Digital EQ Fact & Myth.

[Dubai]

to me all decent eqs sound diferent,
even with near 100% same curves.

there are others that dont.

what sounds best?
ive seen personal taste & music style has strong bias in that decition.

[Laurend]

The user interface is a key point. That's why I like the FLUX Epure.

[elan]

Originally Posted by jdtrbn

I did run the test. I get second and third harmonics and the noise at all levels (you just have to believe me this time). It's interesting that you can get over -13db / 23% harmonic distortion, which is basically all 3rd harmonic and then there's a 2nd harmonic below -70db. You get to about -60db / 0.1% THD before hitting the red light. It may have to do with the bit depth the host is using, so that normal hosts would clip before getting to -13db THD.

There's something very interesting going on here that happens only in analog mode. My guess would be that the noise contains a ground hum peak or something which turns into intermodulation distortion here, see the attachment.

but how it is possible that the OP nulls the SSL and API emulation with a normal eq if it generates harmonics? i mean the sonar eq doesn't generate harmonics

[Paul Frindle]

Originally Posted by elan

but how it is possible that the OP nulls the SSL and API emulation with a normal eq if it generates harmonics? i mean the sonar eq doesn't generate harmonics

Originally Posted by jdtrbn View Post
I did run the test. I get second and third harmonics and the noise at all levels (you just have to believe me this time). It's interesting that you can get over -13db / 23% harmonic distortion, which is basically all 3rd harmonic and then there's a 2nd harmonic below -70db. You get to about -60db / 0.1% THD before hitting the red light. It may have to do with the bit depth the host is using, so that normal hosts would clip before getting to -13db THD.

Are you really honestly talking about a S/W EQ emulation with THD at 23%, before it hits clipping!?

[elan]

Originally Posted by Ethan Winer

* It's probably not possible to overdrive a software EQ assuming it uses 32 bits internally.

i talked with some plug in developer and they told me isn't true that a 32 float bit doesn't saturate. it's true the opposite. a fixed 24 or 48 bit software clips, doesn't saturate, while a 32 float doesn't clip because saturates, because it approximate without clipping.

[elan]

Originally Posted by MHB850

Wow, I'm really glad I didn't bring this up before I got titled a big name, because being called an incompetent deaf troll once in a day is bad enough but 15 times would totally bum me out.

there might be some great eq's that do the same thing but the Air Eq sounds very much like an aphex aural exciter.

michael

sorry i was reading the thread and probably i quoted not the right examples.

i repeat the question, the OP has nulled some HW emulations like Waves SSL and API and others which generates harmonics, he has even nulled the Air Eq all with the Sonar EQ.

I just don't understand how it is possible the Sonar Eq, which doesn't generate harmonics, can null EQs which generates harmonics in different ways.

And i don't understand why the Air Eq sounds so different to me than the other Eq (and not only to me even to Michael Brauer) if can be nulled with the Sonar Eq.

In particularly the filters, the Q sounds so different to any other Eq digital and analog.

I can mute a resonance with some Db with the narrowest Q if i watch the spectrum analyzer see the resonance almost disappear, i mean it goes at the level of the other frequencies or even lower, sometimes with just few db.

I would like to see the values in the Air Eq (if you can post a picture with the plug in with show mode on) because they also seems pretty identical.

I believe in math, graph, etc, but why my ears hear this Eq so different to all the others?

[Ethan Winer]

Originally Posted by elan

i talked with some plug in developer and they told me isn't true that a 32 float bit doesn't saturate. it's true the opposite. a fixed 24 or 48 bit software clips, doesn't saturate, while a 32 float doesn't clip because saturates, because it approximate without clipping.

I'm not sure what you're trying to say, but I proved in my AES Audio Myths video that the 32-bit Sonitus EQ does not change the sound at all, even when passing audio that's 17 dB over Digital Zero. That section is 54:44 into the video.

--Ethan

[anabolique]

Originally Posted by virtualaudio

A as demonstrated by Fader8 in his informative website: Index
The pass filters on the Waves Renaissance EQ have an unusual and significant pass band bump beyond q = 1, which would not be ideal for DC filtering (see above link).

