n-Track Studio Board » n-Track Studio Discussion Forum » n-Track Studio Discussion Forum » Compression

Topic: Compression, Audio Minds has melted my brain!

< Next Oldest | Next Newest >

Ice-9 Offline





Group: Members
Posts: 13
Joined: Mar. 2004

Posted: Mar. 30 2004,12:17

QUOTE


Yes it has, hence why I am confused and about to post hilarious ramblings for your pleasure.

At least I have tried to do some reading though - and not just posted blindly!

I checked out a load of resources at AudioMinds, especially in reference to mastering and the voodoo art of compressing. I read a number of articles and I thought I understood what was going on. I certainly understood each function within compression but it was putting it into practice that I realized I was actually more confused than when I started!

Apologies again, if what I am about to write is moronic and has been posted several times by other users. I gather I have probably got the wrong end of several sticks.


1) "A compressor will turn down volume levels that are above a certain point to a degree that you tell it to, without turning down volume levels that are below this point."

Okay, but if you've mixed it correctly, then no one track should be so much louder than another that it needs turning down, surely?

And on top of this, you want certain tracks to be louder - like the guitar solo - and you want others to be quieter, like backing vocals.
If you compressed these louder signals then wouldn't you be undoing the job you did in mixing it? Wouldn't the tracks be too close to each other, volume wise?

The only way I can see compression being effective is over individual tracks, rather than across the entire song. This way, when applied individually, you ensure that the guitars, vocals, drums etc stay within a certain volume range and so correct the moments when you strummed the guitar too hard or where the singer put too much emphasis on particular words.



2) "Threshold is the control that tells the compressor at what incoming volume level to start compressing."

Again, it makes sense in reading but assuming you have mixed it okay, why would you need to compress it further than it already exists in its freshly mixed state?


3) "In digital, 0db of any meter is the loudest possible volume there is, so the Threshold setting on a digital compressor will always be a negative value."

So you cannot make it hotter, you can only make it quieter. If you've mixed it okay, why would you want to reduce the volume of it all?


4) "With a Ratio of 2:1, any volume at the input that is 2db over the Threshold would come out of the compressor only 1db louder. If the incoming volume was 4db louder, it would come out only 2db louder."

This is a paradox to me. I want the compressor to kick in at a certain volume yet from the definition above, the outputs would be louder than the Threshold - by 1db and 2db respectively.

What is the point of the Threshold limit if you can ratio above the limit?


4) And should you compress the entire song or the individual tracks? As I said before, I can only see it being effective if you applied it to the separate tracks............but I'm probably wrong.


Many thanks.

Back to top


Profile PM MAIL 

 

Bubbagump Online





Group: Members
Posts: 1358
Joined: Oct. 2003

Posted: Mar. 30 2004,12:23

QUOTE


Quote

Okay, but if you've mixed it correctly, then no one track should be so much louder than another that it needs turning down, surely?



Umm, compression allows this 'mixing' correctly to happen. It makes tracks more uniform. Think of compressions as a DYNAMICS tool, not a volume tool. It is used to fix dynamics weirdness. You're not always going to have perfectly done tracks. Folks move closers and further from the mic, there are pops, etc. You can do the same thing as compressions more or less by riding faders, it is just A LOT more work. Volume drawing and your mixer adjust volume.




Quote

What is the point of the Threshold limit if you can ratio above the limit?



Because you want to maintain the transient. What you are describing is brick wall limiting. This just chops off a wave form and kills your transients. Compression keeps the transients in tact unless you over do it and smash the transient to the same level as the rest of the wave form.


Quote

And should you compress the entire song or the individual tracks? As I said before, I can only see it being effective if you applied it to the separate tracks............but I'm probably wrong.



Well, in mixing, you will usually compress select tracks that need it. Then, in a totally separate process called mastering, you may compress the whole mix. Many mastering engineers will chop your fingers off if you compress a whole mix before sending it to them. For us home recording types, we will probably master our own work. But treat mastering as a separate process. Worry about getting a good mix, then in mastering you can compress the mix for loudness.

