With the updated version this was gone for me. Do you have the updated version? A full 10V trigger is also a lot louder, so you can attenuate it also; either by turning drive to around 3 or using 5V or less triggers.
Can you please post an example of the noise and explain what you’re doing to generate it? Also are you talking about the CF100 or the E440?
Thanks @andy-cytomic
It’s the CF100, I was pinging it with a short burst of noise through a VCA although I just realised it’s possible to patch the trigger in directly and that helps as it’s a much higher level. Also, I’ve got two of the filters in series, disabling the second one brings it down to a point where it’s barely perceptible (the first filter pings, the second one is doing standard LPF sweeps). I’m using 96kHz project rate.
Here’s a short recording: https://on.soundcloud.com/PqSch
Also, here’s the patch: Cytomic Filter Pings 2.vcv (7.2 KB)
Cheers,
Ed
Another fun trick: FM pinging … even in hardware it’s not something that works on all filters!
- Ping the filter by feeding a trigger at the input, with res on the edge of self oscillation; it will sound woody / water drop
- use a sine VCO to modulate the filter, play with the two freq knobs, it will turn into metal /bell sounds, with a smooth decay, no eneveloppe needed
9 Likes
Thanks for the patch but there are a bunch of modules I don’t have, and it’s very unclear as to what the source of the noise is since there are many effect modules as well, each of which could be adding noise. Also please note that noise sources modulating the filter frequency will “bleed” through to the audio output - this is what happens in real analog circuits as well, so if that’s not there then the “analog model” isn’t accurate.
Can you please create a mono patch (not polyphonic) with just the stock VCV LFOs for modulation, the Clocked module is fine for gate, and the CF100 x 2 for the ping sound? Please do not use any noise source at all, no randomness, just fixed repeating patters as we are trying isolate noise here, so adding noise is counter productive.
I’ve attached a starting point, I do have the clocked module, and kept the rest of the settings from your project as much as I could see from the partial load, but stripped it back to mono and no effects:
cytomic-filter-pinging-mono.vcv (1.8 KB)
here is a more definitive demo of the patch:
a slow paced pinging garden 
Thanks again for the fantastic sounds of the CF100 Enjoy!
1 Like
Thanks Andy,
Here’s a simpler patch: Cytomic Filter Pings - Simplified Version.vcv (2.4 KB)
It’s absolutely fine on the monophonic version. I think what’s happening is each channel has a tiny amount of noise, but using 16 channels of polyphony is multiplying it, and using two filters is 32x the noise, which brings it up to a noticeable level. The solution is to stick it through a VCA, using the same gate signal that’s generating the ping, and a percussive envelope. Obviously filter pings don’t need a VCA, but it acts like a noise gate and shuts off the background signal when the ping isn’t present.
@pyer that’s a great trick, thanks. I may ‘borrow’ that idea!
Cheers,
Ed
I’m always adding noise in various ways to patches. If the CF100 adds noise because of circuit emulation, that’s genuine fake circuit noise, & I’ll use it with pleasure!
2 Likes
Yep, welcome to analog. I add a tiny bit of noise post input drive to allow for self oscillation without audio input. Uncorrelated doesn’t multiply like you’ve said, so if you add 32 lots of noise you don’t get x32 the noise. You would if you added exactly the same noise sample 32 times, since each and every sample would just add together. When the noise is uncorrelated sometimes it adds, sometimes it cancels etc, so you end up with only a fractional increase in noise power per extra noise source you add, for example if you sum 4 0dBFS uncorrelated noise sources you only get 6dBFS total, not 12dBFs.
sure, a lot of us understand arithmetic, but why is this guy hearing a lot of noise? Maybe if they can’t make a simple patch that shows it they can show it on a spectrum analyzer?
You have the story wrong. The CF100 is super clean. The CF100 adds a small amount of hiss for self oscillation, which if amplified lots with deliberate gain staging will become more audible, just like using analog equipment. Would you like me to post a sample of my Moog One without a gate on the end?