The measurements were done at 96kHZ,24 bit. With low-pass filter, graph was also showing some pass-band (narrow) above 30kHZ. SO, what is that? OK, if its 96kHZ, it will not alias and we cant perceive that. But, what would happen if sampling rate is 44.1k or 48 k? Would that odd hi-freq pass band be on the same place in frequency range? Or worse, lower? Would it be filtered out in antialiasing process? OR, MAYBE as some digital expert guy said, its misunderstood that EQ will work better on sampling rates high as 88.2k or 96k, because it needs to be more precise, so those bumps wouldnt be present at lower sampling rates? This is of course case-by-case thing, some eq plugs can handle , some dont...

[The Listener]

I have already expressed my opinion that all digital EQs are not equal and I obviously use the ones that are coded that way - PlparEQ, UAD Neve1081, 1073, Softube Trident A range, UAD Helios (I hope someone from the Nebula community will sample this device, it is a great tool already in the not so perfect UAD version), etc., but lately I admire the Nebula3 EQ programmes the most - especially VMQ PRO -

AlexB Nebula Programs - VMQ PRO: Vintage Master EQ PRO

Can someone measure and analyze what this EQ does, because compared to anything else I own, it produces "magic"... I am not sure what it does, but for example boosting 3400Hz by several db with this EQ or doing the same thing with anything else digital I have - IS NOT THE SAME.

I would be glad if someone could measure what's the "magic" of this "unit"... and why is it so different from anything else and so sweet...

[Tangible]

Originally Posted by The Listener

I have already expressed my opinion that all digital EQs are not equal and I obviously use the ones that are coded that way - PlparEQ, UAD Neve1081, 1073, Softube Trident A range, UAD Helios (I hope someone from the Nebula community will sample this device, it is a great tool already in the not so perfect UAD version), etc., but lately I admire the Nebula3 EQ programmes the most - especially VMQ PRO -

AlexB Nebula Programs - VMQ PRO: Vintage Master EQ PRO

Can someone measure and analyze what this EQ does, because compared to anything else I own, it produces "magic"... I am not sure what it does, but for example boosting 3400Hz by several db with this EQ or doing the same thing with anything else digital I have - IS NOT THE SAME.

I would be glad if someone could measure what's the "magic" of this "unit"... and why is it so different from anything else and so sweet...

Well, in the Ultimate Plugin Analysis thread, it's been demonstrated that the statement "all digital eq's are the same" is completely false. Nebula and a few other eq's, vary from most digital eq's in that it has analog-like bell curves and phase......and it doesn't deform at high frequencies like most digital eq's. In addition to that, Nebula's eq programs also add harmonic distortion and saturation....based on the actual hardware it's sampling, making it even more unique from the typical algorithmic eq.

[mossie23]

Originally Posted by Tangible

Well, in the Ultimate Plugin Analysis thread, it's been demonstrated that the statement "all digital eq's are the same" is completely false. Nebula and a few other eq's, vary from most digital eq's in that it has analog-like bell curves and phase......and it doesn't deform at high frequencies like most digital eq's. In addition to that, Nebula's eq programs also add harmonic distortion and saturation....based on the actual hardware it's sampling, making it even more unique from the typical algorithmic eq.

rhythminmind, the original poster of the "all digital eq's are the same" thread, clearly stated that dynamic convolution eqs (like nebula) and linear phase eq's are different. he claimed that all other eqs are the same, altough some add distortion.

[aleatoric]

DMG eQuality is the first clean (no added saturation) digital minimum phase EQ I have tried that does not null with the rest of the herd. None of eQuality's phase modes (digital, digital+, minimum phase, analog) null in my tests. Could not even get close. All other clean minimum phase digital EQ's I did this test with I could get to -70dB or below after matching filter shapes. Looks like eQuality has some new technology under the hood. It sounds great too!

[emilision]

I'm still convinced all of Logic's EQs damage sound (including the LP)...they all seem to add graininess/phaseyness...it's more noticeable with acoustic instruments though...

I'm convinced the Sonnox EQ is superior, though that's the only other EQ I have...so it may well be that it's just Logic's stuff that's at fault. I'm tempted to buy into UAD for their Manley Passive; but it's a big investment...and I like my laptop being "self-contained". Nebula is unusable in an active project...great for post-production though.