Korg Forums

TrrP wrote:I'm guessing the Radias/R3 operates at 48khz meaning that any frequencies the Radias/R3 'trys' to produce above 24khz (half the operating frequency) will start aliasing. Basically the aliasing will 'fold' any frequencies above 24khz down to lower frequencies producing unwanted noises. So the actual frequencies of the aliasing are below 24khz.
Example: 25khz frequency will become 23khz. 32khz will become 16khz.
I don´t know for sure but I think osc sync on the R3/Radias will produce even more aliasing because of phasing. I also believe that FM (VPM on the Radias-R3) and cross modulation will also produce aliasing more because of phasing meaning that aliasing will start well below 24 khz.

A 20KHz lowpass filter in front of a D/A converter completely eliminates the possibility you describe, and only dogs would notice the difference, since most speakers don't produce much output in that range and we humans don't pick up anything called music above 16KHz. Phasing is separate from aliasing, and phasing doesn't influence aliasing so long as the original signal is significantly below the Nyquist frequency. For example, the phase of a 8KHz sine wave won't see any significant affect on distortion due to aliasing on the R3/Radias.

However, phasing is a much more common and expected occurance, even desireable, that is often confused for aliasing when heard distinctly and amplified by a filter with high resonance in the right range to pick up the cancellation/summed frequencies. FM, cross modulation and ring modulation intentionally cause phasing, but this has nothing to do with aliasing (which may or may not be occuring at the same time). Comb filtering just so happens to look a lot like the effect of aliasing, even though they're unrelated, the difference being that the frequencies of the filter are based on the distance and Q of the comb bands, which aren't the same as the relationship of the signals that are synced to the sample frequency (as the sample frequency is fixed, so the distortion occurs as a factor of it).

I swear that I can convince people that aliasing is occuring on a digital synth when it's only normal phasing, as I can rig the demo with a digital synth where the actual sound is coming from a completely analog Buchla 200 (not 200e) that's sitting right behind it.

I just tried to reproduce the scenario described by the OP, the Strings patch, an octave shifted up one notch, playing the 2nd and 3rd notes from the left. I only hear normal dissonance between two notes, no aliasing whatsoever. The default noise level for this preset is already 0. There are several parameters that may affect what you hear. It's using pulse waves with modulation, including LFO, several effects including the ensemble and delay effects. Because the LFO is used, the patch is also affected by the tempo. It also has a rather slow release. All-in-all, this is a terrible patch to use to determine whether aliasing is occuring or not because it is full of normal cancellation from dissonance and everything else. Of course two notes right next to each other will sound terrible and cause nasty sounding upper harmonics when the two notes get summed when transitioning between them. That is what it is supposed to do. Even two of the same note an octave apart, like C3 & C4 will have significant phase cancellation with each other on a patch like this. If you build the same patch on a hardware analog poly synth, you'll get similar dissonance. Perhaps there's something more pronounced that you are hearing due that is different than what I'm experiencing, so I can only tell if you upload a sample for me to compare.

You might want to try my BowedStrings patch that you can download from the R3 section here to compare with the Strings preset. I built this program specifically because I didn't like the Strings preset for arpeggios, and I think mine is much better.

Ok so is it just digital synths that have dissonance and they create the cancellation on its own or is this something that analog synths do as well?

creepysuitguy wrote:Ok so is it just digital synths that have dissonance and they create the cancellation on its own or is this something that analog synths do as well?

This will happen to both. it is to do with the interaction between the two notes as you make the transition.

however, we can't be sure from what you have said so far that we are hearing the same as you. it would be useful if you could provide an audio example.

The problem occurs on any polyphonic instrument, including a real piano or guitar when any two note intervals are a half/whole step apart when played at the same time. Try the first 6 notes of Chopsticks, for an example of dissonance. By the way, ambient music often uses dissonance to create mysterious/creepy moods, so you may be able to use it to your advantage.