Filter specs?

[Derozer]

Hi all,

I’m tweaking some freq/gain knobs of a VST’s filter I like (which I’m quite sure is made with JUCE; a basic decompilation reveal lots of JUCE references).

Should be a Peak filter.
Here’s the response analysis at different hz/db (giving an input of 1024 dirac samples).

700hz/0db:


700hz/+40db:


10khz/+40db:


100hz/+40db


Which kind of filter is this? Native of JUCE? Own-made by the developer? I’d like to experiment a lot with it, but don’t know which one to start with.

Tried ask on official JUCE forum, without any success.

Can you help me? Thanks

[rd2play]

I used Juce only very little, but I thought it only offers libraries to write filters, not finished filters. Why can't you experiment with it, what stops you?

[Derozer]

I used Juce only very little, but I thought it only offers libraries to write filters, not finished filters. Why can't you experiment with it, what stops you?

Don't know honestly where to start from. Design peak filter? Biquad? With distortion? Which type? And so on...

The context is so HUGE, maybe there's something out of the box on JUCE, that's why I ask the question...

[rd2play]

Yeah, not sure, but I don't think Juce has anything out of the box. It must have been done by somebody.

You said you don't know where to start from. What are you trying to do? Are you trying to learn about filters? Because for using a filter you don't really need to know if it's a biquad, DFI, or DFII, cascaded, or whatever...

[Derozer]

Yeah, not sure, but I don't think Juce has anything out of the box. It must have been done by somebody.

You said you don't know where to start from. What are you trying to do? Are you trying to learn about filters? Because for using a filter you don't really need to know if it's a biquad, DFI, or DFII, cascaded, or whatever...

I'd like to replicate the responses I've posted above

[Derozer]

For example: what are those "ripples" on highs when I tweak the filter? For a Peak, its Very unusual, isn't?

[mystran]

For example: what are those "ripples" on highs when I tweak the filter? For a Peak, its Very unusual, isn't?

The ripples look like harmonic distortion (and some of it looks like aliasing), but that's about as much as anyone can tell you from the information available to us. The fact that a plugin itself is made with JUCE doesn't mean much, because often JUCE is used for the plugin APIs and GUI rather than the actual DSP.

[juha_p]

Have you tried another analyzer?
Here's one utility you could try - https://www.kvraudio.com/product/eq-cur ... rtom-audio

[hugoderwolf]

For example: what are those "ripples" on highs when I tweak the filter? For a Peak, its Very unusual, isn't?

If it happens only while tweaking, it's likely zipper noise from the plugin updating filters in steps. You should be able to hear it if you run a sine wave through it.

Btw there's nothing special about the filters in JUCE compared to any other standard biquad thingy. What are you trying to achieve though?

[Derozer]

For example: what are those "ripples" on highs when I tweak the filter? For a Peak, its Very unusual, isn't?

If it happens only while tweaking, it's likely zipper noise from the plugin updating filters in steps. You should be able to hear it if you run a sine wave through it.

Nope, its "constant" when the params of the VST are tweaked.

As @mystran said, probably a Peak filter + some kind of Harmonic Distortion unit based on cutoff freq?
In fact, if I move the the cutoff param, the harmonic distortion slide to left/right.

Could a video helpful if I post it here?

What are you trying to achieve though?

Well, this: replicate the behaviour of this VST's FX unit, based on a X/Y pad (freq/db), which does what I've showed above (sound pretty dirty and nice to my ears).

[juha_p]

...

As @mystran said, probably a Peak filter + some kind of Harmonic Distortion unit based on cutoff freq ?
In fact, if I move the the cutoff param, the harmonic distortion slide to left/right.

...

...because of high gain (as you boost +40dB there).

IIRC (I left Windows few years ago), VST Plug-in Analyzer , by Christian Budde, helped in finding these harmonic distortion issues but, it didn't support VST3 at that time ... . Dunno if there are other such tools available ... .

[mystran]

As @mystran said, probably a Peak filter + some kind of Harmonic Distortion unit based on cutoff freq?
In fact, if I move the the cutoff param, the harmonic distortion slide to left/right.

Spectrum alone is not terribly useful in figuring out the nature of distortion. If you try filtering a saw-wave and look at the output waveform in time-domain, then this can sometimes give you clues as to what is going on... but in general you're probably looking at a very long and tedious process, where you're basically trying to guess what might be going on, then implement that and compare, then repeat ad nauseum.

ps. Basically from a purely technical point of view (ie. ignoring legal considerations), just decompiling the plugin of interest is probably much easier, even though that's a long and tedious process as well.

[Derozer]

As @mystran said, probably a Peak filter + some kind of Harmonic Distortion unit based on cutoff freq?
In fact, if I move the the cutoff param, the harmonic distortion slide to left/right.

Spectrum alone is not terribly useful in figuring out the nature of distortion. If you try filtering a saw-wave and look at the output waveform in time-domain, then this can sometimes give you clues as to what is going on... but in general you're probably looking at a very long and tedious process, where you're basically trying to guess what might be going on, then implement that and compare, then repeat ad nauseum.

ps. Basically from a purely technical point of view (ie. ignoring legal considerations), just decompiling the plugin of interest is probably much easier, even though that's a long and tedious process as well.

Well, my target is not to replicate "exactly" what is going on (I don't care), but at least the main process involved in the FX chain.

Tried to decompile the .vst3: Can't intercept neither the labels showed in the plugin (using Ghidra). So will be very hard...

I've followed your suggestion though: here's the saw outputs in time domain; maybe can help me further (or confirm what you already said).

700hz/0db (i.e. no filtering):


700hz/+40db:


10khz/+40db:


100hz/+40db:


Does it helps?

[urosh]

Does it helps?

Resonant BPF -> super heavy clipper/saturator could produce this. I mean really superheavy, this is scorched.

edit: BPF with mixed clean input to be precise

[mystran]

We see hard clipped resonance, but +40dB is probably too far to extreme to make heads or tails about the actual nature of the distortion, which might not be quite that hard with less extreme settings.. but like you should also try with higher input gain and what not to try to build a more complete picture of how the filters behave in different situations... and then you basically start guesswork.