VST EQ with before/after curve?

[Phyl]

Sounds like he's looking for a function like iZotope's Ozone Paragraphic EQ.

It will let you generate an eq curve for any recorded material and save it. The intent is generate a curve from some reference material that you like (e.g. commercial CD), save it, then compare it graphically to your own recording. It will even let you force the reference curve onto your recording, but they don't advocate trying for a 100% match.

I've never used the function. I can see it being useful as a learning tool; sort of a feedback mechanism to see if your assumptions about a mix's problems are confirmed.

 

[clong89]

Why don't you ad a spectrum analyzer to the track first. Then add the EQ after the spectrum. Then ad another spectrum after the EQ. It would eat up CPU usage, but I am pretty sure that would be what you are looking for. The spectrum before the EQ will analyze the raw signal and the spectrum after the EQ will analyze the modified signal. You can see both at the same time. Let me know what you think.

Clong

 

[themdla]

well...

i've got a few friends that do sound and post production for movies and frequently use spectrum analyzers to match up ADR/overdubs across multiple recording sessions to good effect.

primarly to match room "tones" between two pieces of dialouge that may have been recorded with different mics, different rooms, different times, etc...

i think the use of a spectrum analyer can be useful in the right situations for EQ'ing (ie two individual tracks that need to sound the "same" or have the "same" EQ profile) but not as the only method for understanding EQ'ing

as everyone else has said EQ'ing is much more about the ears, and i find myself frequently turning off my computer monitor when i'm acoustically "analyzing" my recording performances, level, pan, EQ, reverb, etc...

but i don't think the use of a spectrum analyzer can't be effective it just depends on the application.

ie) i have no doubt that any "modeler" or "plug-in" that copys the sound of an analog piece of gear has gotten it's "sound" by running pink noise through the original analog piece in tandem with a spectrum analyzer used to "read" it's EQ profile in order to determine what effect/frequency profile/EQ that piece of gear has on the pink noise.

my .02

 

[SouthSIDE Glen]

i've got a few friends that do sound and post production for movies and frequently use spectrum analyzers to match up ADR/overdubs across multiple recording sessions to good effect.

That's an entirely different animal. We're talking mixing music here. To analogize visually, it's like the difference between directing a talking heads documentary and directing a movie.

Even then, I gotta be honest and question their methods, because there is absolutely no reason why they should need to use an analyzer for anything other than the occasional troubleshoot (unnles one is trying to impress a boss or client who doesn't know any better. ) I've done that kind of work, and there is no more need to use an analyzer for that stuff than there is in music production.

The reasons for using a spectrum analysis are for troubleshooting specific problems such as pinpointing harmonic distortion for setting harmonic filtering, finding converter aliasing, IDing sub-bass or ultra-high frequency issues that ones ears or speakers hint at but can't pin down accurately, etc.

Think of them like X-rays. An X-ray can be helpful if the doctor suspects a certain issue and needs to look for it specifically (and even then, the X-ray might not be definitive.). And when that happens, they have to have the medical training both to know that the X-ray is potentially a good tool to use to look for it, and the training to be able to read and interpret what the X-ray is actually displaying. But no doctor in the world can tell if a patient is truly healthy in general from looking at an X-ray.

A spectrum analysis is similar. It can't tell anyone if a given signal is generally right or wrong; it's only useful if the engineer needs to look for some specific problem that they already suspect or know is there and need to pin down some specifics. And even then it takes not only a trained eye to read them, but it takes a trained ear to know what to use the analysis to look for.

And determining whether a musical track has the right level of a given frequency to sound right or not is not one of those things that a spectrum analysis can tell us - no matter what those shysters at Hair-Ball will tell us.

G.

 

[themdla]

That's an entirely different animal. We're talking mixing music here. To analogize visually, it's like the difference between directing a talking heads documentary and directing a movie.

Even then, I gotta be honest and question their methods, because there is absolutely no reason why they should need to use an analyzer for anything other than the occasional troubleshoot. I've done that kind of work, and there is no more need to use an analyzer for that stuff than there is in music production.

The reasons for using a spectrum analysis are for troubleshooting specific problems such as pinpointing harmonic distortion for setting harmonic filtering, finding converter aliasing, IDing sub-bass or ultra-high frequency issues that ones ears or speakers hint at but can't pin down accurately, etc.

