Compressing in "Stages"

[audiovisceral]

Compressing in "Stages"

Stage 1: Adding Control Compression
The purpose of this compression stage is to control any sudden high-level peak that could cause the average level to be brought lower. This compression has much the same effect as "tape compression" in analog recording, and can be omitted if the mix was originally recorded on analog tape.

We need somewhat of a soft knee compressor will be used with a compression ratio moving towards limiting (5:1 to 9:1). Attack time should be fast (1 - 2 ms.) and the release time medium-fast (40 - 50 ms.). When we set the threshold so that gain reduction peaks to -3, so that it effects only the loudest peaks. Makeup gain can be set to restore level after compression (at least 2.0 dB). For mixes that have uncontrolled peaks, the higher ratios (8 or 9 to 1) should be used; slightly higher gain reduction could be allowed on these peaks, as long as there usually is no more than 3 dB reduction.

Stage 2: Adding Mild Compression
A mild amount of compression will give us general increase in apparent volume with very little objectionable effect on the dynamics of the tune. Done properly, the mix will often have twice the apparent level, especially in a mix that has a lot of musical dynamics.

When we say mild, we mean both a low compression ratio (between 1.5:1 and 2.0:1) and mild amount of gain reduction (6 dB). The attack time should be fast, but not at 0 (one or two milliseconds so that the compressor begins to compress right after any transient attack begins. We will be adding additional dynamics processing to control the initial peak of a transient, which is usually formed in the first 2 ms. The release time should be medium (100 ms. range) or a variable-release can be used.

We will be using a plug-in made by T.C. Electronic** called MasterX.** The compression and dynamic processing in this plug-in is done in three separate bands. The bass, midrange, and treble ranges are split, compressed separately, and re-combined back into one signal. With this kind of arrangement will prevent most compression pumping problems and make it much harder to hear the compressor working. .

The plug-in will automatically apply makeup gain after the gain reduction so that the original peak level is attained after compression. As you move the threshold slider down you will hear an increase in the volume of the lower-level passages and instruments, plus an increase in the apparent level of the mix.

Stage 3: Adding The Final Loud Edge
Final increases to the apparent level will occur by applying limiting to control transient peaks. We will often use a Waves** program called L1 Ultra-maximizer to accomplish this task. In previous compression stages we only controlled peaks after 2 ms. With the maximizer the attack time will set to "0" ms, and the release time at the ultra-fast setting of 1.0 ms. With these time parameter settings, the transient peaks will be literally sheared off. If you do not have a Waves Maximizer, use a fast-acting limiter and adjust the attack/release times according to the above times.
Care must be exercised in applying this limiting. Since it is literally clipping the signal, there is perceptible distortion being added. A little bit can cause a nice "edge" to the mix and too much causes it to sound dirty. The changing gain-reduction metering is fast-acting, making it hard to read the amount of gain reduction being applied. Fortunately, there is a maximum gain reduction figure (in dB ) that is displayed below the metering. It is a good idea to set the threshold for 5 dB or less gain reduction at this stage, although the ears should be your guide.

Don't expect to become an expert in one lesson. Carefully following these guidelines, however, should greatly improve your final sound.

Specifically, I am curious about attacks. Is a 1-2 ms attack the standard for average, modern music mastering purposes (where 2 would be a relatively slow attack and 1 would be relatively fast)? It seems a bit quick to me ...

What average range do you tend to work within?

Are there any constructive comments or suggestions to add on other points?

So far that passage has been the best resource I've come across for how, in general, to practically compress/limit music.

 

[Michael Fossenkemper]

welcome to the forum. It is generally good practice to credit a quote so people know who it belongs to.

Compression is kind of like clothes, people have different tastes and depending on the compressor, diifferent approaches work better. I personally don't use compressors as they describe. But that doesn't mean my approach is better or worse. The best thing to do is try it and take with you the things you like about it and modify the things you don't. your eq, compressors, limiters and whatever else you throw in, affect each other. If you place an eq before the compressor, then what worked before might no longer be working.

