Analog EQs

[cpsmusic]

Hi Folks,

I generally do all of my audio work "in the box" using plug-ins, although I have a small collection of outboard gear that I use mainly for recording. The main mic preamp that I use is a Warm Audio WA-273 EQ which is basically a copy of a Neve 1073.

I've recently been experimenting with using hardware on my mixbus and I've found that the WA-273's EQ sounds far better than any plug-in EQ. Without going all "audiophile",  the difference is quite noticeable, adding a sort of "3D" depth to the sound. I suspect that this is due to the EQ's inductors doing things to the signal's phase, as well as slight differences in the EQ settings between the L/R channels. Whatever it is, it definitely sounds better than software (e.g. a Sonimus Burnley 73).

So my question is, is this difference specific to inductor EQs or is it something that all analog EQs have (more or less)?

The reason I'm curious about this is because I'm planning to build a W492 EQ but I'm thinking if this effect is specific to inductor EQs then maybe I should look at some other design.

Cheers,

Chris

 

[Bo Deadly]

IMO, for basic EQ that does not involve subtle distortion, there really shouldn't be any difference between hardware and a vaguely good plugin equivalent. But there are many professional sound engineering types that will vehemently dispute this and that's ok. But you asked so this is my opinion.

Inductor EQs make for simpler circuits that can have better noise performance because of low insertion loss because both the inductors and capacitors are reactive components whereas with resistors there needs to be some active gain that is going to also boost noise. Active circuits consume more power which could result in noisier power rails. Inductor circuits can more easily be lower impedance which helps with distortion and noise. There are technical differences. But, even though these effects might be measurable, they are probably not perceivable by a human being (when operated correctly).

So it is probably that you're just hearing some difference in the frequency response and then your instinct is to decide that because it's different, it's "better".

Another factor is the generally superior experience of working with something that you can touch and look at. The cooler looking it is, the better it sounds. For sure. Blinking LEDs and bouncing meter needles are mesmerizing and generally improve listening experience. Perhaps most important is how expensive and unique it is. I actually would not discount the merits of this phenomenon. There is nothing that will kill some music mojo more than reaching for a computer, blasting your visual cortex with a backlight and then fiddling with little drop downs and fake controls.

If you really want to do a more accurate test that factors out psychoacoustic effects, use FFT software to carefully match a plugin to the 273 and then create an AB recording where you alternate between the two playing different tracks. Then start the recording at different random locations with your eyes closed and listen. Take breaks once in a while to reset listener fatigue (when you listen to the same thing over and over it becomes annoying and you can't really think about it objectively).

The bottom line is that gear and software is 1/1000th of quality a quality recording. The artist's performance is all that really matters. Just make sure you get a nice clean dry track. Don't spend an hour fiddling with compression. That hour would be better spent just playing / singing.

 

[cpsmusic]

It's as if digital signal is in 2D or 2.5D and then you hit some powerful equipment and bam, there's the 3D space you've been looking for.

Yes, that's the effect I perceive. I notice it particularly with things like doubled vocals which seem to separate with respect to front-to-back depth. What's interesting is that I notice a similar effect (although less pronounced) with Acustica plug-ins which are based on impulse responses of real gear.  And it seems similar to the effect I get with my SPL Vitalizer which makes me think it's something to do with phase.

Yes, I probably should do a proper double-blind test of this, and no, I'm not going to rush out and get a hardware EQ. However, it does make me want to build one!

 

[abbey road d enfer]

Both references to phase are unsubstantiated and don't stand scrutiny.
I think squarewave gave a pretty thorough rundown of the mechanics and aesthetics involved.
I just want to add that iron-core or ferrite-core inductors, as used in audio, are significantly non-linear, so they add their own flavour of distortion, like transformers do. Transformers produce their dirt all over the audio spectrum, while inductor-based EQ's tend to generate it in specific bands.

 

[cpsmusic]

I've been thinking about this and was wondering whether the effect could be due to a sort of "stereo-ising" that's happening due to the L/R channels of the EQ not being identical. After all, this unit isn't designed for mastering so I'm not sure how close the EQs of both channels are.

 

[Bo Deadly]

There are so many parts of electronic equipment that change the sound yet this is denied here on the forum to a large degree.

