It’s from a module we are working on. I think it’s mostly extreme phase jitter, with a good amount of aliasing thrown in for good measure. It’s not been easy to get rid of it.
1 Like
Mirrors the picture I saw from several VCO’s in my test. What do you call that “thump thump” artifact in the low end? Once I saw that, and started hearing it, it was very annoying.
very curious that the main aliasing element is constant in amplitude (at around -40dB)
Well, this artifact is a DC offset in the output waveform, so by itself it won’t “thump thump”, but if you apply a fast amplitude envelope it will make a very, very low-level thump. Is that what you are talking about? That DC offset comes from the MinBLEP implementation in VCV SDK when applied to a sawtooth wave which in turn (I believe) comes from the Befaco Even VCO. So you will find it in EvenVCO and Fundamental VCO (mk II, as I call them, since they totally changed at some point), my own EV3, and this new thing, and of course any other VCO that uses the MinBLEP in the SDK.
Note that in these VCOs you get the DC offset with sawtooth, but not with square.
I logged a bug for this about a week ago: https://github.com/VCVRack/Rack/issues/1816. I don’t have enough experience to know if this is just a bug, or if it’s some kind of necessary tradeoff. I suspect since it does not happen with square that the MinBLEP itself is fine, and the DC is due to that fact the sharp transition in the sawtooth is a little different from a perfect band-limited step. Which I just realized writing this reply, so I must go back and notate this in my bug - it’s probably more or less a fact of life with MinBLEP on non-rectangular waveforms.
1 Like
Yes, it is truly a unique and awful artifact. Since I’m not a great DSP engineer (but pretty good, I think!) it took me a week of meditation, sketching, and prototyping to get rid of it. I had similar agony with I added sync to EvenVCO for me EV3. These things “should” be easy.
This is probably one of the reasons I post so much about aliasing. For me it can be a lot of work to get it under control.
Of course there is a happy ending to this story - I did eventually figure it out:
2 Likes
WOW !
It’s just our friend MinBLEP, so the amazing thing isn’t how clean it is now, it’s “what did you do so badly before”? But I’m not telling…
2 Likes
ahahahah 
Yup, that’s exactly it. Once you hear it it’s a quite annoying and fairly loud piece of low-frequency “mud”, that I descriptively call “thump thump” 
It’s interesting… I just went back to my little test-stand, and in the 4 simple VCO’s left standing, the good ones, ALL of them have that DC offset (the thump) on ALL the waveforms. I can make a short video if you think I’m crazy, or just send you the patch…
I believe you. Why don’t you PM me with some examples (of free ones) and I’ll tell you what I see. If any are mine I’ll try to fix them (eventually). fwiw Fundamental VCO-1 on square looks pretty good to me. At least that’s how I remember if. (My “home” computer is tucked away behind my “work” computer ATM.)