DMG EQuilibrium

[DaveGamble]

In FIR mode, for Phase, what's the difference between "analogue" and "zero-latency analogue"? The manual says,

Since they provide the same "cumulative minimum phase response" as each other, how do they differ?

Latency. Analogue Phase is identical to EQuality Analogue Phase. ZL Analogue phase is the same thing with the latency reduced. Same magnitude and phase responses, but since some people had developed a very strong affinity for EQuality's Analogue Phase, I figured I'd keep it in EQuilibrium unchanged.

Dave.

[DaveGamble]

Hey,

Thanks for asking good questions.

If choosing a minimum-phase option, what advantage is there to FIR-mode ("global
minimum", "analogue", or "zero-latency analogue") over IIR-mode ("Digital+Phase" off, or "Digital+Phase" on)?

Ability to design the impulse response in detail. In FIR mode, you can window, adjust phase and generally go wild. If you're just doing regular EQ work, then IIR mode all the way.
If you're doing something more demanding, FIR mode gives you more flexibility.

IIR-mode has less latency and less CPU-demand (no?) than FIR-mode, so that's an advantage to IIR-mode; but what would be any advantage to FIR-mode for minimum-phase?

For regular analogue-style cumulative minimum phase, IIR is great.
If you need the flexibility of the FIR design, it's there in FIR mode.

Dave.

[DaveGamble]

What would be the purpose in choosing the option with latency, when the zero-latency option is available? Why would you ever choose to add useless latency? Why would an equalizer-design include a mode to add useless latency?

Strictly, under certain circumstances, there could be an eccentricity in the centering of the window function in Analogue phase vs ZL analogue, depending on cumulative phase. It's not really clear to me whether this could be audible or not. Nonetheless, it's an option on a menu that allows users to definitely get a sound they've used before and love. It's hardly making the UI more cluttered; it's on a menu with a bunch of options.

Dave.

[DaveGamble]

I just tested, and it's not just a matter of processor power. "Analogue" and "zero-latency analogue" actually yield different audio. I set up two tracks, with the same source material (an impulse) on both tracks. On one track I put Equilibrium in "analog" mode, and on the other track I put Equilibrium in "zero-latency analogue" mode, and they do not null -- i.e., they yield different audio.

Exactly. Could you hear a difference?
I included both options because I didn't want anyone to be without, in case they COULD hear a difference.

Dave.

[DaveGamble]

That will work for the terms "linear phase" and "minimum phase", because those terms have precise meanings in engineering circles, and you can ascertain those meanings with research.

Research will not, however, enable you to learn what DMG means by the terms "IIRDigital+Compensation", "IIRDigital+Phase", "FIR Analog Mode", and "FIR Zero-Latency Analog Mode", because these are terms which were coined by DMG and have no meaning determined by engineering conventions.

Yes and no.

IIR and FIR have precise engineering meanings.
Zero-latency has a precise engineering meaning.

Digital+, Digital+Phase and "Analogue Phase" are descriptions of various bits of tech that form part of the EQ implementation. I try to explain them in the manuals. I have a set of videos underway to explain them in greater detail though.

Dave.

[DaveGamble]

NB: There's inevitably a disparity about how I think about these things as designer vs how you'd think about them as a user, so if any of my answers are bad, just say, and I'll elaborate further.

[Funkybot's Evil Twin]

Dave, since videos are underway, perhaps you can address some of the differences (both practical and sonic) between the options. For instance, I don't know the difference between an IIR and a FIR and don't really care except: which will sound better? And the phase modes? Which will generally work best on what material?

My common Use Case scenarios are:

1. Zero latency mixing EQ
2. High quality "render" EQ (mix)
3. Mastering EQ (master)

Is there one setting that you'd recommend for each of those scenarios for the highest possible audio quality while taking into account the latency concerns? I'd be curious about that, even if the answer was something like "well if you're mixing percussion I'd stay away from X setting because it may smear transients, but that would work well on more melodic material."

[Hermetech Mastering]

Thanks for all these clear explanations Dave, looking forward to the videos too!

[vicoli]

I just tested, and it's not just a matter of processor power. "Analogue" and "zero-latency analogue" actually yield different audio. I set up two tracks, with the same source material (an impulse) on both tracks. On one track I put Equilibrium in "analog" mode, and on the other track I put Equilibrium in "zero-latency analogue" mode, and they do not null -- i.e., they yield different audio.

