I would add that EQs are kinda like cars... if you know how to use it, you will get good results no matter which one you use. The differences are rather marginal, but admittedly they can accumulate. EQ should be used more sparingly than most of the people use it anyway.

 

That's why I use ProQ3 less. I tend to do butchery with digital equalizers. So I only use them to remove resonances. Otherwise, I mess things up and it doesn't work. everything become thin...
As I like colored stuff, I use emulations of Neve or pultech , Etc.

 

Measured with one band at +18dB, 45Hz, Q of 2.5. CraveEQ in Analog mode at the same settings for reference.





So yeah, q632 is using FIR filters with a buffer size too small. This is clearly a bug, and I'm quite surprised it was released in this state.

 

Last edited: Nov 3, 2023

Yikes... ok...

I'll reiterate what I said previously, because it's true:

PluginDoctor (and Bertrom, to my surprise) do not / can not measure q.632 correctly.

I know this because I have a dev version of q.632 --- identical dsp but with a generic Kush AU shell that can host Kush dsp --- and it measures exactly as anyone would expect. The main diff is that the dev q.632 has wwaayyy longer latency.

But those two q.632's null completely --- as in, when monitored with a bit scope, there is no residue whatsoever. None.

And beyond the science, they sound identical both to each other and to the last version of Magpha ever, and they sound phenomenal.

I have no idea why plugindoctor and bertrom can produce two different measurements for two identically-set plugins that null perfectly under all scenarios. But that's the reality on the ground.

I put the above forth, along with my reputation, for the believers, so you can filter out the noise and continue to trust both your ears and your instincts.

Gregory Scott - ubk

#111
https://gearspace.com/board/new-product-alert/1416518-q-632-kush-magpha-eq-reborn.html

 

I finally made some measurements and tests with surprising results:

Spoiler: Q632 in 'newer' PD - nothing surprising

Spoiler: Q632 in old PD - woot?

Spoiler: Q632 quick setting for drums

Spoiler: Kirchhoff matched (more or less precise)

Spoiler: Q632 'superior' phase

Spoiler: Kirchhoff 'inferior' phase (set to Analog)

Even if they didn't cancel out eachother in a phase inversion test (~-60dBFS peak for pink noise and ~ 20dB less for drums), in an A-B double-blind test, I wasn't able to differentiate these two at all.

 

Spoiler: Pro-Q pre-post spectrum comparison

Don't think I need to tell anybody present here what's going on.
No signs of ringing or combing.
Tested all bands, with different freq and Q settings - the result is the same. No cramping either.

 

I guess we need a transient test then, to really see its silky-smooth phase. Preferably with high freqs. Too bad I'm lazy.

 

My Fourier transform of a dirac impulse shows just that (CraveEQ in red that's behaving correctly, q632 in blue with severe comb-filtering) :



The problem comes from the truncated FIR filters the plugin is using (see the impulse response from my previous reply). Basically, a filter needs to ring to do its job correctly. If you use steady-state signals to measure the plugin, such as a slow sine sweeps or white noise, the artifacts will never appear, because the FIR filters will be convolved at a steady rate, and you won't encounter the abrupt end of the post-ringing.

However, if you use fast enough transients, you will encounter the post-ringing, which is pretty much busted in this plugin. The current version of PluginDoctor is using very fast sweeps to measure plugins, and they're fast enough to reveal the artifact. So the only way to hear this, is with extreme settings on fast transients.

 

That would mean that the "smoothness" is achieved with basically delaying the signal and compensating for a phase shift. Sounds deliberate.

 

this is by far the most incomplete GUI Kush ever made he must started sniffing coke lol .....naw he cool peeps but wtf was he thinking its not ugly its just incomplete looks like he started on it for 30 minutes and got tired of fucking with it lol @BlackHaze1986 look at this shit lol

 

Kush Audio Compressor is nice......Try his compressor Novatron you will like it on acoustic drums and if u want that vintage looseness or tightness its there fully he did something under the hood because it really does sound good i think thats one of his best vst's .....it don't sound too good on electric drums for some reason

 

Given that Greg seemed unaware of the issue on the GS thread, I don't think it's deliberate.

Also, you never want to set the buffer length of a FIR filter that short. This creates a copy of the signal that is delayed in time (and with an inverse polarity in some settings), the perfect recipe for comb-filtering. I don't think that the goal of this plugin was "Comb-filtering on transients only when the EQ is set at extreme settings"

 

im dead homey said you have cute nightmares lol thats some type of shit i would say lol

 

What you mean with i should look at this Shit? The GUI of this Plugin?

@Atlantis84 My GoTo's for EQ's are UVI Shade ,FabFilters ProQ3 and Kirchhoff. For analog emulations i use the Nebula stuff.

@Atlantis84 I have never used Kush Audio Plugins so i can't compare the tech side, but GUI Wise not my type compared to the ones i use.

 

Last edited by a moderator: Nov 6, 2023