 

[audiovisceral]

The reference:
http://www.alexandermagazine.com/recordingeq/weeklytip/06tip01-3b.asp

I understand that everyone seems to have a different way of doing things. It becomes difficult to discuss or understand though unless people are willing to talk in specifics or default/average approaches ie. starting points for typical mixes. My music teacher in high school always used to say "First you learn the rules, then you learn to break them."

General approaches like the one I quoted above are a necessity for beginners to learn the basics. I've experimented (and read) like crazy, but without a default approach, I've been essentially attempting to re-invent the wheel. When it comes to mastering, I very clearly realize I'm not (yet) that great an inventor.

So far, this has been the most effective method I've tried for compressing/limiting. I find a single multiband/wideband comes out as too harsh and doesn't give the fullest volume, though this might be my technique. Additionally, I find that using attacks in the range I used to (20 ms or so) doesn't sound as smooth as a 1-2 ms approach.

For me, this means:

1) BBE Maximizer - exciting treble/bass as is needed
2) EQ - balancing as is needed
3) TC NEss Single-band Compressor - Soft knee 'control' compression (3 db)
4) Sonitus/LinMB Multi-band Compressor - 5 bands of equal threshold 'mild' compression (6 db)
5) L3 Limiter - 5 bands of equal threshold limiting (6 db)

Sounds satisfactory so far. But I remain curious what other 'default' approaches or modifications might be recommended. Additionally, I am still curious about the attack and whether 1-2 ms truly is usually the norm.

Any takers?

Let me put it this way:
If you were setting up a basic compression/limiting chain to use on EVERY modern radio-type song from now until the end of time, what would you do with it, including approximate attack/release/attenuation settings?

 

[pingu]

Hi mate.


Way too much processing there.

The goal is to minimise the amount of processing to gain the maximum benefit.

If what you are trying to master isnt coming together with a little eq and compression then you might have to relook at the mix/tracking.


I seem to use attack times between 50 and 120 ms but YMMV.

1-2 ms is way too fast for single band mastering compression.

If you learn how to set a compressor correctly you will find that one size does not fit all.

Say you did set the compressor correctly, by then applying multiband afterwards you are only asking for trouble.

Are you mastering your own music?

If so post a link and ill have a listen to the tune you are trying to master and i can give my thoughts as to weather its salvageable or not.

 

[audiovisceral]

Thanks for the input. I have a lot of songs of different genres - some instrumental, some vocal complete. All are for my own purposes than commercial release, so perfection is not necessary. The best radio-single-style results/approach I can manage are all I am aiming for.

Here's a simple song with two approaches. Suggestions to improve further are very welcome.

Eg. What do you think about the 1 ms attack on my multiband in the first example? I feel like I could be doing something to get more power out of this as a whole. hmm ...

What would you do with the source file? How do you expect something like Outkast's Hey Ya was compressed by comparison, in terms of settings?


1) LOUDEST AND PUNCHIEST
http://www.audiovisceral.com/av_f2b_1i.wav
1) BBE - 7 low, 6 high
2) Sonitus Equalizer - 6 band touch up
3) TCNess X (single band comp) - 5 attack, 40 release, -3.5 threshold, 2.3:1, softknee
4) Sonitus Multiband Compressor - 1 attack, 88 release, -30 threshold, 3:1 ratio, 4 dB knee, 12 dB peak atten in low freq, 3-6 dB in the high
5) L3 Multimaximizer - ~6 dB peak atten


2) SMOOTHER LESS PERCUSSIVE
http://www.audiovisceral.com/av_f4l_full_processing.wav
1) BBE - 7 low, 6 high
2) TCNess X (single band comp) - 2 attack, 45 release, -3 threshold, 7:1 ratio, softknee, 3 dB peak atten
3) Sonitus Multiband Compressor) - 20 attack, 95 release, -30 threshold, 3:1 ratio, 4 dB knee, 6-9 dB peak atten
4) L3 Multimaximizer - ~7 dB peak atten


