The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
Why do hardware EQs and software EQs differ in sound?
Old 1 week ago
  #1
Lives for gear
 
soundmodel's Avatar
 

Why do hardware EQs and software EQs differ in sound?

Why do hardware EQs and software EQs differ in sound? Or do they (in all cases)?

And I already expect:

-"non-linearities"
-"real components"
-"real electricity"

But since these can be simulated using software, then these do not "fully" explain what causes the difference.

Preamps?
Old 1 week ago
  #2
I don't think they can be simulated exactly like hardware in all cases. So, they're...different.
Old 1 week ago
  #3
Gear Addict
They can be simulated, but it's difficult. There's a difference between simulating the function of a circuit, and what that physical circuit board will actually do.

Simulating what that physical circuit board will actually do means taking into account a lot of things, ranging from PCB layout to component thermals, how rigid (or not) the PSU rails are, etc etc etc.

Getting all those effects in will make very small differences, but will require much more processing power. Nobody wants a single EQ that'll max out your CPU.

Chris
Old 1 week ago
  #4
Lives for gear
 
soundmodel's Avatar
 

Quote:
Originally Posted by chris661 View Post
Getting all those effects in will make very small differences, but will require much more processing power. Nobody wants a single EQ that'll max out your CPU.
Hmm this explanation I like. That it's about "side effects". That one can surely recreate the EQ curve, but not all the "nuances" that occur when parameters are turned, when gain is altered etc. that occur in real electronics.

I could try the Acustica EQs sometime. They're supposed to be about trying to model those things, and they are CPU hungry, but I've heard they sound hardware-ish.
Old 1 week ago
  #5
Village Idiot
 
Labs's Avatar
 

Quote:
Originally Posted by soundmodel View Post
Hmm this explanation I like. That it's about "side effects". That one can surely recreate the EQ curve, but not all the "nuances" that occur when parameters are turned, when gain is altered etc. that occur in real electronics.
-"non-linearities"
-"real components"
-"real electricity"

Gustav
Old 1 week ago
  #6
Lives for gear
 
swafford's Avatar
 

It's the knobs.
Old 1 week ago
  #7
Lives for gear
 
kennybro's Avatar
I guess I'll address the "Or do they" part of the post...

Is there some easily identifiable characteristic that separates all hardware EQ from all software EQ? All EQ's sound different to me. A Massive Passive sounds very different from a Klark Teknik, and both sound different from an API. And those are generally very different from graphic EQ's.

Pretty sure I could identify EQ plugs that sound far more like a Pultec than a Tascam graphic would. Is there a sonic characteristic you are identifying that is uniquely shared by Pultec and the Tascam, but not present in any plug? Would you rather use a Behringer hardware EQ than Apogee EQP-1A plug?
Old 1 week ago
  #8
Lives for gear
 

I don't know. I can do recordings using either and if you asked me or anyone else which was used in a blind listening comparison there is no way in hell you can tell the difference.

The much bigger issue which gets overlooked here is the elephant in the room.
Why do you need to use so much EQ to begin with?

I remember back when I first started recording I used to EQ the crap out of instruments prior to them being tracked.
Then I switched to recording them as is and EQing the crap out of them when mixing.
Somewhere along the way I found that method sucked too so I focused on getting the best tones possible before I tracked.

Today, I rarely if ever use any EQ on Bass or guitars. It isn't needed. I use amp modeler hardware that lets me dial up the ideal tones before I hit the record button, especially when recording direct. Prior to that I was using all that EQ trying to make it sound like I was playing through a miced cab. Now I only have to select the cab type to get the ideal tones.

I still use some EQ on vocals. I've never had a great voice but at least with the New Ribbon mics I been using, the tone match for my voice has never been better. A few mild bumps and I'm good to go. Drums are the other. For canned drums or drum machines, I typically need little or no EQ. Real drums are a much bigger challenge. Even there I have reduced the need of using huge amounts of EQ by simply getting the mics in phase so the bleed over between mics isn't causing drums to sound thin as hell. I mostly use Low and High band passing to take care of things I don't need then use a decent buss compressor for the rest.
Old 1 week ago
  #9
Quote:
Originally Posted by soundmodel View Post
Hmm this explanation I like. That it's about "side effects". That one can surely recreate the EQ curve, but not all the "nuances" that occur when parameters are turned, when gain is altered etc. that occur in real electronics.

