At the risk of sounding like an idiot, I’m not sure how to make a judgment when the two sounds I am hearing are so different, and I feel I am hearing the constant noise both left and centre (an illusion?), but the sweep only on the right. I prefer the constant noise FWIW, but it’s a bit like apples and oranges.
1 Like
I have the same… it is hard for me to compare when the two are mixed in one stereo track.
This makes sense, polyphony as implemented in rack sometimes makes me stupid. Perfect example patch!
I just listened on headphones and took one side off then the other. Not much in it, but the right side seemed to have slightly more high frequency content and sounded somehow clearer and more transparent, if that’s possible with noise. I mean they both sound like white noise, so it’s a subtle difference. I wondered for a second if it’s a trick question and they’re both the same! 
Yes, you’re right, but mixed together, both sides on, I only hear the sweep on the right side, and constant noise on the left (and vice versa if I swap the headphone sides). Strange. My hearing maybe.
If you load the mp3 into a DAW or similar you can pan the left to mono and right to mono if you want, of perhaps do that inside VCV Rack and have an LFO swap between the two, I thought it’s just easier having both in a single file with the same envelope applied to both.
VCV Rack handles polyphony by any source being able to generate it, which is very powerful, but a bit “implicit” in that you have to know what you’re doing first since there isn’t a “unison knob” or similar to get more voices.
1 Like
They are not both the same, they have been generated differently.
The effect is so strong sans panning that I thought you were asking us to compare the constant noise with the swept noise!
There you go; I’ve made even more of an idiot of myself than usual
The right side is brighter, but is it also a couple of dB louder?
Just checking! The right side was brighter, but it’s fairly subtle.
as the others already said, it is really hard to hear some differences of the noise sounds on headphones just playing the mp3.
and then, which kind of noise sounds better, and which kind sounds worth?
imho the left channel sounds better, as it has a tiny little less bright noise, so it sounds for me a bit rounder, but maybe this is not the intended kind of good noise? how would it sound in a “real” situation (…mix, or…)?
Sorry, I can’t give a more clear answer.
Ok, thanks for the feedback. I’ll do another test where I make the sound samples mono, one after the other so people don’t have to download the file and mess around with it to get that to happen.
3 Likes
Me too
Me too
Yes please
1 Like
Ok here is the new demo:
https://cytomic.com/files/forums/cf100-noise-example-A-B-A-B.mp3
In this example I switch between the amplified noise of a Moog Matriarch (just the filter with silence at the input, and direct filter output) with the CF100 noise model at 50%. The noise model adds gaussian white noise a bit of pink noise across every resistor, and transistor, correctly scaled for thermal noise with temperature. I think it’s sounding pretty close, but can you guys tell much difference? The pattern is ABABAB, but which one is A and which one is B?
- A is the Moog Matriarch
- A is the CF100
- They are close enough it doesn’t matter
0
voters
1 Like
I took a guess but I think it’s a good idea to model noise. My intuitive feeling from using both hardware and software is that the noise makes a difference on “darker” patches where the filter is closed or closed quickly, because digital seems too flat, while a hardware synth still has some top end.
I read once on KVR or GS that Softube add noise to their modules (as well as oversampling) which is part of why they have a good reputation for sound quality.
Edit: my previous post was incorrect, there is no quantization noise, it was actually VCV Rack doing that by not preserving the floating point input of the audio being sent to it from the DAW input - this is actually a pretty major thing to be aware of, so I’m glad I found it. I’m pretty sure VCV Rack converted it to fixed point and so made it sound terrible. The Softube Module isn’t doing noise modelling, it’s just sticking some noise in at the input to start self oscillation, nothing more interesting or complicated.
Here are the updated examples of the sweeps, first the CF100 for two sweeps then the Softube Modular Prophet for two sweeps:
https://cytomic.com/files/forums/cf100-vs-softube-prophet-noise.mp3
–
Well I thought, why not check this out? I grabbed their new Prophet stuff for Softube Modular, then set a resonance of 2 and added a bpm synced triangle to sweep the filter full-ish range. I did the same for the CF100 and recorded the results in an AB style again. The CF100 is first for two sweeps then the Softube Prophet for two:
< I’ve removed the noise examples as they weren’t correct >
To me it sounds like nothing more than adding some noise at the input of their filter, which is what I currently do but at a much lower level just to kick start self oscillation, not to model analog noise! Also the Softube Prophet sounds really weird and breaks up at low cutoff - most likely some kind of quantization noise in their filter code, eww. (edit: it was VCV Rack doing the quanitzation)
3 Likes
Thanks - I think their prophet filter module is their worst vcf module effort by far (even though the vst synth is nice). I hope it doesn’t signify a drop in their standards. Their earlier doepfer, korg, moog, and rolands sound great to me. Yours are incredible and it seems will have better noise modelling. Awesome.
1 Like
Ok, my LA67 Macca Filter arrived, so I’ve now been able to match the noise levels to a real world version of the CF100, which I’ll set as the 50% level of “analog noise”. I’ll have Off, Input only, 25%, 50%, 75%, and 100% noise levels, where Input only is the current behaviour.
I’ve put a little video together showing the CF100 matched to the Macca Filter (which I’ve based my model off the SSI2140 datasheet schematic which I’m pretty sure was used) with full noise modelling using the LA67 T-ZED osc as the input. The Scope in the video shows both signals at once, and the Spectrum and the Audio hard switch between the CF100 and the Macca for constant ABAB comparison:
10 Likes
Wow that’s impressive
1 Like