[nobug] 4P Diode filter weird behaviour...

Ive already created this topic during beta phase. But i couldnt replicate this bug.
Now i encountered it again.
So if you wanna test it, open attached Redux Instrument and add RES. It will horribly peak and make almost no sound, adding drive will help a bit.

Ive heard something about DC offset problem so ive tried adding DC device before filter. Play with it, it makes interesting things :blink:

Activating Auto DC helps a lot (and create interesting but unwated phasing).
So some simillar function should be implemented in 4P Diode filter maybe?

The problem is with your samples.

For example:
5727 akiz-sample-bad-dc.png

… then you might have better luck.

If you’ve generated these samples yourself, then look at the process that was used to create them, and try to see where this bad DC offset/bias is being introduced in the first place.

The diode introduces dc ofset more quickly then the other filters …

Especialy when the samples start above te dc line .

For example it’s better to draw a saw wave beginning from 0—>1/-1----->0 ( numbers being the amplitude factor )

Instead of 1----->0------> -1

For certain waveforms it’s impossible to start at 0 amp , square for example

WTF! When i re-open uploaded instrument there is no bug iam talking about.

I get you, but no other FX does this:

  1. ive got drum loop with these samples, ive inserted eq, compressor. No problem here.
  2. I inserted moog analog filter. No problem here.
  3. I inserted 4p diode filter. Sound is gone and Redux is showing horrible peaking (moving by cutoff makes some weird sounds on certain frequencies).
  4. I see that 4p filter has default resonance about 20%. So i lowered it, sound was back. When it got above 20% sound is gone, adding drive attentuates a signal and sound is going back, even it is sounding horrible.

If it is feature of 4p diode i thing that it shouldnt have default res bigger than 0 because it will confuse anybody with bad DC samples.
Thank you, it is amazing filter.

I get you, but no other FX does this:

It generally should not happen with the compressors (or other volume based devices) as they are simply shaping the signal amplitude, and not doing anything special with some form of feedback (except if you have insane gain boosts, I guess).

However, plenty of the other DSP effects will do this quite easily.

  • Chorus / Comb Filter / Delay / Flanger / Phaser, with feedback turned up moderately high.
  • EQ 5 / EQ 10 / Mixer EQ, with a moderate amount of gain boost in the lower frequencies (still happens in other bands, but most easily in the low bands where there’s more energy).
  • Reverb with large room size / mpReverb with long duration.
    Of course, each device works differently, and each filter algorithm works differently. Nobody is denying that. Heck, maybe some of the math can be tweaked slightly, I dunno.

Nevertheless, the problem is clearly with your sample material. Crap in = crap out. So just fix the samples and these problems should disappear immediately. :slight_smile:

At the very least, you can process the samples with a simple Butterworth 8n highpass filter with cutoff set to minimum (or insert this as the very first DSP in your chain if you want to be lazy), and it should get rid of most of the problems.

A DC filter is simply a basic highpass after all, so the butterworth will give you pretty nice results.

5730 akiz-sample-bad-dc-hp-fix.png

@dblue, That is some awesome advice! Basically you’re loosing headroom when higher frequences are being “modulated” by lower frequencies, I guess?

That looks pretty neat indeed, thanks dblue! Now time for a ninja FR: Wouldn’t it be awesome if we were able to see those waveforms live in track scopes (a la s(m)exoscope)?

I would just use a gentle 6db highpass .set to lowest value , or bring down the bottom end a bit on an eq

No need for steep filtering

Or even better

Sonimus son eq has a very gentle 1pole highpass …the plug sounds fantastic and is free


Thank you Dblue.
So this topic is usefull after all.