[...]
It might be worth considering that many of the 'nuances' associated with hardware EQs typically take the form of what in an earlier time we called distortion and noise.


Early on, outfitting my then-new project studio, I bought a pair of 5 band parametrics, not expensive, $350 apiece. I never put them on a scope but they sure sounded bad to me. (I bought them new, should have returned them but I was new to the whole gear acquisition thing.) Knowing what I know now, but not having had or used the boxes in question in years, I would say that they sounded like they had substantial crossing point distortion (nasty grittiness).

With the exception of some 1990s DSP software, I would say that virtually every software EQ I've used since then has sounded considerably better, more transparent.

But, for sure, we're not talking apples and oranges, here. More like apples and lemons.
Old 1 week ago
  #10
Lives for gear
 
soundmodel's Avatar
 

Quote:
Originally Posted by theblue1 View Post
It might be worth considering that many of the 'nuances' associated with hardware EQs typically take the form of what in an earlier time we called distortion and noise.

...

With the exception of some 1990s DSP software, I would say that virtually every software EQ I've used since then has sounded considerably better, more transparent.
Yes, but in art "good", "better" are subjective. Thus "lower S/N" is not necessarily "nicer sounding". A matter of taste. Some people like "clean digital", I like it sometimes.
Old 1 week ago
  #11
Gear Guru
 
Karloff70's Avatar
 

Is it maybe a better question to ask why they in fact would sound the same? As they are not the same thing.

One is a thing doing what it does and the other is some lines of code, trying to approximate what the first thing was doing. But not in an 'all encompassing every detail' way. People say because the computing power isn't there yet. Maybe so. I am thinking even if it we had the power, we likely haven't even got all the parameters that matter together yet, to even PUT into code to make a perfect simulation.
Old 1 week ago
  #12
Quote:
Originally Posted by soundmodel View Post
Yes, but in art "good", "better" are subjective. Thus "lower S/N" is not necessarily "nicer sounding". A matter of taste. Some people like "clean digital", I like it sometimes.
Of course.

But when we're talking about tools and tool sets, it can be helpful to understand the factors that contribute to those subjective evaluations.
Old 1 week ago
  #13
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by theblue1 View Post

With the exception of some 1990s DSP software, I would say that virtually every software EQ I've used since then has sounded considerably better, more transparent.
Sure depends what you are listening for, for it to be transparent TO though, eyh. Might not be adding any noise. Or distortion. But that is such a limited 'couple of obvious parameters' way of thought. Might be slyly stealing your image. Your 'place'. That illusion of a space you created with some mics.

I still haven't heard a digital eq that doesn't at least change or even ruin the space in a delicate stereo image recording of something real with good mics. Making it less deep and flatter. Either a little bit, or lots. Always some. I'd rather have a little noise and keep my 'place' intact.
Old 1 week ago
  #14
Quote:
Originally Posted by Karloff70 View Post
Is it maybe a better question to ask why they in fact would sound the same? As they are not the same thing.

One is a thing doing what it does and the other is some lines of code, trying to approximate what the first thing was doing. But not in an 'all encompassing every detail' way. People say because the computing power isn't there yet. Maybe so. I am thinking even if it we had the power, we likely haven't even got all the parameters that matter together yet, to even PUT into code to make a perfect simulation.
A key distinction, of course.

But, actually, while a simulation of a given piece of hardware will be, to varying degrees, attempting to mimic the performance characteristics of a specific hardware device across its intended performance range, a non-sim designed for the dedicated purpose of effecting certain changes to the linearity of frequency response -- not a recreation but a dedicated piece of software designed to primary purpose -- doesn't need to waste design effort trying to simulate what the initial gear designers quite likely saw as divergence from the ideal.
Old 1 week ago
  #15
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by theblue1 View Post
A key distinction, of course.

