EQ plug-ins... I just don't hear it

[jdtrbn]

OK, I've demoed a lot of plug-in EQs. Some free, some demos and some that my friends own. These include plug-ins like Epure, Sonalksis, Sonnox and apEQ. I've really tried to hear the difference in quality and character (though I chose fairly clean EQs that aren't emulations of analog gear).

The result is that I don't hear any significant differences between these. This troubles me because some people are pretty enthusiastic about EQs and I have read a lot about the theory behind them. I can hear the pre-echo of a linear phase filter and such, but the actual character is something I can't hear (except for between LP EQs and minimum phase ones).

It didn't encourage me when I phase-inverted and summed a processed wave with a wave that was processed with another EQ and got something that peaked around -40db, much quieter than the original signal (it was hard to believe at first so I did the test with other settings and got a very quiet signal again). This happened only with certain two EQs filters, though, but the resulting differences were very small with some others too.

So what I did was that I matched the frequency curves as well as I could with VST plug-in analyzer and then matched the RMS. I tried bell filters at >12khz range and somewhere below 1khz, with high Q and low Q, with high boost and low boost. Test material included some music I listen to and some guitar and rhodes sounds, through headphones or speakers.

I don't think I have so bad ears. It's very easy to hear differences between different compressors, for example. I was just thinking if there's something I haven't learned to pay attention to, or if I just really have "wooden ears". I've heard the hint of boosting 14khz but most sound contains so little of this that the boosts are hardly audible. And in real-life usage the boosts would hardly exceed 6db anyway.

So, what am I supposed hear?
Artifacts, distortion, temporal differences, phase shift (this one I actually doubt) or something else? Or should I just give up?

[cooker]

You should try them on cleaner sources like vocal or acoustic guitar. Try playing around 2 khz as for me I can detect differences easyer around there. Are you using decent monitors? the 4 eq examples you give are generally considered as clean sounding eq's but they do have differences. Just for example epure sounds smoother than apEQ when you boost a freq. .

[A440]

I think the differences between the better plugin eq's can be subtle at first, but more apparent the more you use them.

Example. I've been using the Sonalksis eq as my go-to for the last couple of years, I like the interface, the HP filter is pretty clean and all-in-all it sounds very good to my ears.

Then, I bought an SSL Duende recently. At first I thought...'well yeah it sounds good...but is it that much better than the Sonalksis?'...I just wasn't that convinced. Then after a lot of mixing with it I suddenly realised that the right 'sounds' that I wanted to hear were jumping out of the speakers...like, all of a sudden, things were really coming into focus and were 'right' and 'up front' in a way that is hard to get with the Sonalksis.

Don't get me wrong. I still love the Sonalksis too, but it works in a completely different way to the Duende ...if I'm looking for tonal shaping that's not so up front, then I'll use the Sonalksis. If need the track to have more presence and bite then I'll use the Duende.

Very happy to have both these colours on my pallette.

[Jaguar Dreams]

I went through a period lately of shooting out a lot of software eq's and I hear *a lot* of difference between certain ones. The first thing I would wonder is how is your monitoring situation and room treatment? Assuming that's taken care of...

Here's what I look for when evaluating a software eq. Hopefully this may be of some use:

1) Phase distortion grossness on drum overheads when using a hi shelf/boost. Pick a song section where the drummer is working the ride cymbal.
2) How does it handle boosting and cutting at 8K and 5K on those same overheads and in particular, what is this doing to you perception of the snare drum? Again, is it nastifying the overall sound with phasing or just making the snare sound bad? Or is it taming/bring character to the snare in a pleasing way?
3) Throw all the eq's you are testing on the snare drum track and turn them all off. Now, one at a time, how long does it take you to get the sound you are looking for?
4) Take an acoustic guitar track that has a LDC on the body and a SDC on the neck. How well does the eq work for getting a good sound?