Think of them like X-rays. An X-ray can be helpful if the doctor suspects a certain issue and needs to look for it specifically (and even then, the X-ray might not be definitive.). And when that happens, they have to have the medical training both to know that the X-ray is potentially a good tool to use to look for it, and the training to be able to read and interpret what the X-ray is actually displaying. But no doctor in the world can tell if a patient is truly healthy in general from looking at an X-ray.

A spectrum analysis is similar. It can't tell anyone if a given signal is generally right or wrong; it's only useful if the engineer needs to look for some specific problem that they already suspect or know is there and need to pin down some specifics. And even then it takes not only a trained eye to read them, but it takes a trained ear to know what to use the analysis to look for.

And determining whether a musical track has the right level of a given frequency to sound right or not is not one of those things that a spectrum analysis can tell us - no matter what those shysters at Hair-Ball will tell us.

G.

i agree with you, however....

in our analogy of the apple painting, a spectrum analyzer can not tell you anything artistic, ie how much red you need in your painting or what quanity of red will make you rembrandt, but it will tell you the "color" of red your dealing with, ie maroon, fire engine, fuscia, etc...

in our musical example it is correct to ascertain that using a spectum analyzer to get that "floyd" sound would not be best practices.

but if you wanted to analyze the "color" or frequency content of an isolated "floyd" guitar track against your guitar track you could certainly see the difference in the frequency content. (provided all things are equal, ie notes and voicings)

that being said, no amount of EQ will make you david gilmore.

 

[RAMI]

but if you wanted to analyze the "color" or frequency content of an isolated "floyd" guitar track against your guitar track you could certainly see the difference in the frequency content

And that doesn't tell you much. So, you're analizing the frequency content. And? There are a million other parameters that make up a "sound", or are part of creating a "tone".

You can have 2 guitars with the exact same frequency content, and they can still sound totally different. You might have the same amount of mids, lows, and highs (visually), but what about attack, sustain, tone, etc....

If a "tone" or "sound" was only about the EQ, then everyone would have a great guitar sound.

 

[themdla]

And that doesn't tell you much. So, you're analizing the frequency content. And? There are a million other parameters that make up a "sound", or are part of creating a "tone".

You can have 2 guitars with the exact same frequency content, and they can still sound totally different. You might have the same amount of mids, lows, and highs (visually), but what about attack, sustain, tone, etc....

If a "tone" or "sound" was only about the EQ, then everyone would have a great guitar sound.

well i think the tone is fundamentally just EQ, but your right a frequency analyzer isn't going to cover everything (sustain, reverb, effects, player's fingers, wind through the trees) but it can get you closer to matching a tone.

again i never said it would get you 100% there... as there is no way to exactly mimic anything transient in life weather it be a recording or any other matter interaction, just ask scientist about entropy or quantum super positioning

but if you wanted to "sample" a color from a painting and ask the guy at home depot to match it so you can paint your house that color i can guarantee that guy is going to use a frequency analyzer (light not sound) to match it, is it exactly the same as rembrandt's red? no. is it closer than just guessing? yes.

if an analyzer get's you 50% there in a short time is there any harm?

actually i'll take it a step further...

you and i are different physically (prolly i have no idea who you are in real life) and more specifically we both have different ear structures, as well as a different history using said ears (i've listened to tons of super loud music).

ie you and i do not hear exactly the same things... frequency content, timber, harmonics, etc...

so if you say a sound is "pink" and i say it's "red" who is right??

so i turn up the mids because i think it's thin and you turn them down because it's too harsh...subjective

1k is 3.8 db is a bit less subjective

i say the analyzer could tell me more about the specfic frequency content of that sound than anybody on HR could just using his ears and guessing.

OTOH, does it matter that 1k is @ 3.8db artistically? prolly not.... for that it's better to use your ears

 

[themdla]

And that doesn't tell you much. So, you're analizing the frequency content. And? There are a million other parameters that make up a "sound", or are part of creating a "tone".

You can have 2 guitars with the exact same frequency content, and they can still sound totally different. You might have the same amount of mids, lows, and highs (visually), but what about attack, sustain, tone, etc....