But, actually, while a simulation of a given piece of hardware will be, to varying degrees, attempting to mimic the performance characteristics of a specific hardware device across its intended performance range, a non-sim designed for the dedicated purpose of effecting certain changes to the linearity of frequency response -- not a recreation but a dedicated piece of software designed to primary purpose -- doesn't need to waste design effort trying to simulate what the initial gear designers quite likely saw as divergence from the ideal.
It might not be simulating a specific hardware eq. But it is still 'simulating' what EQ does. It's code. The code describes things we know about what eq does to the sound and then we hear that description. But our descriptions in code are not specific enough yet in some way, or else no space would get lost as it still does, and it doesn't matter whether the plug is a 'virtual analog' plug or a straight digi eq algo. They all steal space. My point is that if we had all the relevant parameters to code there wouldn't be that loss of 'the place'.
Old 1 week ago
  #16
Quote:
Originally Posted by Karloff70 View Post
It might not be simulating a specific hardware eq. But it is still 'simulating' what EQ does. It's code. The code describes things we know about what eq does to the sound and then we hear that description. But our descriptions in code are not specific enough yet in some way, or else no space would get lost as it still does, and it doesn't matter whether the plug is a 'virtual analog' plug or a straight digi eq algo. They all steal space. My point is that if we had all the relevant parameters to code there wouldn't be that loss of 'the place'.
Well, to be sure, a basic digital FIR filter DOES in a sense simulate what a conventional analog EQ does, at core, feeding back a slightly time-shifted image of the current signal onto that signal to create cancellation in desired frequency ranges, more or less directly parallel to the processes in an analog FIR filter.

https://en.wikipedia.org/wiki/Finite_impulse_response

That said, I'm not really sure what you mean by the term 'space.' You allude to a phenomenon/loss I'm not familiar with from my own experience, though I've used analog filters consciously for a half century and various generations of digital filters for maybe half that.

(Now, the first DSP EQ software solutions I used, in the 1990s, were, indeed, disappointing. But the EQ's I've come to prefer over the last 15 years or so I felt have tended to be what I would describe as transparent, clean, and effective -- in the sense that they were largely capable of delivering the EQ changes I wanted to make without degrading the overall sound or adding noticeable noise or distortion.)
Old 1 week ago
  #17
Lives for gear
 
soundmodel's Avatar
 

Quote:
Originally Posted by Karloff70 View Post
Is it maybe a better question to ask why they in fact would sound the same? As they are not the same thing.

One is a thing doing what it does and the other is some lines of code, trying to approximate what the first thing was doing. But not in an 'all encompassing every detail' way. People say because the computing power isn't there yet. Maybe so. I am thinking even if it we had the power, we likely haven't even got all the parameters that matter together yet, to even PUT into code to make a perfect simulation.
Philosophically, epistemologically, one could ask, whether it's possible that physics theory and real physics could be "nearly the same". I.e. that it's possible to write an algorithm that recreates the "functions" that occur in real electronics (assuming purity of the incoming electricity, low environmental variables etc.).

Again, this is why I posit this question. Problems in "purity" of electronic signal, low environmental variables etc. are often a design-goal in electronic design. Thus this would suggest that electronic gear designed that way should perform closely to "idealized". Now then if we have "idealized physics theory", then wouldn't it be possible to model it to high degree of accuracy? If the theory and the real world highly correspond?
Old 1 week ago
  #18
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by soundmodel View Post
Philosophically, epistemologically, one could ask, whether it's possible that physics theory and real physics could be "nearly the same". I.e. that it's possible to write an algorithm that recreates the "functions" that occur in real electronics (assuming purity of the incoming electricity, low environmental variables etc.).