I use this approach when evaluating software eq for the following reasons. First, it seems like most software eq has an easier time doing reasonable cut/boost down in the sub 1K range. From 1K up it seems like it gets much harder to make a good eq, and if the eq is good, it seems like you hear more of the differences in character from 1K on up. Second, snare drum just has so many interesting freq zones and possibilities for cut and boost depending on the tone you are aiming for, that the versatility of the eq is really pushed to its limit on snare. Third, I think of acoustic as being one of the least agreeable instruments to take software eq well, so that also makes for a useful test.

I happen to think Sonnox and Sonalksis are both very good eq's, and they fall into a more "transparent" category - they do the job really well with no noticeable phasing artifacts but they don't necessarily stamp the sound with their signature. When trying out Sonnox, it struck me that it would be an excellent choice for anyone mixing modern country or singer/songwriters.

I notice you didn't list URS. They make my current favorite software eq. Try them out. I think you will find, for instance, that the A series EQ has quite a different character than their S Series EQ. The whole URS line is aimed at being EQ with attitude. I ended up buying their Channel Strip Pro and I love it for mixing rock music.

I also happen to really like the UAD Neve 1073 EQ and the UAD Pultec Pro EQ, primarily for vocals and drum buss, in that order, but the URS has taken over as my primary eq and compression tool.

You mention that you were using some already completed songs plus just some individual tracks - I'm not sure if that's going to be the best way to compare eq's. At the end of the day, the primary use of eq is to sculpt individual tracks or groups to work with each other in the context of the mix, so evaluating them in any other context is probably not the best choice.

Best of luck!

Cheers,
-Jaguar

[Doc Mixwell]

Digital [itself] adds almost nothing to the equation; EQ's of similar layouts, should hopefully sound the same on you're tracks....one would hope.

Digital EQ is [should be] void of phase shift artifact, which [going on a long shot] is what makes an analog EQ offer up a certain character.......[besides the chosen parts used] which is why you can tell a MAJOR difference between them when tweezing the audio. Probably something to do with Phase, internal power supplies and the go-go juice that comes out of you're wall.

With digital compression; the artifacts are much more apparent; and the scale of what you can do with gain reduction and time [enveloping of the compression] in the digital realm vs what you can do in the analog realm is the shootout. I like to test each box with equal compression to see how much they "offend" the sound on the track.

[jigsawlogic]

Roc, seems to me that digital EQs change phase unless they're specifically phase linear. Try multing your signal, adding EQ to a channel, and then recombining while listening for phase-related dropouts.

[art_is_hard]

when i mix at school, i can hear a huge difference in the many different EQ's. when im mixing at home, not so much. the main difference is the school is using custom adam monitors, and im using infinity speakers/headphones.....not to mention my room isnt treated properly, and the school's room is nicely paded with aurelex foam --- so i would say that the listening environment is a huge difference when you want to hear subtle differences like that.

[cooker]

Originally Posted by art_is_hard

when i mix at school, i can hear a huge difference in the many different EQ's. when im mixing at home, not so much. the main difference is the school is using custom adam monitors, and im using infinity speakers/headphones.....not to mention my room isnt treated properly, and the school's room is nicely paded with aurelex foam --- so i would say that the listening environment is a huge difference when you want to hear subtle differences like that.

Custom adam monitors?

[Doc Mixwell]

Originally Posted by jigsawlogic

Roc, seems to me that digital EQs change phase unless they're specifically phase linear. Try multing your signal, adding EQ to a channel, and then recombining while listening for phase-related dropouts.

I guess I just meant having overlapping of bands....and boosting the same way I would with an analog equalizer to adjust the tone.

My bad....

[vernier]

The result is that I don't hear any significant differences between these. This troubles me because some people are pretty enthusiastic about EQs and I have read a lot about the theory behind them.

Your discovery here seems to correlate with the recent "Daw sound" thread.
'

[art_is_hard]

Originally Posted by cooker

Custom adam monitors?

yeah, they are similar to the A7's but they are a bit bigger. had them made for the recording classes.

[Jack Ruston]

Try some eqs that are designed to have a bit more charachter...waves api for example? you'll feel the love a bit more.

Overall I do agree with you.