--------------
www.krististrauss.com

Check it out and let me know what you think.

Back to top


Profile PM MAIL WEB 

 

Ice-9 Offline





Group: Members
Posts: 13
Joined: Mar. 2004

Posted: Mar. 30 2004,12:29

QUOTE


Bubbagump,

When you say

Quote

"compression allows this 'mixing' correctly to happen. It makes tracks more uniform."

Do you mean individual tracks - so that each track doesn't vary too much or are you talking about making the tracks more uniform together?

I didn't understand at all your explanation about ratio. Sorry!

Back to top


Profile PM MAIL 

 

Bubbagump Online





Group: Members
Posts: 1358
Joined: Oct. 2003

Posted: Mar. 30 2004,12:36

QUOTE


This is for individual tracks. Say there is a spot where the vocals sort of jump out and get harsh. You would use compression on this. The compressor does not kick in until this overly loud part. You can also ride faders as I said. (Now it is volume drawing.)

The key here is dynamics. Some time the soft in a track is too soft and gets lost in a mix, or the loud is too loud and sticks out. You use compression to even this out. Otherwise, you would have to bump the fader up at the quiet part and down at the loud part. Compression does this automatically for you.
   Compression also can add attack to things. Say I have a wimpy acoustic guitar track. I can compress it to add more 'chunk' to it. What you do here is make the attack slow enough to compress after the initial attack to make the transients stick out more in relation to the rest of the sound.

--------------
www.krististrauss.com

Check it out and let me know what you think.

Back to top


Profile PM MAIL WEB 

 

phoo Offline





Group: Members
Posts: 732
Joined: Oct. 2003

Posted: Mar. 30 2004,12:42

QUOTE


One thing to keep in mind: The FASoft 1.1 Compressor controls don't act like traditional compressor controls (analog old fashioned....). This drives me bananas, but I like the compressor so I live with it.

Don't use it and then use the threshold and ratio sliders and expect to hear what a traditional compressor would do...and this is when compensation is turned off. You can look at the graph and see what it does, but putting what your eyes see into you mental ears is not that easy, if not completely impossible. Threshold does not change just the threshold and ratio does not change just the ratio.

Other compressor plug-ins have similar problems, but not all. I've been messing with SoundForge's WaveHammer this week, and it's compressor is close to analog compressors, but it's still not perfect.

There are many I haven't checked out though. I still use FASoft 1.1 Compressor for most things.

--------------
Shut up and play yer guitar  -  FZ

Back to top


Profile PM MAIL WEB 

 

learjeff Offline





Group: Members
Posts: 785
Joined: Oct. 2003

Posted: Mar. 30 2004,12:50

QUOTE


Quote

The only way I can see compression being effective is over individual tracks, rather than across the entire song.



Then you're half right.

Like Bubba said, compression changes the dynamics (reduces it).  You can think of it as making the too-loud parts quieter, but I often find it more helpful to think of it as making the too-soft parts louder.

You can do this on individual tracks, and it makes any performance more "even" in volume.  You can also do it on the whole mix, and it makes the whole song more even in volume (and generally, louder overall).  The latter has to be done with even greater care, though.  For example, let's say we have a keyboard pad sound that's steady in volume, and a guitar or vocal part that suddenly soars in volume.  If we compress this mix, the soaring part will cause the gain to be turned down on the whole thing, and what we'll hear is the pad track "duck" when the lead part soars.  Sometimes this is just what we want.  Other times it sounds unnatural.

This is one reason why folks often use a multi-band compressor for compression on the final mix.  That way, a leaping guitar part won't cause the bass line to duck, because the different frequency bands are compressed separately.  You can still have the problem, but it's limited on a band--by-band basis.

But to prove to yourself how valuable a little compression on a mix can be, try using peak Compressor on a mix, with the default ratio and threshold set to say, -6 dB, and with a final gain of about 3 dB.  If your original mix doesn't clip, it probably won't with these settings (and if it does, just lower the peak threshold -- I think it's a blue horizontal line).