Exactly. Could you hear a difference?
I included both options because I didn't want anyone to be without, in case they COULD hear a difference....

I didn't listen for a difference. I was just doing an objective null-test.

[vicoli]

Thanks for the explanations, DaveGamble.

In FIR mode, for Phase, what's the difference between "analogue" and "zero-latency analogue"? The manual says,

Analogue and Zero-latency analogue provide the same cumulative minimum phase response as hardware.

Since they provide the same "cumulative minimum phase response" as each other, how do they differ?

Latency. Analogue Phase is identical to EQuality Analogue Phase. ZL Analogue phase is the same thing with the latency reduced. Same magnitude and phase responses, but since some people had developed a very strong affinity for EQuality's Analogue Phase, I figured I'd keep it in EQuilibrium unchanged.

So, Analog mode gives the same magnitude-response and phase-response as ZLAnalog mode, but with added latency for the benefit of people who prefer unnecessary latency. It seems strange to me that anyone would prefer unnecessary latency.

[Krzysztof Oktalski]

It seems strange to me that anyone would prefer unnecessary latency.

It's not a question of preference, it's about consistency. There are thousands of EQuality users that moved to EQuilibrium, we needed to offer them analogue phase with identical latency so they could reproduce their previous setup phase perfect. Some users work in environments with no automatic latency adjustments and are calculating and adjusting shifts manually within their chain. As Dave's already stated, the ZL Analogue Phase mode was different to the original EQuality mode so we included both to ensure continuity.

If you think there's a better way of tackling this problem then we're open to suggestions, but it's imperative we support users that have supported us from the start and that's why we chose this two mode solution.

[vicoli]

It seems strange to me that anyone would prefer unnecessary latency.

It's not a question of preference, it's about consistency. There are thousands of EQuality users that moved to EQuilibrium, we needed to offer them analogue phase with identical latency so they could reproduce their previous setup phase perfect. Some users work in environments with no automatic latency adjustments and are calculating and adjusting shifts manually within their chain. As Dave's already stated, the ZL Analogue Phase mode was different to the original EQuality mode so we included both to ensure continuity.

Thanks for the additional clarification. As I understand things now, I personally am going to pass on FIR Analog mode, since:

1) I personally have no need for compatibility with EQuality, and
2) if I understand DMG's explanations correctly, the magnitude-response and phase-response differences between Analog and ZLAnalog modes are very small, and probably inaudible in general. (The responses are different, as I described above when I measured the null-test on an impulse-response -- but the difference isn't by design.)

If my reasoning is faulty, please let me know? Otherwise, I think I finally understand, and my confusion is resolved. Until now, I've been hesitant to disregard the Analog option, because I was concerned I might be missing out on something useful to me personally.

If you think there's a better way of tackling this problem then we're open to suggestions, but it's imperative we support users that have supported us from the start and that's why we chose this two mode solution.

Thanks for your interest in my opinion. As far as your design and implementation: totally good, in my opinion. As for documentation: Why not take the explanations DMG has provided to me here and include them in the official documentation? Copy / paste from this thread into the manual, take 3 minutes to finalize the new text, and then publish the revised manual in place of the older one? (I see you're working on videos too, but manuals work better for me.) The cost would be 10 minutes; the benefits would be more users getting the best possible results in less time, with less guessing; and also less minutes of users asking for customer-support pertaining to the DSP modes.

I appreciate DMG's work to provide excellent DSP and also excellent flexibility in the form of Setup options. I appreciate that you put DSP above everything else. I recognized this before, which is why I rejected people's suggestions above that I should bail on Equilibrium and resign myself to Pro-Q or Equality. But if you want to take your excellent DSP to the level where users can make the best of it, my respectful suggestion would be to review your KVR and GS threads, note the questions asked by users, and add answers to the manual.

Thanks for your help.

[scintillator]

+1 for improvements of the manual. Concise is good, detailed is better.

[Lathomas42]

I have a basic question about this EQ,

I have the Processing mode set to FIR and the phase set to Linear. My understanding of FIR filters is that Linear Phase FIR filters must have a Symmetric (or anti-symmetric?) Impulse response however the display of the impulse response is not symmetric or anti-symmetric. Wouldn't this result in a nonlinear phase response?

[djanthonyw]

Are presets not loading in 1.52 for anyone else?