J

[thethrillfactor]

Originally Posted by jdtrbn

So, what am I supposed hear?
Artifacts, distortion, temporal differences, phase shift (this one I actually doubt) or something else?

If you mix with your eyes which are the people who plug in EQ's are really designed for you should hear everything.


(Relax guys this is a joke...sorta).

Originally Posted by jdtrbn

Or should I just give up?

The "digital" dark side clouds everything.

"Impossible to see the future is".

[Clem Snide]

Originally Posted by art_is_hard

...the school's room is nicely paded with aurelex foam...

Mmm, padded with foam: nice! Those low frequencies will be quakin' in their boots.

[jdtrbn]

Hi again.

As some of you suggested, Waves API or URS EQs have character and I can hear that. I just don't understand why there are so many clean EQs and why some are considered better than others (except for features/GUIs). When we start to make emulations and add distortions and such deliberately, you get differences even in the different implementations(Waves/URS/UAD/Liquid Mix).

I have made all my tests with single sounds, not in a mix. I used music as test material because I hoped the large amount of harmonics would bring the differences up, and because Epure is considered a good mastering plug-in.

My listening conditions aren't good. I'm using $200 speakers and the room is untreated. That's why I listened with my AKG K240 headphones. Because of the talk there is about equalizers, I thought I should be able to tell the difference without professional monitoring, and if I couldn't, the differences were probably so small that the end-listeners wouldn't hear (or feel) them either.

When the frequency curves are matched, the differences I could think of were pre-echo and echo-like temporal differences, distortions and phase differences. For example Ethan Winer thinks that phase differences can't be heard when the original signal isn't played at the same time (or overhead or other recording of the same source). Also Massive Passive user manual says that the differences between EQs aren't likely to be because of phase differences. Fader8 has a website where he has listed different echos or "ringing" in minimum phase filters. However, this has been completely inaudible to me except for in linear phase filters at low frequencies. At high frequencies the phase is so short that the echo is almost non-existent.

You know what a lot of audio pros think about audiophile cables? This clean EQ plug-in stuff seems a bit similar to me, though I'm open to arguments and such. As I said, I haven't tested in a mix, but even if I did, that wouldn't explain the different mastering EQs and secondly, no-one would ever have time to try all those plug-ins. And as far as I know, people often draw conclusions without matching the frequency curves, which is OK when EQs have obvious character, like API.

Maybe I'll still try with clean sources (shouldn't the differences be there with less clean sources too?) and some mid-range boosts.

PS. If someone hears a difference between Sonnox bell filter I and Sonalksis filter III with matched frequency curves, I'd like to know.

[nightjar]

If a plug-in is an EQ, give me an EQ... not a "pleasant-distortion generator".

Ultimately, all well-designed EQ's should sound the same if the parameters are set up the same.

Now if you want some character....

Give me a different plug-in that offers the parameters to add or subtract the specific quality that is being changed. Don't disguise it as an EQ with pleasant distortion. Let me be informed as to what I'm really doing to the signal.

[lllubi]

Originally Posted by nightjar

If a plug-in is an EQ, give me an EQ... not a "pleasant-distortion generator".

Ultimately, all well-designed EQ's should sound the same if the parameters are set up the same.

Now if you want some character....

Give me a different plug-in that offers the parameters to add or subtract the specific quality that is being changed. Don't disguise it as an EQ with pleasant distortion. Let me be informed as to what I'm really doing to the signal.

what creates then the "magic" of a GML or Hammer in the top end?
thought always it is their distortion

[Doc Mixwell]

Originally Posted by lllubi

what creates then the "magic" of a GML or Hammer in the top end?
thought always it is their distortion

OR no distortion, if we're talking about the GML.......

The magic is in the circuit.....and the power supply....as far as I can tell.....

We can talk about digital EQ's, or emulations; there is you're difference.

Digital EQ adds nothing except precision. Most have zero mathematical code that adds "mojo" or "vibe" to the frequency response, unless you program it in there. I could be wrong....I dunno.