3) RAW, UNEDITED
http://www.audiovisceral.com/av_f4l_raw.wav
Zero processing.



mike@audiovisceral.com

 

[heathen22]

As said each track should be listened to for problems which are apparent,there is no single chain setup which will work everytime. Though a good starting reference point is needed I agree but this comes with practice. 1-2 ms is way too fast for me too on 99% of songs.
I usually use multiple stages of compression but very mildly.Say only 1-2 db from each unit. Quite often (not always) I'll use a digital software eq for surgical eq if anything in a narrow bandwidth is really bothering me then out to an Avalon ad 2055 eq (or other)then to an Al smart c2 comp and then to a Pendulum valve limiter. C2 is usually set to 10ms attack and 100 release @ 1.5-1 ratio with as minimum as possible GR say 1-2 db,Pendulum is used to control slower more energetic transients such as bass and any other loud peaks,20- 30 attack and maybe 300- 400 release.With only 1-2db GR
Attack, release times are all dependant on the program material,so is the ratio.All units without denteted pots should be calibrated before use
After this process if someone still wants me to belt the hell out of the mix with an L2 I will,but prefer not to.

 

[dcollins]

The reference:
http://www.alexandermagazine.com/recordingeq/weeklytip/06tip01-3b.asp

What a bunch of crap!

A formula indeed.................

DC

 

[pingu]

It seems ive underestimated you.

Nice song and nice work.

 

[Michael Fossenkemper]

I would love to tell you there is a formula but I personally don't have one. If you listen to your raw mix, and then listen to your processed mix's, what do you hear right off the bat? besides it being louder of coarse.

You destroyed your mix. all the vide turned into a compression pumping nightmare.

So the first thing you need to do is take that forumla and throw it away.

take a simpler approach to begin with. don't mistake louder as better. Listen to the result of what you do. I personally would like to listen to something a little quiter and retain the intention of your mix. It takes some time to learn how to listen to what you are doing with stereo processing. use less compression, slow those attack times down.

 

[pingu]

I would love to tell you there is a formula but I personally don't have one. If you listen to your raw mix, and then listen to your processed mix's, what do you hear right off the bat? besides it being louder of coarse.

You destroyed your mix. all the vide turned into a compression pumping nightmare.

So the first thing you need to do is take that forumla and throw it away.

take a simpler approach to begin with. don't mistake louder as better. Listen to the result of what you do. I personally would like to listen to something a little quiter and retain the intention of your mix. It takes some time to learn how to listen to what you are doing with stereo processing. use less compression, slow those attack times down.

Its good of you to download and comment.

cheers.

 

[audiovisceral]

You destroyed your mix. all the vide turned into a compression pumping nightmare.

take a simpler approach to begin with. don't mistake louder as better. Listen to the result of what you do. I personally would like to listen to something a little quiter and retain the intention of your mix.

Volume and compression are, as most ME's would say, artistic decisions. In my case, I really don't want anything quieter. I would like to retain/increase the volume with a better sound and minimal pumping.

When the first mix is listened next to standard pop successes of the past few years (Hey Ya and Black Eyed Peas - Shut Up were my two references), it is comparable, but quiet. I am posting for advice on how to best obtain those kinds of levels.

One idea that's in my head is considering that the single band compressor might be one of the bigger contributors of wideband thus audible pumping in my chain. Two multiband compressors with the same crossovers might increase separation between the activity of the five channels, but minimize this effect ...

What would you do for a client who wants hyper-compression, after you've explained to them why they shouldn't, and they are adamant about wanting it (besides send them away)?

What would you improve in my approach to maintain volume/compression, but better manage adverse effects?

 

[headchem]

This is a messy arena, but I'll put my 2 cents in...

First of all, like the others have said, take everything I say as a possible starting point and then spend the next few months experimenting until you can really hear what's going on. Here are some basic starting points.