Here is the CF100 with its self noise, at 0, 5, and 10 resonance, you can already see the noise getting boosted by the noise at 5, which then becomes self oscillation at 7, the scale is down to -120 dB:
If I boost the output +24dB then I see the noise I add at the -96dB line, I’ve added the noise at -120 dB to keep it super quiet.
Also, I’m curious. Which modules do you have in your physical Eurorack setup? Could you record the output of one of your filters directly into your soundcard and compare the hiss to the CF100 when fully open with silence going in and resonance at around 90% (of self osc) at 1khz cutoff?
I don’t have any Eurorack hardware, hence the user name! 
I do have a few old hardware synths though, a Juno-6, Novation BassStation and Elektron SidStation, which is really noisy!
Thanks for clearing up my total misunderstanding on noise levels, reading that article now. The noise certainly isn’t a problem, just seemed really audible in that one patch which I’ve now fixed with a VCA. It’s worth it for self oscillation, which is great.
Adding more “good” noise is actually an area I’d really like to investigate further, since it’s just not the same as adding some noise to the input alone. Every component in a circuit and every bit of energy circulating in the circuit has a little bit of noise in analog, and it sounds great. It’s nice to be able to make it optional, but it would be great to have it accurately modelled as well. Doing accurate noise models unfortunately would take around x4 the current CPU usage, but working out some lower cpu methods that get most of the way is something I’d love to sort out.
4 Likes
That’s great to hear you understand the situation and have something you are happy with. Just like in analog you need a gate for some patches where the gain staging, modulation, and sheer number of noise sources being adding together makes the noise more apparent.
1 Like
"Adding more “good” noise is actually an area I’d really like to investigate further, since it’s just not the same as adding some noise to the input alone.
It’s fascinating. Jeff Mills kept a channel on his mixer unconnected with the gain cranked up. He’d bring it up in the mix until it added presence without being overt. That bit of noise interacts with every other sound in the mix.
Yeah, this is where analog is king, grit and noise. This is especially interesting when you consider that the noise alters the signal at every point inside the circuit, which does subtle modulation of the every part of the circuit, which all gets lumped together in what we finally hear.
I agree that noise makes the difference in analog.
It is well-established that many fail to discern the difference between an analog device and a good analog model on a blind listening test, but there is a cumulative effect in an analog setup that does make a (significant and preferable, IMO) difference.
Most people’s hardware setups are combined analog/digital and I sometimes wonder whether all those individual ADCs/DACs also contribute to the general noise in the system.
There is ‘bad’ noise though, especially digital noise (I’m looking at you, Mimeophon), and some people spend their lives trying to remove it. Make Noise are certainly aware of the complaints and their new case has 4 separate power zones to try to allow people to better isolate their digital modules from their analog ones.
Using spring reverbs is a particularly fine balancing act of trying to control noise in hardware. BBDs struggle to contain the cumulative noise. pt2399 based delays (e.g Befaco, Strega) just throw in the towel and turn the noise into a feature.
There is also a problem of voltage differences between cases, for which the mooted solution is to connect the 0v rails together, removing DC offsets, via copper or aluminium bars. This is something I am considering on my own rig. I can actually feel the voltage difference between my Make Noise skiff and 4MS pods, an actual (unpleasant) tingling.
The point of my rambling post - noise is what adds the auditory magic in an analog/hardware setup, but it can be a fine line finding just how much noise you want, and trying to control that noise.
1 Like
I’m definitely with you on the pleasant/unpleasantness of different kinds of noise. E.g. vinyl crackling can be quite pleasant whereas digital aliasing is grotesquely unpleasant to my ears. I’m an odd duck, I prefer no noise at all 
1 Like
This track uses the sound made when I stacked my EMS VCS 3 Putney on top of my desktop computer. When I wasn’t playing any sound on it/through it, the spring reverb picked up mechanical noise from the computer, and combined with the bleed through from the Putney made a mysterious sound.
All the ‘melodic’ parts of this track are Roland TR727 fed through the Putney ring modulator, with the carrier oscillator frequency modified by a square wave LFO. This track is 25 years old!
6 Likes