[nightjar]

Originally Posted by lllubi

what creates then the "magic" of a GML or Hammer in the top end?
thought always it is their distortion

You're right... this coloration is a pleasant distortion of the original signal.

What I'm saying is that was then and this is now...

Let's not disguise our actual signal processing under false labels based on pleasant artifacts.

Let's label stuff what it really is....

An EQ plug-in should be an EQ.... neutral in it's character.

A colorization plug-in should be something else. Call it "magic" if you want.

[Ethan Winer]

Originally Posted by nightjar

If a plug-in is an EQ, give me an EQ... not a "pleasant-distortion generator". Ultimately, all well-designed EQ's should sound the same if the parameters are set up the same. Now if you want some character....

Give me a different plug-in that offers the parameters to add or subtract the specific quality that is being changed. Don't disguise it as an EQ with pleasant distortion. Let me be informed as to what I'm really doing to the signal.

Give this man a ceegar.

[Paul Frindle]

Originally Posted by jdtrbn

Hi again.

Maybe I'll still try with clean sources (shouldn't the differences be there with less clean sources too?) and some mid-range boosts.

PS. If someone hears a difference between Sonnox bell filter I and Sonalksis filter III with matched frequency curves, I'd like to know.

This discussion raises some interesting points about design - in that you are trying to compare EQs based on near identical responses. A product design does not rest on this factor alone. However compelling it may seem on the surface of it, one of the biggest fallacies is the notion that the true nature of an artistic tool can be determined by response comparisons and 'null tests'.

It is true to say that, providing an EQ does only EQ (and not distortion or novel phase stuff), matching the exact same response you would imagine ought to produce the same result.

Issues where they do not match might range from non-conformance to displayed settings, non-conformance to the actual ideal response - and in older fixed point designs, noise and limit cycling at low levels, which can change response with level, change Q with level and content and cause all sorts of low level mayhem which you would hear in real life. For instance your nulling to only -40dB represents a 1% error, if this is only a freq response error that's not too grave, but if it is noise, distortion and limit cycling or something else it is very significant.

But all of the above is actually less than half the story when it comes to product design. Just like any creative 'instrument' what people actually get from it is heavily determined by the kind of user feedback that goes on when the the engineer is actually locked into a creative feedback loop with the sound of the program, his own taste and the application. The whole business is an artistic pursuit with a human in the cycle, where how it all 'feels' and interfaces with the user and the 'flow' of the experience is probably more important than any other single factor.

So the design of an EQ intended to be creative is not just about 'numbers' or 'responses', although they are important - it's a about the whole sonic experience, encouraging intended use and outcomes based on fashion and style, creating the right user involvement and feedback, making people 'feel' right when they are trying to create their art. This level of design cannot be achieved only in terms of numbers, responses and specifications - and it is certainly not something that can be achieved without loads of experience, by simply employing processing savvy alone. It is largely this level of design that separates technically 'good' EQs from celebrated artistic tools people are comfortable with. There are many of the former, but many fewer of the latter.

This isn't something that can be appreciated by measurement or null comparisons after ages trying to set up the exact same responses. It is something you can only appreciate when you actually use these things, under pressure in a real artistic environment - when you are actually intending to create art.

[Ethan Winer]

Originally Posted by Paul Frindle

your nulling to only -40dB represents a 1% error, if this is only a freq response error that's not too grave, but if it is noise, distortion and limit cycling or something else it is very significant.

I don't disagree entirely with your premise, but the OP said he matched EQ parameters by the knob settings. So he might have achieved -80 dB nulling if he spent more time on it. If it were me I'd do the nulling in real time to make it faster and easier to hone in on the null. That is, I'd add the same clip to two tracks, invert the polarity of one, then tweak one EQ to get the lowest signal level at the output.

--Ethan

[lllubi]

as the starter mentioned there are compressor plugins that generate
pleasant harmonics close to the originals

what is it exactly that limits EQ plugs to create "magic" in the top end?
do we need higher sampling rates or else?

