Has anybody ever heard of thr Analog 4 DCOs aliasing!?
As in, and this is understandably a controversial and complex topic, there is apparently a difference between some DCO’s, where some do alias and some do not, and where some DCOs have a digital component in the actual sound generation part of the chip.
See this kind of thing discussed around the Behringer Deepmind DCOs
Also more details discussed in this video
Any thoughts on the matter would be gratefully received!
The Analog four uses Digitally controlled analogue oscillators. They’re not capable of digital aliasing, because they’re analogue.
The source of any aliasing is likely to be coming from the effects, as they’re digital.
I’m sure that some people use the terms quite fluidly (incorrectly?), but it’s as @Fin25 says, to further clarify:
VCO : Voltage controlled oscillator, it’s all analogs baby
DCO : Digitally controlled oscillator, i.e. clocked and stabilised using bytes and bits but generated using them sweet sweet analogs
DDS / DSP etc. : Direct digital synthesiser, Digital sound processor etc. sound sources that produce sound digitally and are controlled digitally
As Fin indicates I don’t think it’s possible for a DCO to alias, as it’s not producing sound from a sample, it’s just stabilised by a digital clock - aliasing is most often associated with low bitrates or undersampling.
Disclaimer: I am not a certified synthologist
Yes, hopefully that’s the case with the DCO sound generation.
Looking again at that lengthy thread, maybe it’s just because there was so much conjecture around the topic which was suggesting that this could be the case, maybe it is in the Deepmind instance because the square wave is software generated in some manner.
From memory, that YouTube explanation does yield some interesting, and possibly enlightening detail around the matter (the DCO digital / analog debate).
Analogue oscillators can’t alias. Doesn’t matter if they are DCO or VCO. As already discussed, the analogue circuit needs control to change pitch, that can be done via voltage control, or digital control.
At a stretch of the accepted definition for sound, one could propose that OSC syncing can be a sort of aliasing effect. Regardless of being digital or voltage generated. So yeh, I vote you can alias the heck out of the A4.
Or it could be the envelopes, like in the Dreadbox Nymphes.
Let’s remember what aliasing means:
To have aliasing, you need sampling [digital recording] going on. A DCO in an analog synth isn’t doing that.
A digital effect (e.g., the onboard reverb in a synth) does sample—it converts an analog signal to numbers through an A-to-D converter at the input, pushes the numbers around to create the effected audio, then sends that to the output through a D-to-A converter, and back out into the world as an analog audio signal. If that process goes wrong in a certain way, you can end up with unintended changes to the original timbre, generated through aliasing.
So like @Fin25 sez:
If I am missing something here, I am open to correction…
There could potentially be aliasing artifacts on digital control signals, like the oscillator output itself isn’t aliasing, but the pitch control signal is, since it is digitally generated. This would only apply with audio rate modulation of pitch though, so a very very edge case situation.
While I don’t mean to add to any conjecture regarding DCOs being aliased: this also sounds like some post I read online about how the Oberheim Matrix-6 (DCOs) use a (digital) clock generator in some clever/cost-saving way to generate some part of its sound (presumbly the D part of the DCOs, or a clever way to be polyphonic with less circuitry, but I am not an electronic engineer).
While looking for that post, and not finding it, I did find this Yamaha patent from the mid-70s detailing a method of digital control over analog oscillators. Surely great reading for anybody looking for something to read on the internet today.
There’s been discussion about what source the different OB Matrix6 family synths (6, 6R, 1000) use as a starting point for their DCO process, and whether that accounts for small core sound differences among those three synths: Sequence 15: Matrix 1000 vs. Matrix 6 DCOs
The video in the original post talks about this kind of stuff, although it doesn’t use the Oberheim synths as an example.
I don’t think this has anything to do with aliasing … but @warpigs330 raised a point earlier about DCOs and aliasing that’s over my head, so maybe there are aspects here I’m missing.
a4 that i use definitely is analog and is not aliasing.
but my nord lead 2x is.
Yeah, my point only really applies for weird situations like cross modulation or FM or things like that, and I don’t think many DCO synths do those things.
The key is knowing if the modulation affects the signal directly or the digital control of the signal.
And if the modulation itself is analog, btw.
In A4, I guess AM is full analog.
Not sure about the LFOs on oscillator (pitch/frequency) or filter frequency.
But in such cases, it’s not the oscillator that is aliasing, per se, if I understand correctly.
If you have digital control in the synth there is always the chance that aliasing can make its way into the audio signal somehow. For example you can hear some aliasing here on the kijimi (which has analog oscillators) because they used low-res DACs for the VCA envelopes (the adapter demoed in the video is a filter to rectify the problem, so you can hear the difference with the adapter in use and no aliasing):
You can also have aliasing with the encoders, the position of which are linked to a digital value. A filter sweep with high resonance for instance can show this on DSI synths, iirc.
I think the Analog Four is aliased all the time: some people call it the “Analog 4”, others call it the “A4”… There might even be other aliases I’ve never heard
DCOs are a great example of why Gearspace discussions can get dragged down by people who know nothing about what they speak, over and over again. It can be just like AI!
This is absolutely not a concern for anyone even in the hypothetical.
I don’t know if that counts as “aliased” so much as an intentional design choice for tuning the filter musically, unless I’m misunderstanding the aspect you’re focusing on 
Analogue four.
Because spelling matters.
Alpha Foxquat