Again, this is why I posit this question. Problems in "purity" of electronic signal, low environmental variables etc. are often a design-goal in electronic design. Thus this would suggest that electronic gear designed that way should perform closely to "idealized". Now then if we have "idealized physics theory", then wouldn't it be possible to model it to high degree of accuracy? If the theory and the real world highly correspond?
Well, I think we ARE modelling to a 'high degree of accuracy' now. Only, that's a relative statement. And not at all the same as saying therefore they are now the same.
Old 1 week ago
  #19
Lives for gear
 
soundmodel's Avatar
 

Quote:
Originally Posted by Karloff70 View Post
Well, I think we ARE modelling to a 'high degree of accuracy' now. Only, that's a relative statement. And not at all the same as saying therefore they are now the same.
What software EQs are an example of high-degree of accuracy then?
Old 1 week ago
  #20
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by theblue1 View Post
Well, to be sure, a basic digital FIR filter DOES in a sense simulate what a conventional analog EQ does, at core, feeding back a slightly time-shifted image of the current signal onto that signal to create cancellation in desired frequency ranges, more or less directly parallel to the processes in an analog FIR filter.

https://en.wikipedia.org/wiki/Finite_impulse_response
Just a little caveat to your otherwise accurate post: Digital EQs are not inherently simulations of analogue EQs. Both digital and analogue EQs are applying the exact same mathematical principles. One approach does it with electronics, the other does it digitally but neither is a simulation of the other. Don't let the Luddites and ignorami confuse the issue.

Alistair
Old 1 week ago
  #21
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by soundmodel View Post
Philosophically, epistemologically, one could ask, whether it's possible that physics theory and real physics could be "nearly the same". I.e. that it's possible to write an algorithm that recreates the "functions" that occur in real electronics (assuming purity of the incoming electricity, low environmental variables etc.).
They are the same and are approached in exactly the same way. The first analogue EQs were designed by people that implemented the mathematical formulas needed to affect the signal, in electronics (by lack of other options). We can also implement the same (and more) maths (with more accuracy) digitally. The principle is EXACTLY the same: Implementation of mathematical formulas.

Quote:
Again, this is why I posit this question. Problems in "purity" of electronic signal, low environmental variables etc. are often a design-goal in electronic design. Thus this would suggest that electronic gear designed that way should perform closely to "idealized". Now then if we have "idealized physics theory", then wouldn't it be possible to model it to high degree of accuracy? If the theory and the real world highly correspond?
We already do but modelling is the not the right term as both approaches are doing the exact same thing in different ways. Digital is NOT a simulation of analogue. (I am not talking about specific products of course. I'm talking in general).

Alistair
Old 1 week ago
  #22
Gear Guru
 
Karloff70's Avatar
 

Principle same, outcome different. In reality, not same then.
Old 1 week ago
  #23
Lives for gear
 
soundmodel's Avatar
 

Quote:
Originally Posted by UnderTow View Post
We already do but modelling is the not the right term as both approaches are doing the exact same thing in different ways. Digital is NOT a simulation of analogue. (I am not talking about specific products of course. I'm talking in general).
I'm talking about "analog-modelling EQ". E.g. Acustica stuff.

It's like "digital filter theory" + "perspectives and implementation on how to mimick 'analogueness' using digital technology"
Old 1 week ago
  #24
Lives for gear
Quote:
Originally Posted by UnderTow View Post
They are the same and are approached in exactly the same way. The first analogue EQs were designed by people that implemented the mathematical formulas needed to affect the signal, in electronics (by lack of other options). We can also implement the same (and more) maths (with more accuracy) digitally. The principle is EXACTLY the same: Implementation of mathematical formulas.



We already do but modelling is the not the right term as both approaches are doing the exact same thing in different ways. Digital is NOT a simulation of analogue. (I am not talking about specific products of course. I'm talking in general).

Alistair
+1

But
What did you come in this galley?
Old 1 week ago
  #25
Quote:
Originally Posted by Karloff70 View Post
Is it maybe a better question to ask why they in fact would sound the same? As they are not the same thing.

One is a thing doing what it does and the other is some lines of code, trying to approximate what the first thing was doing. But not in an 'all encompassing every detail' way. People say because the computing power isn't there yet. Maybe so. I am thinking even if it we had the power, we likely haven't even got all the parameters that matter together yet, to even PUT into code to make a perfect simulation.
That might be true if you’re talking about a digital plugin emulation of a piece of hardware, but if you’re talking about a digital eq that’s simply changing frequencies, not attempting to model the box distortions etc, it’s not.