For songs with many transients (transients being bass drum thuds, snare hits, spikes in volume, etc) I like to use attack times between 30-60 ms, release times between 150-300 ms. Ratio of 2:1 to 3.5:1 Gain reduction no more than 6 db, but preferably between 2-4 db reduction. This won't get you commercial volume, but just wait. I'll get to that.

For "smooth" songs without those transients, I prefer a very slow attack time from 60-120 ms, with equally slower release, from 250-400 ms. Low ratios from 1.3:1 to 2:1. Again try for no more than 2-4 db gain reduction. You'll have to set your threshold much lower for this one.

Now for volume. I use a maximizer / limiter. The limiter comes first, and chops the occasional peak by maybe 1 bd. Now pass it on to a soft clip maximizer. These have an output volume knob you can turn up until you hear the music distort, but the soft clip function adds pleasant harmonic distortion, which is much nicer sounding than digital distortion, which would occur if you just increased gain past 0db in the digital domain.

If you listen to the commercial tracks you mentioned, you'll hear they've been severely compressed, and are very distorted... Record companies like music LOUD, even if it means sacrificing all dynamics and sound clarity. A classic case of those in power not understanding what they're in charge of. Don't forget, people have volume knobs, and they're not afraid to use them. I'd recommend not getting caught up in the loudness wars, and making louder synonomous with better. Mastering can improve sonic quality, but it is more frequently used for evil in commercial music.

Take it easy at first, and you'll be thankful a year from now when you listen back to your first mastering attempts that you didn't go overboard trying to acheive commercial loudness through compression and maximizing. A lot of the loudness in commercial mixes starts in the recording phase with sidechain compressors on various instruments, and stereo imaging placement of instruments.

Hope this helps, and don't ever think of making what I just wrote a formula! I'm just passing along a few of my preferences.

 

[Michael Fossenkemper]

First, I've never had a "Record Company" ask for more level. It is always from individuals trying to second guess what the record company might like. They figure in order to stand out from the other with in the company, louder is better.

Second, you don't have the gear. As much as you would like to get this loud and clean volume, it's not going to happen with the gear you have. I know it's not something you want to hear, but there is a reason why top shelf mastering gear costs a lot.

Third, it takes years and years to get to this level of being able to slam something this loud and not make the listeners run. I could tell you exactly how to do it but I guarantee it would be 100 times cheaper and 1000 times quicker to hire someone.

You have the tools to make it reasonably loud, even very loud. I have a hammer and a table saw, but that doesn't make me a master carpenter.

Why do you want it that loud to begin with? what possible pressures could there be to make it as loud as black eyed peas? Not every record has to scream. I personally don't think a song like this should be that loud.

I feel your frustration. But there is no easy answer. Like ANYTHING.

 

[pingu]

First, I've never had a "Record Company" ask for more level. It is always from individuals trying to second guess what the record company might like. They figure in order to stand out from the other with in the company, louder is better.

Second, you don't have the gear. As much as you would like to get this loud and clean volume, it's not going to happen with the gear you have. I know it's not something you want to hear, but there is a reason why top shelf mastering gear costs a lot.

Third, it takes years and years to get to this level of being able to slam something this loud and not make the listeners run. I could tell you exactly how to do it but I guarantee it would be 100 times cheaper and 1000 times quicker to hire someone.

You have the tools to make it reasonably loud, even very loud. I have a hammer and a table saw, but that doesn't make me a master carpenter.

Why do you want it that loud to begin with? what possible pressures could there be to make it as loud as black eyed peas? Not every record has to scream. I personally don't think a song like this should be that loud.

I feel your frustration. But there is no easy answer. Like ANYTHING.

Agreed.

Black eyed peas loudness is absolutely ridiculous

 

[headchem]

You're right, it's not record companies who want it loud. I guess I should have said that people think talent scouts, A&R, people, etc, will pay more attention to the music if it's louder.