[Paul Frindle]

Originally Posted by Ethan Winer

I don't disagree entirely with your premise, but the OP said he matched EQ parameters by the knob settings. So he might have achieved -80 dB nulling if he spent more time on it. If it were me I'd do the nulling in real time to make it faster and easier to hone in on the null. That is, I'd add the same clip to two tracks, invert the polarity of one, then tweak one EQ to get the lowest signal level at the output.

--Ethan

Yes - that is possible. But the point is that nulling in this way tells you little if anything about the product itself.

A whole discussion like this once broke out concerning actually nulling the OXF-R3 against ProTools or some such! This was also a pointless exercise that told people nothing of value to them. The whole concept of these applications could not have been more different.

As a reverse example - the fact that our latest DSM plug-in will null to itself completely to silence was actually thought by some people at one point as a potential fault! The point being that almost all other apps these people had would not even null with themselves (apparently) - and it was therefore thought that the DSM could not have actually been working!

But the deeper point with all that is - what are people actually expecting to find out?

For instance if you took a couple of LA2As or Fairchild tube compressors and did the same test, you would be hard pressed to get even a 12dB null. What would this tell you about these units? Would they be any the less classic, somehow invalidated and have their long standing reputations for creativity quoshed?

[Paul Frindle]

Originally Posted by Ethan Winer

Give this man a ceegar.

Hmm. yes, I agree that all clean EQs (of the same basic phase response) set to produce exactly the same response really should sound the same.

But that is the point - the parameters you set do not necessarily give you the same results across EQ types and applications - depending on their intended artistic purpose. This is part of what gets you the character - and is part of what I mean.

Some of this aspect of things is described here (well worth a read):

Sonnox Oxford Plugins

and here:

Sonnox Oxford Plugins

Of course there is more subtlety not described publicly, like control laws and parameter interactions etc.

Hope this helps :-)

[James Lehmann]

I think given the wealth of high-quality EQ tools out there, a good GUI can be a deal-breaker.

[dannygold]

Originally Posted by vernier

Your discovery here seems to correlate with the recent "Daw sound" thread.
'

Good point. Time for someone to say "even if it phase cancels perfectly in a null test is still SOUNDS different".

[nightjar]

Originally Posted by James Lehmann

I think given the wealth of high-quality EQ tools out there, a good GUI can be a deal-breaker.

I agree...

As the sonic quality of clean EQ gets down to splitting hairs, here's what matters to me:

1) Well-designed GUI
2) CPU efficiency
3) System compatibility
4) Cost

[Ethan Winer]

Originally Posted by Paul Frindle

Yes - that is possible. But the point is that nulling in this way tells you little if anything about the product itself.

Well, let's say one EQ comes free bundled with your DAW and another sells for $1,500. So before dropping the $1,500 you go to a friend's who has the expensive one, and uses the same DAW, so you try different EQ settings on a few tracks. If you can null one against the other, and your main concern is sound quality, I'd say that proves the $1,500 would be better spent elsewhere.

if you took a couple of LA2As or Fairchild tube compressors and did the same test, you would be hard pressed to get even a 12dB null.

I agree. But then we're talking about intentional color and grit, which are effects other than EQ.

I agree that all clean EQs (of the same basic phase response) set to produce exactly the same response really should sound the same.

Yes, that's all I'm saying.

Though I do agree with the comments about GUI and usability.

--Ethan

[Nishmaster]

Originally Posted by dannygold

Good point. Time for someone to say "even if it phase cancels perfectly in a null test is still SOUNDS different".

Ridiculous. The very fact that something cancels completely in a null test means by definition that it cannot possibly sound different.

There is also another weird myth that I've seen propagating lately around here, and that's that somehow some eq's can exhibit more or less phase change than others. The way eq works is by altering the phase relationship between the original signal and the derived signal and adding them back together, thus producing a frequency specific change in gain.

Given two equal eq settings (Q, Center Frequency, Cut/Boost), any two eqs will produce an identical phase response. Of course, in a digital world, the steep sampling filter will introduce ringing and pre-echo, but with many digital eqs oversampling internally, I would say that this problem is mostly academic. Linear phase eqs, are the exception to all of this, of course.