The reason I like this site is because there are people here who are EEs or at least understand EE well enough to have a grounded technical discussion of electronic devices. I am not an EE but I have some familiarity with what Acoustical Engineering people do, I've "worked" in their labs, I've been in the large anechoic chamber at Bell Labs down the road from where I live today and I've gone to a few Acoustical Society conferences (to sit by the pool and drink but you'd be surprised what you can learn doing that). A LOT of work has gone into studying the differences in how people perceive sound, their thresholds for detecting subtle changes in pitch, distortion, levels and so on. There are people on this site that appreciate that work as science. So please try to understand that many of your proclamations about "3D" sounds and "phasing" are not going to universally appreciated. I'm not calling anybody a "phool". I'm just not going to participate in subjective discussions that are not evidence based.

 

[JohnRoberts]

There are so many parts of electronic equipment that change the sound yet this is denied here on the forum to a large degree.

Adam

Huh?

I am surely repeating myself but I have been writing about this stuff for decades (in the 80's I wrote an audio column "Audio Mythology").  If you can reliably*** hear and identify a specific sound characteristic, it will have a physical manifestation that can be measured. If we can measure it we can manage it in circuit designs (or nowadays in DSP coding). Back in the 70s I rolled some of my own semi-custom modified test equipment to better measure stuff missed by conventional tests.

I learned decades ago that I could measure pretty much anything that I could reliably hear, BUT I could also measure things on my test bench that I could not hear. Maybe along with my sloppy thinking I also suffer from tin ears.  :

A useful technique to isolate hard to identify sound path differences is to null them with each other... What is the same will cancel out leaving only what is different. Note: this can be a little challenging with digital paths because conversion delays can prevent deep HF nulls.

I do not find it productive to argue with people on the WWW about what they claim to hear (actually arguing about anything on the internets rarely works).

JR



*** proved by double blind listening tests with statistical certainty

 

[gyraf]

From a GS post back in december, on the pultec and how it behaves, related to it's load impedance:

On a side note, I think the complex and occasionally very-low input impedance of the Pultec circuit actually contributes to its subjectively perceived qualities: This is why you can boost a fair amount of top into just about anything at high Q and still not have it screaming at you: When there is no significant signal content in the area you are boosting the filter is still easy to drive, but whenever spectral parts turn up within the boost range, limited large-signal drive capabilities combined with heavy loading tends to more-or-less invisibly limit your boost amount - in effect working a bit like a deesser or ever-so-slightly like a dynamic eq, suppressing large-scale swings within your boost area. This also why driving your Pultec with "should-be-incorrect" output impedances like hifi outputs often works surprisingly well, specially for that beloved high end boost. A requirement is off course that your actual output impedance behavior is different from small-signal to large-signal.

Just mentioning this so that you don't run yourself into the ground by coming up with ultra-low-impedance drivers to cancel out imaginary problems, killing hidden niceness in the process. Been there, tried that actually.

I am aware of at least one well-known passive eq that runs a high-current transistor amplifier stage to drive its filters at very-low impedance, enabling frequency-determining inductances to be much smaller, but voiding the pultec'ish high end. Tinkering with that one made me realize that most of it's VERY nice midrange coloring (absolutely great for vocals) came from this power transistor stage, not filters or makeup gain. Win some, loose some.

/Jakob E.

 

[JohnRoberts]

From a GS post back in december, on the pultec and how it behaves, related to it's load impedance:

/Jakob E.

Ding ding ding,,,, yes signal chain termination interactions matter....

Some old school gear that was designed anticipating 600 ohm input and output terminations can have unexpected consequences interfacing with modern gear.

JR

 

[Bo Deadly]

When there is no significant signal content in the area you are boosting the filter is still easy to drive, but whenever spectral parts turn up within the boost range, limited large-signal drive capabilities combined with heavy loading tends to more-or-less invisibly limit your boost amount - in effect working a bit like a deesser or ever-so-slightly like a dynamic eq

I don't quite understand. The only way I can see how this would happen is if the source impedance went up over all frequencies even though it was only loaded at over a narrow band. What kind of output would do that? A tube with supply sag?

 

[ruffrecords]

  If you can reliably*** hear and identify a specific sound characteristic, it will have a physical manifestation that can be measured. If we can measure it we can manage it in circuit designs (or nowadays in DSP coding). Back in the 70s I rolled some of my own semi-custom modified test equipment to better measure stuff missed by conventional tests.

I think this goes to the heart of the matter. Analogue gear has a very complex transfer function which varies with signal level, spectral content and input and output load impedances. It is relatively straightforward to measure transfer function under static conditions but it is extremely hard to characterise it under all dynamic conditions and I think this is the aspect missed most by software emulations of analogue equipment. The result is their emulations are only an approximation to the real world. There will therefore be measurable but not necessarily attributable differences between the plug in and the actual analogue equipment.

These differences may well be audible.