If a "tone" or "sound" was only about the EQ, then everyone would have a great guitar sound.

also, i think if two guitars had, EXCATLY, the same frequency content, then yes they would sound very much if not exactly the same.

i'm pretty sure that's how modeling amps work, by copying the frequency, compression, gain, and other interactions via real world analyzers of all flavors.

is a modeling amp an exact copy of Jimi's 1969 Marshall stack? no. is it closer than an crate practice amp? yes. could you make your crate practice amp sound exactly like a marshall stack with EQ? no. could you get closer? yes.

 

[SouthSIDE Glen]

but if you wanted to analyze the "color" or frequency content of an isolated "floyd" guitar track against your guitar track you could certainly see the difference in the frequency content. (provided all things are equal, ie notes and voicings)

And seeing that will actually tell you almost nothing of any value. Trying to get your guitar track to generally match the Floyd track will not make it sound the same or even necessarily anything close to it. Not even the same *tone*. And an analyzer won't even get you 50% there.

The devil is in the details, most of which the analyzer doesn't make readily apparent. The ADSR shape of the waveforms, the distribution of resonants and harmonics, the overall envelope, the amount and character of various distortions, etc. None of which are readily visible in a spectral analysis, but all of which in total have as much to do with what you hear as the general frequency distribution, and most of which cannot be replicated by applying a spectral curve match to a different signal.

It's like the difference between your fingerprints and mine. They share a lot of similarities; pretty much the same size, similar general patterns of whorls and ridges and such, even the skin is almost the same color. So what? Unless you match mine exactly in every last detail, there's no way yours would be mistaken for mine. And spectral analysis simply cannot provide that kind of detail for musical tracks.

And let's not forget that the "tone" (god I hate that word, it's such a useless concept) Gilmour picked for any given song he picked *using his ears and brain* to determine what worked the way he wanted it to for that particular song. Just copying that "tone" - even if it were possible using an FFT analysis - which it isn't - does not mean that's what will work for whatever song you're working on. And the only way to determine that is by listening, which bring you right back to square one anyway.

There's a whole lotta people out there that don't want to hear that one has to actually have both musically analytical ears and musically creative ears in order to be any good at this stuff, and that it just doesn't slice any other way, no matter what visual tools one has. But that's the way it is. Ya ain't gotta like it, but ya gotta accept it.

G.

 

[themdla]

And seeing that will actually tell you almost nothing of any value. Trying to get your guitar track to generally match the Floyd track will not make it sound the same or even necessarily anything close to it. Not even the same *tone*. And an analyzer won't even get you 50% there.

The devil is in the details, most of which the analyzer doesn't make readily apparent. The ADSR shape of the waveforms, the distribution of resonants and harmonics, the overall envelope, the amount and character of various distortions, etc. None of which are readily visible in a spectral analysis, but all of which in total have as much to do with what you hear as the general frequency distribution, and most of which cannot be replicated by applying a spectral curve match to a different signal.

It's like the difference between your fingerprints and mine. They share a lot of similarities; pretty much the same size, similar general patterns of whorls and ridges and such, even the skin is almost the same color. So what? Unless you match mine exactly in every last detail, there's no way yours would be mistaken for mine. And spectral analysis simply cannot provide that kind of detail for musical tracks.

And let's not forget that the "tone" (god I hate that word, it's such a useless concept) Gilmour picked for any given song he picked *using his ears and brain* to determine what worked the way he wanted it to for that particular song. Just copying that "tone" - even if it were possible using an FFT analysis - which it isn't - does not mean that's what will work for whatever song you're working on. And the only way to determine that is by listening, which bring you right back to square one anyway.

There's a whole lotta people out there that don't want to hear that one has to actually have both musically analytical ears and musically creative ears in order to be any good at this stuff, and that it just doesn't slice any other way, no matter what visual tools one has. But that's the way it is. Ya ain't gotta like it, but ya gotta accept it.

G.

ok maybe my floyd example is bad so....

what if i went back in time to 1973 when floyd was getting ready to track guitars and...

1) ran pink noise through there rigs with all the mics setup
2) tracked that pink noise
3) flew back to 2010
4) analyzed that pink noise
5) asked gilmore to play the same parts on similar guitars and amps
6) eq'ed the tracks to match the frequency analysis exactly

i think i'd get pretty close then

or should just try to get a different guitar sound

 

[RAMI]

well i think the tone is fundamentally just EQ,

Wow! I disagree.

i say the analyzer could tell me more about the specfic frequency content of that sound than anybody on HR could just using his ears and guessing.

I disagree. And it seems you disagree with your own statement, since you also said this:

it's better to use your ears

So, which one is it? It can't be both.

if two guitars had, EXCATLY, the same frequency content, then yes they would sound very much if not exactly the same.

Another Wow! I disagree.

i'm pretty sure that's how modeling amps work, by copying the frequency, compression, gain, and other interactions via real world analyzers of all flavors.

You just added 2 parameters to your argument. Actually, you added hundreds of others with your vague "and other interactions".

could you get closer? yes.

Closer than what? Closer than using your ears? I would say no. And judging by what you yourself typed here:

it's better to use your ears

...it seems you're not really sure if you're convincing yourself, either.

I'm not trying to be a prick with my "I disagree" all over the place. But, the point is, as long as we disagree on these fundamental issues, neither one of us will convince the other that a Frequency analyzer is useful/useless.

 

[themdla]

haha

ok how about this?

can anyone post a pair of GTR tracks that are the same part but recorded by two different GTR's or two different amps?

i will try to use a spectrum analyzer to get the two sounds close.

both tracks have to either be clean or distorted, not mixed (one clean, one distorted) and no other effects (delay, reverb, comp, eq, etc...)

this should be fun

if i can't i will concede defeat!

 

[themdla]

Wow! I disagree.
I disagree. And it seems you disagree with your own statement, since you also said this: So, which one is it? It can't be both.


Another Wow! I disagree.

You just added 2 parameters to your argument. Actually, you added hundreds of others with your vague "and other interactions".

Closer than what? Closer than using your ears? I would say no. And judging by what you yourself typed here: ...it seems you're not really sure if you're convincing yourself, either.

I'm not trying to be a prick with my "I disagree" all over the place. But, the point is, as long as we disagree on these fundamental issues, neither one of us will convince the other that a Frequency analyzer is useful/useless.

ok one more time to summ it all up!!!

i'm going to play you a GTR track...

by ear tell me EXCATLY how many db of 1k there is @ min 1:23... oh you can't, your not a frequency analyzer you say, hmmm that's odd

OTOH, hey Rami is there too much mids in my GTR for this mix i'm doing in, insert genre here, music??

one is subjective, the other is....

 

[RAMI]

ok one more time to summ it all up!!!

i'm going to play you a GTR track...

by ear tell me EXCATLY how many db of 1k there is @ min 1:23... oh you can't, your not a frequency analyzer you say, hmmm that's odd

OTOH, hey Rami is there too much mids in my GTR for this mix i'm doing in, insert genre here, music??

one is subjective, the other is....

Yeah, I got it the first 3 times you said it. You contradicted yourself too many times for me to even want to keep playing ping-pong here.

An Analyzer will show you a curve. How does that tell you whether there's "too much" of something??? It's just as subjective. So, it tells you how many db of bass there is....so what???? You're going to to turn the bass down on a bass guitar because the pretty little picture shows there's way more bass in a bass guitar signal then there are mids and high's??? hmmm....that's odd.

 

[themdla]

Yeah, I got it the first 3 times you said it. You contradicted yourself too many times for me to even want to keep playing ping-pong here.

An Analyzer will show you a curve. How does that tell you whether there's "too much" of something??? It's just as subjective. So, it tells you how many db of bass there is....so what???? You're going to to turn the bass down on a bass guitar because the pretty little picture shows there's way more bass in a bass guitar signal then there are mids and high's??? hmmm....that's odd.

no duder that's not what i'm saying at all, fuck?

how much bass???? subjective
how to mix it??? subjective

exact frequency content of a bass gtr???? Objective!!!!!!!!!!!!!!!!!!

 

[RAMI]

exact frequency content of a bass gtr???? Objective!!!!!!!!!!!!!!!!!!

I get it. But how does that help you?

Knowing the exact frequency content of a bass is all great and everything...but what you do with that information......subjective.

 

[themdla]

two bass guitars are playing an open low E whole note...

the basses and amps are identical except that one bass amp has a different EQ than the other.

what's the exact difference between the two sounds?

 

[Bristol Posse]

ok one more time to summ it all up!!!

i'm going to play you a GTR track...

by ear tell me EXCATLY how many db of 1k there is @ min 1:23... oh you can't, your not a frequency analyzer you say, hmmm that's odd

OTOH, hey Rami is there too much mids in my GTR for this mix i'm doing in, insert genre here, music??

one is subjective, the other is....