This isn't how you'd normally dial up compression -- usually you listen as you draw the threshold down and wait to hear what you want.  But just use these example settings and try it, and do an A/B comparison of the result.  You'll find you hear more detail in the compressed mix, but there will be less dynamics too.  The compressed mix will sound louder at the same gain level.


Quote

3) "In digital, 0db of any meter is the loudest possible volume there is, so the Threshold setting on a digital compressor will always be a negative value."

So you cannot make it hotter, you can only make it quiter. If you've mixed it okay, why would you want to reduce the volume of it all?



By making it all quieter, we can push up the gain (master fader, or the gain knob in the compressor).  So the loudest parts say at the same level (nearly 0 dB), but the quieter parts are now louder than they were before.

4.  The compressor only BEGINS to kick in at the threshold.  It reduces the volume more the higher the input volume is above the threshold.

Also, note that once you go above the threshold, the compressor's gain gets turned down according to level and the ratio.  But it doesn't spring back up immediately when the volume goes back down -- it drifts back up more slowly.  This is one of the trickier parts about adjusting a compressor to the source material and desired effect.

And frankly, it doesn't even get turned down immediately -- there's also a rate governing on how fast it clamps down (turns down the gain).  That rate is usually much faster than the release rate (to drift back up).  On a peak limiter, that rate is much higher -- it clamps down much faster, sometimes immediately (with "look ahead").

--------------
"I never knew I had one, 'Til I saw yours shine" - Johnny Clegg

Back to top


Profile PM WEB AOL 

 

loauc Offline





Group: Members
Posts: 41
Joined: Oct. 2003

Posted: Mar. 30 2004,13:04

QUOTE


Here's a real world example of using a compressor:

I have a drum track with snare, hi-hat and some kind of cowbell. The song starts with little beats on hi-hat and cowbell and I set my level to where it fix nicely in the mix. Then, suddenly, comes a pre-chorus and there are some full snare hits. I can trim them down with some envelope-drawing. But then this song reaches the chorus, which is a very busy part with cowbell, h-hat and snare being hit in a rather complicated pattern. I wouldn't want to draw envelopes for some 20 measures of those hits. Instead I put a compressor as a track insert and do as follows:

- Hear until I find the sweet spot
:p (kidding, my ears need some visual feedback)

- First I set the ratio, around 5:1 since differences in volume are big, with the cowbell being a subtle click in comparison with the big and loud hits of the snare.

- Then I start to move down my threshold control until I start to notice a dropdown in volume

- Next I compensate with the out gain for the perceived loss in volume

- Finally I adjust attack and release trying to visualize how much a snare hit lasts (you can do that zooming your waves) and setting the attack to lag at the start of the snare hit around 20% so I can keep the initial bang, and letting it decay for a little less than the time it takes another beat to come.

If I had recorded in separate tracks the snare, the hi-hat and the cowbell I certainly wouldn't be a home-recordist.

:cool:

Regards,
Loauc.

Back to top


Profile PM MAIL 

 

Mac Offline





Group: Members
Posts: 4218
Joined: Oct. 2003

Posted: Mar. 30 2004,20:13

QUOTE


In digital recording, the audio compressor is literally everything.  

Use it on the single track to yield a good mix fit.  There are different basic settings that can be good starting points for each type of instrument or vocal track, use of studiobuddy.exe on yer desktop is a great reference for the beginner.  

Use it again on the finished and mixed down stereo tracks as a Mastering Compressor.  This is almost always the use of what we call a 'soft knee' compression, and it is not uncommon to do the job in more than one pass in order to make the final volume as high as can be but also musical sounding.  Compressors can also be used as Limiters in this situation.  