The digital is not inherently inferior in this circumstance. Arguably it could be technically superior.

2 tools, trying to do the same job, different methods based on the same theory, different results. Although it is often possible to match the same result, given enough time and effort.
Old 1 week ago
  #26
Quote:
Originally Posted by Karloff70 View Post
It might not be simulating a specific hardware eq. But it is still 'simulating' what EQ does. It's code. The code describes things we know about what eq does to the sound and then we hear that description. But our descriptions in code are not specific enough yet in some way, or else no space would get lost as it still does, and it doesn't matter whether the plug is a 'virtual analog' plug or a straight digi eq algo. They all steal space. My point is that if we had all the relevant parameters to code there wouldn't be that loss of 'the place'.
This is interesting. What makes it a requirement for an eq change to be done using resistors and capacitors, what makes that more of a “real eq” than something digital (other than it came first).

After all - you can eq a drum kit by sticking a baffle in front of it. Acoustic EQ. Is that superior or inferior to an electrical eq?

It’s funny how our prejudices come into play! My digital eqs don’t seem to “lose space” at all. At least, no differently to my analogue ones.
Old 1 week ago
  #27
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by psycho_monkey View Post
This is interesting. What makes it a requirement for an eq change to be done using resistors and capacitors, what makes that more of a “real eq” than something digital (other than it came first).

After all - you can eq a drum kit by sticking a baffle in front of it. Acoustic EQ. Is that superior or inferior to an electrical eq?

It’s funny how our prejudices come into play! My digital eqs don’t seem to “lose space” at all. At least, no differently to my analogue ones.
What sample rate do you normally run at? I am usually at 48, and have noticed that running some plugs inside Metaplugin to x4 them makes the space stay intact a whole lot better.
Old 1 week ago
  #28
Quote:
Originally Posted by UnderTow View Post
Just a little caveat to your otherwise accurate post: Digital EQs are not inherently simulations of analogue EQs. Both digital and analogue EQs are applying the exact same mathematical principles. One approach does it with electronics, the other does it digitally but neither is a simulation of the other. Don't let the Luddites and ignorami confuse the issue.

Alistair
Right, I was thinking that very thing reading the brief exchanges just above.

Perhaps I was trying to put too much weight on my qualifying phrase, in a sense.

But, definitely, you do a better, more succinct job.


For those having trouble accepting the ideas above, it might be helpful to spend a little time with some basic (analog) circuit design instructional materials -- you will likely quickly note that math underlies the entire process.
Old 1 week ago
  #29
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by theblue1 View Post
Right, I was thinking that very thing reading the brief exchanges just above.

Perhaps I was trying to put too much weight on my qualifying phrase, in a sense.

But, definitely, you do a better, more succinct job.


For those having trouble accepting the ideas above, it might be helpful to spend a little time with some basic (analog) circuit design instructional materials -- you will likely quickly note that math underlies the entire process.
Of course it does. So what. The actual outcomes are still not the same.
Old 1 week ago
  #30
Lives for gear
 

Many forget that with plugs the developers have to deal with the finite limits of real time audio. They have limited CPU processing resources due to hard latency limits of working in real time. As such the developers have admitted they have to cut corners / take short cuts to function within those hard limits. While having an increase in available cores to spread out some of that work you still have the finite time limitations as before and developers still need to have their plugs work well on on older / less cores computers for the widest possible sales.

It's not just about applying impuses of certain settings on hardware EQ to model or mimic it. While you may get actual EQ curves that way, what is not taken into account is the huge amount of variable data of the frequencies different phase and interactions to other frequencies. Anyone who has ever used a phase pedal can see the effect of phase and how we percieve sound frequencies. In fact we even use 180 degree (reverse phase) to cancel out frequencies. So you see the vast amount of variables / calculations a true model / copy of a hardware EQ needed by the plug to perform in a very short finite space of time that is real time audio. Now add to that the other variable of how PCM digital changes the sound also.

Much of this is already covered in prior threads here if you want to get into more specifics. The reality through hands on testing is they will sound different and that is not likely not going to change in the near future so embrace whatever you are using to get the best out of it.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump