Help with Multi-Band Compression

  • Thread starter Thread starter Seafroggys
  • Start date Start date
Seafroggys

Seafroggys

Well-known member
Okay I really need help understanding how to use Multi-Band Compression. I know what frequencies I want to compress, but I don't know how to target that frequency. I use the GMulti plug-in.

http://www.gvst.co.uk/gmulti.htm
 
From the looks of it it's a 3-band compressor. Freq 1 sets the crossover frequency between the low and the mid bands, and Freq 2 sets the crossover frequency between the mid and high bands.

So, in the screenshot, the low band will be set to everything below 200Hz (marked by Freq 1), the Mid band is set to be between 200-2000Hz (marked by Freq1 and Freq2) and the High band is set to be anything above 2000Hz.

Question is... why do you need multiband compression?
 
Actually, I figured it out as soon as I loaded the webpage and looked at the manual....that always helps ;)

As for using it, I use it during the "mastering" process when I treat the stereo file to tame some frequencies that get out of control at times. For example, with the demo I was just working on, the toms are overbearing in the 80-140 hz range, so I just apply some compression there. If I did EQ, that drops everything down, which I don't need since the rest of the mix is fine.
 
So if the toms are giving you grief, why don't you sort them out in the mix, before you get to the 'mastering' stage? It's a lot better to fix the individual tracks rather than mangling the stereo mix
 
^He's very right. remember that you're compressing ALL frequencies now, and not just the toms...
 
actually, its not so much the toms that are giving me grief, its the frequencies themselves that give me grief. The toms sound fine and sit well in the mix, its just everything up to like 180-200 hz gets pretty hot at times so I need to tone it down to even it out to make it more comparable to the other music I listen to (which tends to be mid-upper mid ranged in dominance).
 
Even with all the discreet bands on that thing, It's very rare that you can find a frequency range that will let you correct one thing without stomping another. Go back and fix it in the mix.

It is a tool of last resort.

If you are trying to fix it because the mix is fighting you , remember what the mastering engineer is up against , working only with a 2 track mix....

Bob Olhsson's classic quote,


"Mastering is the art of balancing objective degradation against subjective improvement."

:D
:D:D
:D:D:D
 
^He's very right. remember that you're compressing ALL frequencies now, and not just the toms...

The point about it being better to fix the toms on their individule or group tracks vs trying to fix them in the Stereo mix is of course right on. I just want to address the above statement, because it goes to the core of MBC abuse. Used correctly for this application, this statement would not be true. Only a single band of compression should be employed, the width no wider than necessary to encompass the offending frequencies, with the rest of the spectrum left flat. It is true that all frequencies within the band would be compressed for the duration of the threshold to release time, which can be bad enough. But worse is the common notion that you "need" to use all the bands, just cause they're there.
 
In my experiments with multiband compression, it seemed as though individual tracks may sound great, but it is impossible to predict exactly how the frequencies will constructively or destructively add in the MIX; so the multiband compression really does serve a very valuable tool for "naturalizing" and even-ing out the sound of the mix almost automatically, that would be almost impossible if trying to track down the sources in the individual tracks.

If you have a multi-band compressor that shows you how much compression is actually being applied, such as on a Behringer DSP9024, or the effects cards of a Roland digital recorder, you can set the compressor levels such that there is no compression most of the time, and it only comes into effect when the levels of any of the chosen frequency bands exceed some chosen level. As usual, the choice of level, and how you like the sound are all just matters of opinion. The MB Comp is a very valuable tool.

In this scenario, all the bands are "on" all the time, and all frequency bands are being monitored. But they are only actually being COMPRESSED when the spectral energy level of each band exceeds the chosen threshold. The thresholds of each band must also be individually set. The Roland units have three stereo bands, while the Behringer has nine. Both units were valuable. The Nine independent stereo bands of the Behr take a long time to set up, but they really work exactly as supposed to once set. Those are just examples, you would have to work with the settings available on your MB comp.

Good luck.
 
In my experiments with multiband compression, it seemed as though individual tracks may sound great, but it is impossible to predict exactly how the frequencies will constructively or destructively add in the MIX;

Why do you need to predict? You are hearing it in real time as you monitor the mix buss.

so the multiband compression really does serve a very valuable tool for "naturalizing" and even-ing out the sound of the mix almost automatically,

Danger Will Robinson!

....that would be almost impossible if trying to track down the sources in the individual tracks.

Not impossible at all, just more time consuming. I'm usually the resident proponent of the much maligned MBC, but I don't advocate using it as a shortcut to avoid properly defining problems, tracing them to their source, and treating them at the source whenever possible.
 
Why do you need to predict? You are hearing it in real time as you monitor the mix buss.
What you are hearing is THE MIX.

If you then feel like there is about 20 mSec of too much 4.4 kHz at 2:21 in the mix, and .5 sec of too much 192 Hz at 3:14 in the mix, and....

Do you take extensive notes ?
Or do you tackle each instance individually ?

Then you still really don't know which instrument caused the frequencies.....was to the bass ? Or was the bass still resonating when the drummer hit the toms, ? Maybe it was a temporary interaction of the bass, drums, baritone sax, and vocals. etc..etc..etc...How acn we blame any one of them as being "a problem" ? The digital waveform of the mix doesn't know "who" caused the "problem".

It can all certainly be done your way, and maybe you are so artful that your mixes end up sounding great...that is all a matter of taste.

But a properly set up MB comp realloy can do exactly the same thing on autopilot in one pass. I would also claim that the digital automation actually is less error prone and more thorough simply because it removes the human error. Should you correct for 30 mS ? ....what it it needed 33 mS ? ...or only needed 27 mS ? The Comp dtermines this with digital precision, while your ear will have a hard time deciding exactly when the start and cutoff times should be.

Bottom line is that whichever ends up sounding good to you really is good. But there is nothing about the MB Comp method that is going to cause anything "bad" to happen, unless you simply choose inappropriate settings. Of course chosing inappropriate settings is easy to do, just like making a bad choice in the "hunt and correct" method.

Of course there are lots of musical classics recorded with NO digital technology at all, and very little "hand" adjustments in the mix,....just live sound and they sound better than anything I can create. :D I just want to point out that once you enter the world of making precise "corrections" to audio data, the automated way achieves more consistent and accurate results.
 
Last edited:
Of course there are lots of musical classics recorded with NO digital technology at all, and very little "hand" adjustments in the mix,....just live sound and they sound better than anything I can create.
This is the loose thread that unravels the whole thing for me every time: MBCs are only about 10-15 years old (they grew out of modern hearing aid technology.) This coincides with two phenomena; the rise of modern home recording, and the collapse of commercial grade audio quality.

How is it that for almost an enture century the recording industry got along just fine without having to lean on MBCs? And how is it that for just as long those that make the classics knew that one of the secrets was to create tracks thatvirtually mix themselves and require practically nothing in the way of sonic shaping after mixdown?

And how is it that just when tools come along that make it "easy" - translation: where one can seemingly phone in the tracking and mixing and just fix the 2mix - happens to coincide in time with the largest degredation in the quality of commercial music production inthe past 100 years

The fact is that yes, most of those who don't lean on MBCs CAN tell what to do in the tracking and mixing in order to fix *or avoid altogether* problems that show only in summing. The problem with the regular use of MBCs on the 2mix to fix problems in the tracks is that it allows one to become sloppy with the tracks.

The general rule of thumb always used to be that the only real work done on the 2mix was stuff like fixing fades, setting final song lengths and levels, and setting some EQ curves as required by the master medium. Except for maybe a little shaping EQ, the 2mix was pretty much sonically set. Frontloading the burden of sound quality was always what worked best, and still is. OTOH, the new trend of relying on advanced tools like MBCs to try and fix bad tracking and mixing in the shrink wrap tends to becone a crutch that makes one slack just that little bit more on the front end of the process, and the overall results will never be as good.

G.
 
This is the loose thread that unravels the whole thing for me every time: MBCs are only about 10-15 years old (they grew out of modern hearing aid technology.) This coincides with two phenomena; the rise of modern home recording, and the collapse of commercial grade audio quality.

How is it that for almost an enture century the recording industry got along just fine without having to lean on MBCs? And how is it that for just as long those that make the classics knew that one of the secrets was to create tracks thatvirtually mix themselves and require practically nothing in the way of sonic shaping after mixdown?

And how is it that just when tools come along that make it "easy" - translation: where one can seemingly phone in the tracking and mixing and just fix the 2mix - happens to coincide in time with the largest degredation in the quality of commercial music production inthe past 100 years

The fact is that yes, most of those who don't lean on MBCs CAN tell what to do in the tracking and mixing in order to fix *or avoid altogether* problems that show only in summing. The problem with the regular use of MBCs on the 2mix to fix problems in the tracks is that it allows one to become sloppy with the tracks.

The general rule of thumb always used to be that the only real work done on the 2mix was stuff like fixing fades, setting final song lengths and levels, and setting some EQ curves as required by the master medium. Except for maybe a little shaping EQ, the 2mix was pretty much sonically set. Frontloading the burden of sound quality was always what worked best, and still is. OTOH, the new trend of relying on advanced tools like MBCs to try and fix bad tracking and mixing in the shrink wrap tends to becone a crutch that makes one slack just that little bit more on the front end of the process, and the overall results will never be as good.

G.

+10000

Like I was told as my hand was being slapped away from the 33609 at the START of the mix "mix with the faders, NOT the rack"
 
This is the loose thread that unravels the whole thing for me every time: MBCs are only about 10-15 years old (they grew out of modern hearing aid technology.) This coincides with two phenomena; the rise of modern home recording, and the collapse of commercial grade audio quality.

I think we may be confusing some concepts here:

I belive that:
1. The REAL "quality" of an audio recording lies in the EMOTIONAL CONTENT of the PERFORMANCE. The ARTIST controls what all listeners will call the "quality" of the recording. Lots of GREAT recordings have been done with very "poor" quality audio equipment. They still SOUND great.

2. Having said that, I believe that if the same artist were to record a classic, done on "old" "lower quality" technology" on new technology, it would be a "better" audio quality. Or at least it could POSSIBLY be a better audio quality. See #3:

3. The current degradation in commercial audio quality is because consumers don't demand any higher. They buy overcompressed rap and techno, and listen on MP3 players ! That has nothing to do wth what audiophiles COULD produce, and appreciate, with modern digital technology if motivated to produce it.

4. So, in light of all the above, I am only providing my opinion that the MB Comp, with proper settings, can automatically accomplish control of audio defects such as were originally requested by the thread starter, who wanted "to tame some frequencies that get out of control at times".
The MB Comp will do that faster and more accurately than any human mastering ear possibly can. If you have it set such that it is changing the funmdamental mix or the overall tone of the mix, then you simply don't have it set correctly to accomplish the goals stated in this scenario.

I am also horrified by the ABUSE of compression. But the useage of any tool is an artistic decision....neither inherently good or bad. They all have pitfalls and can be misused, abused, or simply fouled up.
 
4. So, in light of all the above, I am only providing my opinion that the MB Comp, with proper settings, can automatically accomplish control of audio defects such as were originally requested by the thread starter, who wanted "to tame some frequencies that get out of control at times".
I don't think I actually seriously disagree with any of what's in your post.

What I am saying , related to your point 4, quoted above, is that the key is to aviod or remove the audio defects from the get go, not just accept them and fix them in the shrink wrap. OK, sometimes that can't be avoided if someone else did the mixing and tracking. But somehow, that didn't seem to be a problem before MBCs came along, did it? We got along just fine without having to use them after the 2mix was done when they didn't exist, which was not that long ago at all.

It's the creeping dependance on their use in the 2mix as a crutch that is slowly eating away at the quality that is put into the tracking and the mixing. The idea that if one hears an anamoly in the mix, that they then can't hear to fix it in the individual tracks is ludicrous. For decades, engineers had no problem hearing what needed to be done. But now, the home recorder with the MBC thinks he/she doesn't have to hear it, because they think it's easier to just fix it later. The problem is, that just lets their ears go slack, and the quality of the tracking and the mixing gets worse, not better. And when that happens, no MBC in the world is going to bring back any sonic potential that was lost due to lacksidasical tracking and mixing.

You're right, the decline in quality over the past 15 years is due to fad - the idiotic RMS wars - and not due to the gear itself. Ironic how we have more dynamic range and better S/N at higher resolution available to us than ever before, yet everybody wants to squash their entire production into the last 10dB with 3 dBs of clipping.

It may have been wrong of me to imply that tools like the MBC are the *cause* of such idiocy, just like it would be wrong to imply that AK-47s are the cause of radical jihad. But that is sure where they idiotically get used the most. There's nothing wrong with MBCs in and of themselves anymore than there is anything wrong with a rifle in and of itself. It's how it's used, misued, and abused that needs to stop. And the general idea that regularly using MBCs to essentially do in premastering what should have been already done in tracking or mixing is a malpractice that needs to stop.

Are there times when there is no other choice, or is maybe even the best choice under bad circumstances? Sure. I'm not arguing that. I'm arguing that the disease of the idea that it should be standard engineering practice is rotting the practice from the bottom up.

G.
 
I dont think we should be blaming the consumers for the reduction in audio quality. They seemed perfectly happy to pay more and more to get better stuff thru the ages.

I didnt see broadcast quality go down until I saw a RADICAL increase in pretengineers and an exponential increase in the number of pretengineer magazines and schools

I have never had a consumer ask me if I could make the extra money they spent on their hifi system worthless by making a crappy album
 
What I am saying , related to your point 4, quoted above, is that the key is to aviod or remove the audio defects from the get go,...

I don't think we disagree;
but I think we are making different assumptions about what is going on and what the question the original poster is asking.

Certainly, if there are audible defects in an individual track, that should be removed "from the get go". I have assumed that is already done, or at least is done by most recording engineers. You seem to be assuming that it has been ignored (but I don't want to be making statements for you; that it just the way I am interpreting your statements.)

But the situation also occurs that perfectly fine tracks, none of which is really a problem, add together in unpredictable ways to form unexpected phenomenon, such as a period of boominess harshness IN THE MIX. The problem could not be removed on the track, because it simply was no defect in the track. The problem only exists in the mix. That is what I am assuming the poster is referring to.

Although the MB comp could be successfully used to lightly compress only problem areas of the individual track also....again, the key is the appropriate settings.

I am just pointing outy that PROPERLY SET, the Comp is doing exactly the same thing that your hand on a fader would be doing ! It is just doing it with digital 96 kHz sampling rate precision and 24 bit amplitude accuracy. Maybe your ears and fingers are equally accurate :D

This is all pretty theoretical and philosophical. I like to use compression VERY lightly, such that it is really not an audible "effect" at all. If what you are doing sounds good, it is good.

I dont think we should be blaming the consumers for the reduction in audio quality. They seemed perfectly happy to pay more and more to get better stuff thru the ages.
But the simple fact is that DVD-A and SACD (superior audio formats) are struggling to find consumers willing to pay, yet MP3 is thriving ! (I personally can't stand the sound of MP3. Why even bother !)
I didnt see broadcast quality go down until I saw a RADICAL increase in pretengineers and an exponential increase in the number of pretengineer magazines and schools
I actually believe that modern audio engineers, in general, have a much deeper understanding of what is really going on than the engineers of prior eras. Are they as "artistic" ? That is purely a matter of opinion.
Rap and Hip-hop (usually very poorly engineered) sell because people buy it ! If they were demanding higher quality, it simply wouldn't sell.
I have never had a consumer ask me if I could make the extra money they spent on their hifi system worthless by making a crappy album
But I have heard numerous stories of artists wanting their mixes louder an more compressed !

....and a lot of consumers boost the bass to the point where the sound bears no relation at all to the way it was originally "engineered" :D

Bottom line: I think every generation has had far more "pretengineers" than great artists. But we forget the crap, and only remember the great works. So it SEEMS like there is a lot more crap these days and only good engineering in the past. Also, vinyl was simply more forgiving of errors than digital, so a pretengineer of bygone days could still produce something that was at least listenable. :D A modern engineer, armed with a bank of compressors, can produce homogenized white-noise with a beat.
The sad thing is that people still buy it.
 
Glen and Axis.
Not to buttkiss but as usual (mostly Glen but Axis is new to me) I have highlighted and printed your posts into my scrapbook at the studio. Glen we've butted heads a couple of times but your knowledge is one of the most valuable on this sight. Kudos!
 
I don't think we disagree;
but I think we are making different assumptions about what is going on and what the question the original poster is asking.

Certainly, if there are audible defects in an individual track, that should be removed "from the get go". I have assumed that is already done, or at least is done by most recording engineers. You seem to be assuming that it has been ignored (but I don't want to be making statements for you; that it just the way I am interpreting your statements.)
Of course I have no way of knowing which is the case in this specific instance. But what I can say with pretty fair confidence - based upon my own experiences talking with budding and even somehat experienced home recordists on the past few years, both in person (as both associates and clients), and on this board - is that there is a pretty strong trend towards a)the belief that sophisticated gear can effectively replace engineering technique, and b)that most rookies just don't grok the idea of frontloadling importance in the process (i.e. that performance trumps tracking, that tracking trumps mixing, and that mixing trumps mastering.)

When I was growing up and cutting my teeth in this racket. the big Bozo-no-no catch phrase was "fix in the mix". You could always tell that when someone said, "We'll fix it in the mix", that it meant one of two things; either the guy was a pro under a time or budget gun and just had no options, or much more often, that it was a rookie that didn't understand the importance of getting it right in the tracking first.

IME, that has now evolved in the past 10 years to not only "we'll fix it in the master", but even "we'll MIX it in the master." If I had a buck for every time in the last two years on this board I saw or heard someone who basically phoned in the mixing by compressing every track flat, layering them like lasagna, and then taking their MBC and kneading the hell out of their 2mix to try and accomplish what they should have accomplished in mixing, I'd be able to buy a new premium gold channel of analog mic and pre for my rig ;).

Now I know that's not quite what you're describing with your technique at all, and I never meant to imply that. You're taling about a light sanding, not a complete mangling. And that's not bad. I'm not convinced that it's necessary, for the very reason that it hasn't been necessary for many decades until now, but it's not an awful abuse.

It just amazes me how most of the best-sounding productions in history ever to stick to disc or tape were accomplished without even the invention of MBCs or even the need to so anything other than the actual technical premastering itself in premastering, yet today they are considered a gotta-have go-to device to be used at a point in the production process that used to be considered as downstram of that kind of alteration. We certainly haven't seen an improvement in production value by that change. A change, BTW, that never really came from the pros, but really was started mostly that the grassroots home recording level.
But the situation also occurs that perfectly fine tracks, none of which is really a problem, add together in unpredictable ways to form unexpected phenomenon, such as a period of boominess harshness IN THE MIX. The problem could not be removed on the track, because it simply was no defect in the track.
But the mix is nothing but the results of the mixing process. If the tracks sound fine, but the mix doesn't, that means there's something worng in the mixing process or inthe tracks. You can't blame the 2mix itself, it's the tracks and the process that made the 2mix that are at fault, and it's there where they should be addressed.

A basic tenant - I think you may have actually said it yourself earlier, I'm not sure - is that one should not judge how a track sounds solo, but rather how it sounds in the mix. To say that the individuaal tracks sound "fine" is meaningless when you know they don't sit right in the mix. The mix engineer should have good enough ears and knowledge to know which component tracks are contribution to a problem in the mix, even if they sounds pretty OK solo, From there it's a simple matter of adjusting the mix parameters to account for it.

Usually, IME anyway, it's a case of concurrent peaks on seperate tracks which are fine alone, but when summed are too pronounced. The solution there is simple; manually or via automation, knock down some individual peaks by only a dB or two (high threshold compression or limiting usually doesn work, because the individual peaks are relative harmless and perhaps even invisible until the mix). The real advantage here is that the individual changes to the tracks are virtually inaudible, only changing a peak or two a dB or two, and not targeting any frequencies to change any timbre on the track(s). But the end result is a tamed mix where the harsh frequency is no longer bothersome and no loger requires frequency-based dynamics on the overall mix.

Now, you might say that your way is a little less time consuming and a bit easier. And you might even be right. But it also is a lot less skillful. I don;t mean that in a negative way specifically agianst your technique; what I mean is that it requires less ear, less intimacy with the tracks and less mixing skill. All of which, IMHO, leads to a general degradition in mixing skills altogether. Better to get one's hands and ears and even eyes into the open chest of the mix, than to antiseptically spray the mix's wound with an MBC after it's sewn back up. I may be old school by saying this, by I truely feel that it makes AND keeps one a better engineer AND does a better job on the patient..
But the simple fact is that DVD-A and SACD (superior audio formats) are struggling to find consumers willing to pay, yet MP3 is thriving !
That's because, unfortunately, convenience trumps sound quality...at least for a while. MP3 players are the 8-track tape players of the 21st century, and even moreso when their sound quality is, while still crippled, a couple of magnitudes better than 8track ever was, and when one can hold the equivalent of an entore record store full of 8-tracks (or CDs or SACDs) on one little hard drive the size of a couple of Ritz crackers, people would be foolish to move to the SACD format..
I actually believe that modern audio engineers, in general, have a much deeper understanding of what is really going on than the engineers of prior eras.
That I'm not so sure of. I doubt most of today's engineers could do D2D that sounded half as good as the boys at RCA and the BBC could do with 1/10th the gear. Of course course, it doesn't help that there's no musiciaship anymore either.
Ridgeback said:
I have highlighted and printed your posts into my scrapbook at the studio. Glen we've butted heads a couple of times but your knowledge is [edited out of false modesty :p]
Thanks, ridge! Very kind of you.

And I hope you realize that if we did butt heads (I honestly don't recall), I am usually doing it as a terse method of trying to figure out what the real deal is, and not out of any kind of animosity. Sometimes I learn things best from debate.

Ahhh, I know you're just keeping that at your studio so you can blame me when something goes awry: "Hey, it's all Glen's fault!" At least that's what everyone else does :p ;) :D

G.
 
It's a tool. Use it when necessary.

People "got by just fine" with tier dynamics of the organ and the harpsichord before the piano was invented. But try playing Rachmaninoff on a harpsichord... on the other hand, I hate hearing Bach on the piano. Yuck!

Sooo, people used to get by w/o the MBC during the Beatles era... People also made a lot of lopsided and "weird" mixes once Stereo just came out... I mean drums in ONLY the Left channel? Hurts my ear! Just as they abused Stereo for fun when it first came out, people are now abusing MBCs just because it's new. Eventually we will all come to realise yet again that less is more, and there is nothing inherently "evil" in using MBCs when done "properly".

Now, if you'll excuse me, I am going to go abuse some audio. Maybe I'll even strangle it with some MBC followed by Guitar Rig where I have an LFO modulating the size of the speaker cabinet in rhythm... try THAT with your hardware Mess Boogger.
 
  • Like
Reactions: XLR
Back
Top