Cheers

Ian

 

[abbey road d enfer]

I don't quite understand. The only way I can see how this would happen is if the source impedance went up over all frequencies even though it was only loaded at over a narrow band. What kind of output would do that? A tube with supply sag?

The output impedance of an active stage varies with level; actually the variation is dependant on the value of the instant signal. NFB tends to minimize the variation. Many vintage units have little or no NFB.
Add to that slew-rate induced effects.

 

[john12ax7]

You can view plugins as a sort of simulation, with most focusing on fancy GUIs and computing efficiency.  Would the public even want accurate simulations if you could only use 1 before the CPU  ran out of gas? There are other issues as well,  poor handling of aliasing, poor non-linear approximations,  over reliance on FIR filters,  poor implementation of dither,  and so on.

So it's not really a surprise that they don't tend to sound as good as well designed hardware. 

 

[emrr]

My impression upon diving into DAW world years ago was the EQ's drove me crazy.  I couldn't 'hear' them because they lacked the artifacts created by hardware, and those artifacts were my mental cue for the response effects.  Once I adjusted, which took awhile, it was fine.  Sometimes you want EQ, sometimes you want EQ with artifacts. 

 

[Michael Tibes]

So my question is, is this difference specific to inductor EQs or is it something that all analog EQs have (more or less)?

In my experience: no, it is not specific to inductor EQs. I don’t know why, but digital works, analog rocks.

Michael

 

[abbey road d enfer]

My impression upon diving into DAW world years ago was the EQ's drove me crazy.  I couldn't 'hear' them because they lacked the artifacts created by hardware, and those artifacts were my mental cue for the response effects.  Once I adjusted, which took awhile, it was fine.  Sometimes you want EQ, sometimes you want EQ with artifacts.

"years ago" may be a clue to why you had this experience. I was the Steinberg distributor when they came out with their first VST. I thought the EQ's were weird, so I checked them out in the AP and found out that, although the nominal points (amplitude/frequency) were correct, the curve was not. Typically a 1kHz boost was accompanied with a 500Hz cut and a 250 Hz boost. It turned out the developper who had written the code had no audio experience, had it all aout of a book, and did not have the equipmnet to measure, and no one at Steinberg thought of checking, like a hardware mfgr would have done. It took two releases to correct the problem.

 

[abbey road d enfer]

In my experience: no, it is not specific to inductor EQs. I don’t know why, but digital works, analog rocks.

Michael

I would say it depends very much on how you use EQ.
Most of the tone control I do at tracking, with analog gear.
I use the DAW EQ to do "surgical" stuff such as damping tom resonance, kick pedal noise, string noise, whatever.
Then I may do a final EQ pass in the external digital mixer. I've never felt the need to insert an analog EQ in the digital chain.

 

[SWAN808]

Hi Folks,

I generally do all of my audio work "in the box" using plug-ins, although I have a small collection of outboard gear that I use mainly for recording. The main mic preamp that I use is a Warm Audio WA-273 EQ which is basically a copy of a Neve 1073.

I've recently been experimenting with using hardware on my mixbus and I've found that the WA-273's EQ sounds far better than any plug-in EQ. Without going all "audiophile",  the difference is quite noticeable, adding a sort of "3D" depth to the sound. I suspect that this is due to the EQ's inductors doing things to the signal's phase, as well as slight differences in the EQ settings between the L/R channels. Whatever it is, it definitely sounds better than software (e.g. a Sonimus Burnley 73).

So my question is, is this difference specific to inductor EQs or is it something that all analog EQs have (more or less)?

The reason I'm curious about this is because I'm planning to build a W492 EQ but I'm thinking if this effect is specific to inductor EQs then maybe I should look at some other design.

Cheers,

Chris

I know what you mean, although I'm not sure many people really no why. Majority tend to say its placebo but I've passed blind tests in the past (I believe I passed some on the 1073 from a certain developer who claimed their model was indistinguishable)...so I can't help with the suggestion why except say I've noticed differences in software and hardware processing that I put down to the very subtle differences between signal processing in the analogue and digital domain...it's quite rare that I prefer the digital processing from a sound perspective but also rare that I prefer the convenience and impact of my wallet in the analogue processing...

 

[rock soderstrom]

I don’t know why, but digital works, analog rocks.

+1

 

[abbey road d enfer]

I don’t know why, but digital works, analog rocks.

I don't want to be rude, but to me it's similar to saying "Bud sucks, Coors rule".
It does not really help others making an opinion.
Perhaps I'm too much a quibbler nitpicker engineer and not enough a salesman politician TV expert artist, but this kind of assertion should be backed by comparative audio abstracts. Or a demonstration of the superior workflow.