I get what you're saying but I don't get why it matters. Music is more or less entirely subjective.

If it was a universal absolute that guitars would always sound perfect with an RMS value at 1K of -11.8 dBFS when a C5 was played and so on there would be no need to mix at all.
just set up a computer to force all of the instruments into those universal absolutes

I absolutely can't tell you exactly and specifically how much sound (whether it be dBSPL, dBFS, dBa, dBVU, dBu, or dBv) is at any given frequency, and I don't know what the benfit would be if I could anyway since, Peak vs RMS at that frequency, time to reach peak and decay, other frequency interaction and levels as well as odd and even order harmonics to name just a few things that potentially you could see on a spectrum analyzer are all major facters in what makes an instruemnt sound a certain way. so just knowing that in a vacuum guitar 2 had a level at 1kHz of -4.98 dBFS, for 6 miliseconds at 1min, 05 secs 12ms on the timeline doesn't help me at all as far as I can think it through. Heck even running the exact same bar of music through a spectrum analyzer can give you almost limitless different graphs depending on the settings of measurement you use in the analyzer.
What I can say with absolute certainty is whether I like the sound of something, and if I am the one making the mixing decisions, assuming I have learned to trust my room, monitors, ears and instincts, that should be all that matters

 

[themdla]

I get what you're saying but I don't get why it matters. Music is more or less entirely subjective.

If it was a universal absolute that guitars would always sound perfect with an RMS value at 1K of -11.8 dBFS when a C5 was played and so on there would be no need to mix at all.
just set up a computer to force all of the instruments into those universal absolutes

I absolutely can't tell you exactly and specifically how much sound (whether it be dBSPL, dBFS, dBa, dBVU, dBu, or dBv) is at any given frequency, and I don't know what the benfit would be if I could anyway since, Peak vs RMS at that frequency, time to reach peak and decay, other frequency interaction and levels as well as odd and even order harmonics to name just a few things that potentially you could see on a spectrum analyzer are all major facters in what makes an instruemnt sound a certain way. so just knowing that in a vacuum guitar 2 had a level at 1kHz of -4.98 dBFS doesn't help me at all as far as I can think it through. Heck even running the exact same bar of music through a spectrum analyzer can give you almost limitless different graphs depending on the settings of measurement you use in the analyzer.
What I can say with absolute certainty is whether I like the sound of something, and if I am the one making the mixing decisions, assuming I have learned to trust my room, monitors, ears and instincts, that should be all that matters

ok but if you were going to model a steinway in your new line 6 "STEIN" box and wanted to copy that steinway the best you could it would behoove you to analyze it on as many levels with as many real world metrics as possible.

it does not suffice to say... modeling a steinway huh? just add more "highs"

what kind of "highs"?
how many "highs"?

but you could OTOH hand say that....

"you know i've noticed on all the steinways we've measured it seems that C5 really resonates at 12k, i fact if you look at this graph you'll also notice..."

two different ideologies.

how much 12k does it take get good mix in this jazz song???

again that is a question for your subjective ears in which the answer "whatever sounds good bro" will work just fine.

 

[Bristol Posse]

ok but if you were going to model a steinway in your new line 6 "STEIN" box and wanted to copy that steinway the best you could it would behoove you to analyze it on as many levels with as many real world metrics as possible.

it does not suffice to say... modeling a steinway huh? just add more "highs"

what kind of "highs"?
how many "highs"?

but you could OTOH hand say that....

"you know i've noticed on all the steinways we've measured it seems that C5 really resonates at 12k, i fact if you look at this graph you'll also notice..."

two different ideologies.

how much 12k does it take get good mix in this jazz song???

again that is a question for your subjective ears in which the answer "whatever sounds good bro" will work just fine.

Yep I get you
But in the context of a mix rather than modelling a perfect steinway C5, you may need to Change a few of those fundamental things the make a steinway sound steinwayish (either modelled or recorded for real) so that it works with everything else. And whether, for example, it's competing with vocal clarity at 3.5kHz and needs to be cut, or is complementing it is really a subjective decision based on how the mix sounds to the person doing the mix and not how the spectrum looks whether you look at the spectrum of the entire mix or the individual components

As an aside I think we've collectively managed to frighten the OP away