Periodic Noise in Distortion FX?

Hi everyone - I just received my Witch a couple of days ago, and have been trying it out. One thing I noticed is that when overdrive is the selected FX on one of the stock patches (SubTract by default, for example), there is periodic noise - not distortion - that lasts for 10 seconds or so. It sounds like sizzling. After 10 seconds the sizzling sound goes away, and then it comes back for another 10 seconds, etc. This even occurs when the level of the effect is turned all the way down using Knob E.

When I change the FX by setting CC89 to something other than overdrive (Delay in 32-63, or Chorus in 96-127, for example), the periodic sizzling goes away. Is there something that I can do to avoid this periodic noise when using overdrive?

Thanks in advance!

Hmm, weird, has anyone else noticed this?

Is the periodicity related to the LFO1 speed at all? The Overdrive also adds a modulated sample offset, or bias, which is tied to LFO1. But when the FX amount is all the way down it shouldn’t do anything.

How do you have your Witch hooked up, with USB power and headphones?

Hi there - thanks for your response!

It is a good question regarding relationship to LFO1 speed - I will check this out. In answer to your question, I did have the Witch hooked up with USB power and headphones. I did also try it with a separate power supply, but IIRC the issue was still there. I will re-confirm this, as well.

I’ve done some quick tests. Using a pair of Sennheiser HD-25 (70 Ohm) with the device volume turned all the way up I can just about hear a bit of fuzz from the overdrive, even with the FX amount turned all the way down. It coincides with the ‘LED on’ phase of LFO1. Note that at this volume, playing notes is way too loud on my headphones!

With the default output volume I can’t hear a thing, but maybe it is partly masked by my tinnitus :ear:
I wonder if the type of headphones make a difference, mine are neither fancy nor low impedance.

The noise is also visible, and measurable, when recording the digital output over USB. Audacity puts it at about -80dB. There’s some measurable noise on the Vosim patch also, and it’s a bit louder at -78dB, but the frequency content is different and it’s not at all audible to me.

Here’s a screengrab from audacity, recording three parts of ‘silence’ (dB scale):

  • Poly Sub, Overdrive, zero FX
  • Poly Sub, Overdrive, high FX (lots of DC offset)
  • Vosim, Phaser, zero FX

We can see both on the dB waveform and the spectrogram below it that some amount of high frequency content increases and decreases along with the LFO.

(note that the audio outputs on Witch are AC coupled so the low frequency DC offset will never get to the headphones)

I will do some more work on the DSP to reduce it, certainly with FX amount at zero there shouldn’t be any noise from the overdrive. Thanks for bringing this to my attention!

Adding a DC blocking filter to the Overdrive takes about 6dB off the noise, but it’s still there. It bothers me a bit because the should be complete silence, zero, when FX is off, and I don’t quite understand why that’s not the case :exploding_head:

Thank you very much for checking this out so quickly!! I was just performing my own tests when I saw your response.

Here is a link to an audio file I just created. This was made playing a single note in the default SubTract patch (non-MPE version), with the overdrive (Knob E) turned all the way down. The Witch was powered via USB connected to my Windows tablet, and the audio recorded into Renoise using audio over USB - this basically matches what I heard over the earbuds I had plugged into the tablet. You will hear the noise at the beginning of the file (I didn’t time the start correctly), but if you listen to the end you will hear the noise cycle off, back on, then off again. I believe the noise was much less when I plugged the headphones directly into the Witch. I think this confirms your findings, as well.

In addition, you can hear the noise increase in graduated steps/levels until it shuts off again - in other words, it is not at a constant level, and does not change smoothly. Maybe this would help you identify the source?

My guess is that the noise you identified may be exacerbated by noise from powering the unit via USB - I have heard USB-powered noise issues on other hardware (IK Multimedia UNO, Korg NTS-1), both myself and anecdotally on the internet.

Thanks again for your help, and please let me know if there is any other testing you would like me to do!

Noise File

Edit - please let me know if you can’t access this file for some reason. I can’t figure out how to upload a .wav file into this message.

1 Like

That certainly sounds like some DC offset in ADC inputs that doesn’t let you reach 0.