is the point where some artifact occurs in the distorted version, which does not exist in the clean version. Seems to line up with the peak of the filter envelope, and is mostly affected by resonance amount.

 

Sounds like the signal is being clipped somewhere along the synth's internal signal flow. As resonance adds gain it's highly likely it's distorting there. I'd go with lowering the oscs volume.

 

AFAIK Spectrograms won't help much in understandig the source of the problem in harm. distortion, except in clipping, clearly visible.

THD meters don't use music or generic audio material, though it's theoretically possible knowing infinitesimally the reference.
A pure clean sine reference at freq. f is fed through the unit under test, the output is read filtering out the fundamental, the remaining partials/harmonics energy is part of the harmonic distortion.
Then f is sweeped/increased and all new energy results are root mean squared with the others, giving the THD.

For VST shoud be the same principle: provide clean sine at differet freqs, feed through VST under test, read result with VST THD meter (that I didn't even know to exist in the VST realm, I'll have a look).

 

Last edited: Jul 25, 2022

As I said, you can see a difference (btw even in an analyzer) but can you identify this as distortion just by looking at it???

 

Yes! By definition, what is visible there is distortion. The question is, what kind of distortion are we dealing with and what exactly triggers it?
And here I agree with you, the sonogram does not help us.