Now that you've read a bit about it, do something hands on for a bit, like download the n-Track demo song and try your hand at the compressor on a selected track, following the basics from StudioBuddy.  tweak and listen.  save files with different settings applied and give 'em names that tell you what is what, then come back tomorrow and re-audition the sounds of each one each way.  Start to get a feel for what your ears are telling you a setting situation sounds like.  Give the things names of your own that describe the sound --  harsh, cold, warm, icy, steely, anything that means something to you alone and can assist the recall of what made it sound that way, whether it was the ratio, the threshold, etc.  

Takes time to learn this business, use the time wisely and get there faster, avoid the temptation to simply select some default setting and accept whatever it might do to your music right or wrong, although you might find a default setting that is close:  so tweak it, one control at a time, one way or the other, and LISTEN to the result in context with your mix.  

Keep reading, a good idea is to Google up some of the audio tutorials that take you thru someone else's pro tracking and mixing sessions (prorec.com has a few good articles that should help out by giving you a birds eye view of this from beginning to end of several sessions, search for articles by Rip Rowan among others) and just do a little each day and take it to the DAW and do some hands on with it each time.  In a matter of a few weeks you will have a much better understanding on what to do, and then you can practice making your own recordings.  Each recording should be a tad bit better than the last, and again it's just like anything else in this life:  it's the practice that makes the expert.  

--Mac

--------------
"Our constitution was made only for a moral and religious people. It is wholly inadequate to the government of any other."  -- John Adams

Back to top


Profile PM 

 

Teej813 Online





Group: Members
Posts: 1242
Joined: Oct. 2003

Posted: Mar. 30 2004,22:01

EDITQUOTE


but Mac!  This is the age of instant gratification.  You're saying we have to WORK at this?


sheesh..  :;):



tj

--------------
For God so loved the world, why can't we?

New to digital recording?  Start here!  www.audiominds.com

Back to top


Profile PM MAIL ICQ MSN YIM 

 

Ice-9 Offline





Group: Members
Posts: 13
Joined: Mar. 2004

Posted: Mar. 31 2004,16:34

QUOTE


Thanks for the advice guys though I am still at a loss in places. I can see you're really trying to get this through to me though! :D

Let me just recap some issues.

1) Clarifying the facts

Learjeff you say:

Quote

I often find it more helpful to think of it as making the too-soft parts louder

yet http://www.recordingproject.com/article....&page=2
states:

Quote

"If the music is above a certain level, it turns it down. If the volume is below the point that is desired, it does nothing to it."



So who is right and who is wrong?


2) Making bits louder/quieter

I understand the concept of compressing individual tracks to iron out where the vocal/instrument jumps out but I don't get the concept of whole track compressing because even if it does make the quiet bits louder, as Learjeff says, doesn't this undermines the point of mixing in the first place?
I mean, if I wanted certain tracks to be louder, I would just tweak the volume as I am setting the levels.

Lemme explain my process:

a) Record individual tracks.
b) Pan them as accordingly.
c) Mix the individual volumes to suit.

At the moment, as I am new to the scene, I am staying away from aux channels, effects and the more complex features.

Let's say my song features drums, bass, an acoustic guitar strumming in the background, overdriven rhythm guitar, lead guitar and vocals.

Bearing the above process in mind, when I mix, I will have set the acoustic guitars quite low in the mix and just to fill out the sound, the rhythm guitar in the middle and the lead guitar and vocals even louder.

I can't see why I would want to compress this and make the acoustic louder whilst lowering the volume of the lead guitar and vocals. To me, that just undermines the entire mix I have set.


3) Setting the Threshold

I know this will vary from song to song but I am having difficulty getting my head around this.

Let's assume that all tracks have been recorded and I have set each volume and pan in the mix, it all sounds great and I am ready to compress.

I know what the Threshold does, I just don't understand at which point, and why, do I set the Threshold?
As I can't go above 0db, and the lead guitar is the loudest track, I assume that this will be 0db whilst the rhythm tracks will be about 6db lower - that's the difference between my rhythm and lead channels on my amp, anyway! :D

So, why would I set the Threshold lower than 0db and thus bring down the lead guitar volume?

And again, if the volumes are all set in the mix, why would I want to lower and/or raise certain track volumes? I just can't get my head around this.

Nor can I understand what level to set Threshold at. -3db? Why? Why not -1db or -5db or -10db? How do you reach whatever figure you come to? More importantly, why?

As I have said, I understand compression gives more punch and that sense of loudness. I understand why you'd use it on individual tracks.

It's these other areas I am struggling with.

Back to top


Profile PM MAIL 

 

duncanparsons Offline





Group: Members
Posts: 57
Joined: Nov. 2003

Posted: Mar. 31 2004,17:23

QUOTE


Quote (Ice-9 @ Mar. 31 2004,15:34)

Let me just recap some issues.

1) Clarifying the facts

Learjeff you say:

Quote

I often find it more helpful to think of it as making the too-soft parts louder

yet http://www.recordingproject.com/article....&page=2
states:

Quote

"If the music is above a certain level, it turns it down. If the volume is below the point that is desired, it does nothing to it."



So who is right and who is wrong?


well.. both..

Look at it this way: I have some audio:
|
|                                  /\
|                                 /   \
|                                /     \
|               /--\           /       \
|              /     \         /         \
|   --       /        \      /            \
|  /   \    /           \--/               \
| /      \/                                   \
|/
-------------------------------------

Now then, that last peak is causing some problems, so I compress it, at the threshold of the previous peak:

|
|
|
|                                /----\
|               /--\           /       \
|              /     \         /         \
|   --       /        \      /            \
|  /   \    /           \--/               \
| /      \/                                   \
|/
-------------------------------------

That last peak has been compressed leaving the other two minor peaks alone. This sounds OK now, however, it doesn't stand out in the mix properly, so I use the gain control of the compressor to raise the levels:


|
|                                /----\
|               /--\           /       \
|              /     \         /         \
|   --       /        \      /            \
|  /   \    /           \--/               \
| /      \/                                   \
| |
|/
||
-------------------------------------

So what we did:
1: brought the large peak into the same range as the other peaks
2: brought the whole sound up, keeping the new dynamics.

Notice: The loud bit got softened by compression, the quieter bits got louder by the increased gain.

Sorry about the ASCII art - it is only an outline of what goes on.. [the thought of drawing some gifs, posting them to a site and pointing to them was all a bit much!]

HTH
DSP

Back to top


Profile PM MAIL 

 

loauc Offline





Group: Members
Posts: 41
Joined: Oct. 2003

Posted: Mar. 31 2004,19:50

QUOTE


Quote (Ice-9 @ Mar. 31 2004,15:34)

As I can't go above 0db, and the lead guitar is the loudest track, I assume that this will be 0db whilst the rhythm tracks will be about 6db lower - that's the difference between my rhythm and lead channels on my amp, anyway! :D

So, why would I set the Threshold lower than 0db and thus bring down the lead guitar volume?


The threshold control won't make your track quieter as if it was a volume fader, it will just set the level where the compression takes place. And note that compress is not the same as "turn down" as it  depends on several factors.

The way signals above threshold are affected is defined by your ratio control.

Try to imagine the threshold as the level where the peaks in your signal are becoming flat (take a look at DuncanParsons' drawings).

And about your assumption of your loudest track being at 0 db it isn't right. You have to watch your VU-meters. And how do you know the level your amps are giving in db's? What the knobs say are not db's.

This leads us to the subject of compressing the whole mix. I can see your concerning of having spent a good deal of time balancing it and then asking why to use a compressor to trash everything. Well, for compressing the whole mix you usually use different settings than what you used for individual instruments.

Generally speaking, you compress the whole mix just to reduce peaks, it enables you to push up the volume a little more. And don't worry, reducing peaks isn't the same as lowering the volume. If you compare your mix before and after the compression you'll almost certainly only notice a louder signal, unless it gets really squashed and you'll know you overdid it when you hear it.

Hope it helps,
Loauc.

Back to top


Profile PM MAIL 

 

jimbob Online





Group: Members
Posts: 200
Joined: Oct. 2003

Posted: Mar. 31 2004,21:37

QUOTE


Quote

1) Clarifying the facts

Learjeff you say:
Quote  
I often find it more helpful to think of it as making the too-soft parts louder  
yet http://www.recordingproject.com/article....&page=2
states: Quote  
"If the music is above a certain level, it turns it down. If the volume is below the point that is desired, it does nothing to it."  


So who is right and who is wrong?



When Teej says he thinks of it as making the quiet parts louder, he is assuming that you either have the "compensate gain" button selected or that you will be raising the level since your headroom has been increased. The "compensate gain" button automatically increases the gain based on an estimate of how much the peaks are reduced. Under these conditions the soft parts are louder but it is not the compressor per se that does it. I also find that the "compensate gain" function will occasionally allow clipping within the compressor so you may want to turn it off and adjust the output gain fader directly or use the channel fader to bring the track to the right level.

I think you are a little confused about the time scale. Compressors are used to reduce the dynamic range for short events. Faders are used for long-term adjustments and volume envelopes for intermediate time scales (solos etc.). I sometimes use volume envelopes instead of compressors when I want more precise control but it is very time-consuming to do it for a whole piece. I also use envelopes to alter the inflection of notes to improve the expressiveness of the part.

Overall compression has advocates but most of us use limiting more than compression on the total mix. It is mostly done to keep short duration peaks from clipping when l the overall level is increased. A limiter is a compressor with very fast attack and release times, a very high ratio and (typically) a relatively high threshold.

Some people use multi-band compression to reduce pumping but I'm not a big fan.

Jim

Back to top


Profile PM MAIL 

 

learjeff Offline





Group: Members
Posts: 785
Joined: Oct. 2003

Posted: Mar. 31 2004,22:10

QUOTE


Iceman, you're thinking too hard.  Do some experimentation and listen -- the results of compressing a mix are very significant and it would take a LOT of careful tweaking of individual parts to get the same effect.

There are purists who agree with you that we should avoid compression on a mix, since it can ruin the dynamics of the piece.

The truth, of course, isn't black or white: it's a tool, use it wisely.  That said, if you post tracks online that aren't compressed, folks will be annoyed that they always have to crank the volume way up when they play your tracks, after playing others.  (A few purists who bother to download mp3's may applaud you ... but they're definitely a small minority.)

You'll also want a CD to master out so that you can play it after playing commercial CDs without adjusting the your stereo's volume.

But finally, you should concentrate on MIXING first, and leave MASTERING for later.  If you compare your results with similar material ripped from CDs, you'll start to notice the differences.  Eventually, when you have a bunch of good mixes, you'll want to get into mastering.

Compression of a mix is generally part of mastering.  If you're going to have a pro mastering house do it for you, then don't do any compression on the mix yourself.  Instead, just do your best mix, mixdown to 32 bit format, and send off for mastering.

So bottom line, it doesn't matter if you don't get it now.  You don't need it now -- focus on the first part, mixing.  You seem to have a good handle on that, but there's no end of things to learn!

:)

--------------
"I never knew I had one, 'Til I saw yours shine" - Johnny Clegg

Back to top


Profile PM WEB AOL 

 

Ice-9 Offline





Group: Members
Posts: 13
Joined: Mar. 2004

Posted: April 01 2004,15:44

QUOTE


Hi all,

Thanks for the replies. I get some of it but some I don't. I guess i will just try it and see where it gets me - and check back should I hit any more problems.

Before I go though, some issues:

I tried to DL the demo song but when I went to open it I was told "old format". I am using N Track 2.2.

As for compressing, can I undo any compression and try again?

And when do I add the compression? I take it if I add to individual tracks, this is when I can  see all the tracks in N Track - before I mix it down?

If so, when I attempt whole song mix down, is this on the same point or do I compress after it's been mixed down into one wav?

People are talking about peaks. Is there are way I can visualize these peaks and the compression effect in N Track?

Back to top