I was surprised that the TZFM VCO had so much aliasing, but I guess that’s a common trait of the line. I’m surprised.
Is that always a bad thing? I like properly anti-aliased oscillators but it adds an interesting edge to others.
1 Like
I thought it was interesting here because these modules are (I thought) supposed to be accurate models of the analog originals. In fact the manual for tanh is just the manual for the analog module.
It’s certainly a good Q “is aliasing always bad”? In my reading I would say most people say it’s bad, a minority say you can’t really hear it, so it doesn’t matter. And then a very small minority like the sound - I guess that’s you.
myself, I’m always suspicious of a module like this, because we know that std::tanh(input * gain) is literally one line of code, whereas a wave shaper that does not alias it perhaps 1000 lines of code (or at least 100?). So I tend to suspect “they didn’t bother” when I see something like this.
If the addition of aliasing is indeed an intentional enhancement over the analog original, that’s cool. Maybe some users would like a control so that they could turn off this enhancement though, to get the classic sound.
2 Likes
My favorite trick with Vaz Modular was to downsample with antialiasing off & then high pass filter it to isolate the aliasinng noise .
Maybe VCV can find it’s “niche” as the go to place when you want a lot of aliasing?
I feel like I’ve heard this conversation before… 
3 Likes
I definitely noticed the TZFM VCO didn’t sound great. To the ear I can’t tell if that bad sound was because of the aliasing or what, but I definitively wasn’t influenced to buy hardware after hearing it.
2 Likes
In my ears, yes. But people are funny that way, some just like noise 
I’m pretty certain it’s not very good for your sense of hearing though. With a background in HIFI, and as someone who really trained their hearing, I’ve contemplated several times writing an in-depth article about audio quality in here, but I shy away from it and think it’s probably a bad idea because it seems to be such a contentious issue and I don’t need that. I shouldn’t be too surprised though, I guess, because I remember back when working in HIFI shops, that musicians definately were not the ones, in general, to go for pristine sound. They all had their idea about what sound they wanted, never mind if it was HIFI or not. I was quite surprised about that in the beginning, but I think it slowly dawned on me that what sounded “right” to them was not necessarily what sounded “good”. In other words, they were going for a particular subjective experience they desired, rather than “what’s right”, kind of like an instrument. Life’s funny.
5 Likes
A good part of them just cannot here those pristine and/or aliasing frequencys anymore. I lost a good part of my hearing above 12kHz in front of my 4 x 12’’ speaker cabinet when I was young. And a brutally loud Drum & Bass Party some 20 years ago added a nice Tinnitus to that… I would love to talk about aliasing, but I only hear it, when it is very present.
That said, the Saw and Pulse Output of Neoni does have anti-aliasing. And again: Internal Sample Rate doubled to at least 96kHz and all Instruo Modules are sounding much much better. Jason Lim did a video for Omris first Event and he used an internal sample-rate of 192kHz with 96kHz at the audio-output demonstrating the CSL. That’s also a way of anti-aliasing, but they are a “little” bit CPU-hungry that way.
2 Likes
By internal sample rate you mean running Rack’s engine higher right? If so the problem with that approach is that it quadruples (or whatever) the CPU usage not just for the modules that benefit from it (like the Instruo) but for those that don’t benefit from it at all (the other modules in your patch). If the Instruo modules benefit a lot from being run at a higher sample rate, It would be great if that over-sampling could be turned on for modules individually in their respective right-click menus.
2 Likes
I little bit of aliasing in an analog model may not be a big issue, since it’s hard to hear. But having a lot of aliasing, for me, is not pleasant. Unless, the oscillator is intentionally digital aimed to get those raw sounds.
Not having alias reduction makes the oscillator code much simpler and more efficient. Some of the anti-aliasing techniques could be expensive. But if well implemented they can be very efficient. Take as reference Vult Bleak which has literally zero-aliasing and it is fast enough.
Judging by the name TZFM it is an FM oscillator. Avoiding the aliasing when doing FM with two complex waveforms it’s challenging. In both Vessek and Noxious (which have FM) I manage to reduce it but using the spectrum analyzer some of those can still be spotted.
Some years ago I made a video about it:
7 Likes
It might not be easy to hear aliasing, but it’s much easier to hear the absence of it. You know, when a synth just sounds great.
I compare it to mp3 encoding. The best way to hear that is in your car. If you keep turning up the volume, that’s because the mp3 just does not cut through the sound of the engine (as a musician, I have, of course, an old car with a loud Diesel engine). Take a wav, and you will be able to hear much more of the music.
Same for aliasing. The synths you keep on using, just because they sound great.
But maybe I am wrong, and there’s aliasing all over the place 
Yes that is true. It’s just the FM, the Wavefolder, and the sync that don’t. But for sure it’s not accurate that none have alias reduction, clearly some do have some.
Yes. And as we saw in that excellent video that gets re-posted, running everything faster isn’t as good as internal oversampling. Basically if you have a chain and module 1 adds aliasing, and then module 2 adds aliasing to that aliasing you still get alias products all the way down into the audio range.
That said, a lot of digital filters (mine included!) have their high frequency response “warped”, and running them at higher rates can also reduce this. It’s a very expensive technique, CPU wise, however.
I didn’t know about aliasing a few years ago, I was not conciously hearing it, but I kept wondering why some of my music stressed me out and fatigued my ears much more than other music with similarly agressive or distorted sounds, there was something sharp in it that felt unhealthy… When I finally learned about it I began to hear it for what it is and it became clear to me that I should be aware of it and try to avoid it (at least in most cases). Since then I try to make sure to always use high quality plugins, especially for distortion.
I haven’t noticed it in Neoni yet but now I have to be cautious with it, and I really like this oscillator, so I really hope that if it needs some fixing it will get it.
I think in most scenarios aliasing will not be conciously identified, but it will still disturb the experience of the music.
1 Like
Yes, it will just turn up as listening fatique, headaches, numbing to high frequencies or muddled sound. But it’s also a case of once you’ve really heard it you can’t un-hear it, and now you’ve trained your ears to correctly detect that specific kind of distortion. And a visual aid such as a spectrum analyzer can greatly aid in hearing it consciously the first times, because now you have it through two senses, and they can correlate the experience, it’s very effective. But you’ll lose your “viginity” so watch out
1 Like
it really need good ears to balance the sound between aliasing and the way it s reduced, i can tell in many plugins oversampling technics sound like crap and i would rather get more aliasing insteed when it s low enought
I’m glad I heard right… the oscillators sound very good and now I looked at them and they seem to be anti-aliased very well. I run everything at 48kHz and they only show aliasing above 20kHz
The tanh Waveshaper is a problem though 
So that one should definitely get some good anti-aliasing. Maybe you can help them out with that, inspector Squinky? The Oscillators are just innocent bystanders and all of Instruo shouldn’t be accused with false accusations because of some black sheep.