https://relabdevelopment.com/mea-2/

 

I don't quite understand this part. look at the website. anybody cares to explain. looking weird there isn't it ?

 

I've started the demo, let's see how it sounds in the afternoon and we'll try to explain it

 

It's freaking software, it's ALL about technical formulas, ones and zeros, zeros and ones is how they're expressed.

 

Basically they're showing that the Maselec MEA-2 isn't working properly in the sense of constantly. It's changing its Q in an asymmetrical way depending on the freq.
I'd like to compare this to AA Ivory but sadly, Relab's version is iLok'ed.

 

did it looks like the EQ is cramped at the Nyquist to you ? they used machine learning tech, Tone Empire machine learning EQ cramped at the Nyquist too as if they didnt use high enough sample rate. I think somebody forgot to teach the AI about oversampling in DSP.

does the Ivory behave that way ?

 

In the gif the freq is not going high enough.
I remember that Ivory cramps down shortly before Nyquist.
I'll test how Ivory's Q compares to a pure digi EQ, never done this before.

 

So it seems that after all the de-cramping talk of the past years, cramped EQ features are now being marketed as something desirable ("artistically calibrated curves")?

Is this really how the Masalec hardware behaves and if so, why would this be useful?

 

That's an astute observation. Thank you.

Note: Asymmetric curves are not cramping. It's intentionally emulating what the (analog) hardware does, and is a part of its sound.

 

Oh interesting. Endorsed by the creator of Maselec himself. Keen to try it

 

Well, yes and no! Saying that it's intentionally emulating what the hardware does, doesn't negate the fact that this boils down to a cramping-like curve. So, whether we refer to it as "intentional emulation" or as "cramping", the underlying phenomena we are referring to remains the same.

In fact: badly coded EQs that produce cramping could be marketed as "intentional emulations" of "asymmetric behaviour of artistically calibrated curves", if we applied the same logic!

 

From what I can gather, cramping causes asymmetric curve and considered bad because it's... asymmetric. From hundreds of plugin developers that made EQ, the only dev that openly said the asymmetric shape of the cramping doesn't matter (and get slaughtered on the comment section) is those that made Harrison plugins, if one familiar with the Dan Worrall, Harrison and that one attention seeker youtuber I don't remember his name, conflict last year. The video already brought down I believe because of shame after SSL bought Harrison not long ago.

 

Wonder if one day they'll start incorporating the use of big and small zeros and ones in coding... Just like in alphabets

 

Maybe I'm wrong but I always thought cramping is what occurs near nyquist and even some pure digi EQs are doing this.
AFAIR, Trident HW EQs are close to being famous for asymmetric curves.

So, compared it to Kirchhoff. Other than the advertised Relab version, Ivory is more asymmetric in the lower freqs and less in the higher ones. So basically the opposite of what the gif is showing.
Also the Q/Gain relation is different from Kirchhoff's.
No chance to match any setting precisely. At the same LUFSi, Kirchhoff had a higher TP and therefore sounded a bit more open.

Lonesome Sucker, no wait, Lonely Rocker, that's his name.

 

"Timeless music is more art than math"

from the creators of

"every plugin is a distortion plugin if you seek art before the math"

 

Testing, liking it.

Interesting to see mr. John Maselec himself involved in this, Relab acknowledging that endorsements are BS and being all "this time it's different trust me bro". The plugin delivers though, just kinda pricy for what it does.

Yup, it's basically the EQ curve forcing itself back to unity at nyquist. Doesn't seem to be the case with this EQ though the gif implies it does. Bad choice of example i guess.

Isn't that how Winrar works? Making zeros and ones smaller so they use less space?

 

Would be interesting to compare Tone Projects Michelangelo against Indigo as well mmm

 

No - cramping distorts the desired/intended EQ curve at the high end around nyquist because of tradeoffs in the DSP coding techniques used. It's essentially an artefact of poor/naive/CPU-light DSP. And yes, the effect of that cramping results in an asymmetric curve, in that high end response only.

Asymmetric curves *by design* are not an artefact of poor coding, but are a carefully and intentionally designed-in feature to match the behaviour of the original, and occurs across the full range of frequencies.

So both behaviours involve some degree of curve asymmetry, but they are not the same thing, and shouldn't be confused as such.

 

Man, the sales pitch vibe on the webpage is a bit much...

It reads like a 8-week life-changing course that was made through a Clickfunnels website

 

1) if it's by design, for a 10k Mastering Grade Equalizer that advertised (and printed on the faceplate) as "Precision Stereo Equalizer" it's kinda odd

2) asymmetric shape, theoretically sounds bad (whether it can be heard or not is different story) because of unbalanced gain distribution for affected frequencies around the center-frequency of the boost. Simply put low Q boost associated with smooth sounding while sharper high Q boost commonly sound unnatural. The affected surrounding frequencies played a huge role in the overall boost sound. Also kinda odd especially for a proportional EQ topology to be like that by design as the Q value is variable depends on the amount of boost. More like manufacturing flaws caused by inaccuracies of the component values which shouldn't happen at the first place because of its price tag. Consumer/professionals would've assumed they are built with carefully selected high-grade parts.

3) Asymmetrical curve can still happen on hardware equalizer because it's close to impossible to get 100% precise shape but it should be around standard acceptable +/- tolerance. From the graphic on Relab's website I believe if that is below the tolerance of manufacturing standard, it's a little too low quality for a mastering grade equalizer.

4) I don't think the hardware reference unit measured is that bad curve asymmetric-wise, I still think there's something wrong with the measuring technique used (machine learning) to gather the information.

5) first time seeing this in an emulated legendary mastering grade EQ plugins, nobody talks or even do this before, probably because they view it as a flaw as a proper engineer would, don't know for sure.

6) Neve 1073 is bad design according to engineering community and even the late Rupert Neve hated it till his last moment and he kept building a new EQ because he embarrassed by it (the man admitted it), but people love 1073 for its flaw, which again a different story. Still doesn't make it a good design from engineering point of view, well it is depends on where are you looking from, sales perspective, it's genius.

 

Last edited: Oct 26, 2023