I hear-by pose the question...how does one compress audio more efficiently?

[Jerome Niko]

Hi folks,

I'm having a serious issue with how to compress an entire mix of video footage audio. I suppose that the reason I am such a stickler about well compressed audio is because I enjoy music very much and audio quality is just as important to me as video quality. My current issue is as follows:

I edit with Adobe Premiere Pro on a PC and the problem I am having is that I want to compress an entire audio mix-down of my whole video project (completed). However, the only way I can think of doing so is by putting a compressor effect on the Master Fader in the Audio Mixer window thus bringing the quietest parts of the video closer to the loudest parts of the video in terms of dynamic range. In theory this method would work just fine, but unfortunately it's a completely different outcome in practice.

I'm am at a loss when it comes to resolving this issue as it seems to be a lost cause at this point. I don't know of any Master Audio Mix options and so I thought I'd try asking here to see if anyone had any advice on this issue. Another thing I've noticed after playing around with compression a little bit is that even when compression is being applied it does not have any effect on preventing audio clipping.

I appreciate any advice that someone might have on this issue. THANK YOU IN ADVANCE AUDIO EXTRAORDINAIRE!!!

Sincerely,

~L.J.~

[Steve House]

Just why do you want to compress the audio - what are you trying to achieve and why do you think that's what's needed? IMHO the idea that "loud = good" is a false god and compression is too often overused, especially in pop music. Remember what it does ... it decreases the dynamic range by selectively reducing gain on the loudest passages while at the same time using 'make-up gain' to increase the average level. But the contrast between loud and soft passages is a dramatic element that gives vitality to a performance and the result of overuse of compression can be a performance that is loud but lifeless, like a used car commercial.

If you have clipping either your source material is already clipped or you have your overall gain too high to start with - tackle that issue first. Once it's clean, if the dynamic range is such that soft passages get lost when the loud passages aren't clipping, THEN think about adding a smidge of compression to help bring them up.

If your output consists of a mix of dialog, music, and effects and you elect to use compression, it's not advised to apply it at the final mix stage but rather to the individual stems before they're mixed.

[Brett Sherman]

Originally Posted by Steve House

But the contrast between loud and soft passages is a dramatic element that gives vitality to a performance and the result of overuse of compression can be a performance that is loud but lifeless, like a used car commercial.

In an ideal world everyone listens to what we produce in a dead silent room with no distractions. But we all know that's not true. So IMO compression is useful to allows quieter sections to be heard in less than perfect situations. That's why pop music is mixed the way it is, to be heard in the car or iPod on a noisy street.

That being said, improper use of compression can suck the life out of audio. I manually level dialogue first (a simple form of compression) and then apply light compression to it, usually UAD-1s LA-2A compressor.

The other thing is that not all compression plug-ins are made equally. Some sound better than others. My first suggestion would be to ditch the ones included with your software and try out some third-party ones. UAD-1 is great but you have to buy a card. I like Sonalksis too.

[Chris Soucy]

hesitant as I am to venture into this speciality with scarcely a smidgen of knowledge to back it up.

If you are such a keen music lover, why on earth do you want to compress it? If it's been recorder right in the first place, and everything else ditto, what needs to be compressed?

If the tops blow out your base drivers, then so be it, turn it down a smidge. If the vocals are too soft, turn 'em up in the mix.

Hey, if you're in the habit of watching your videos on a NY subway at rush hour, I concede you could have a problem with the sound, but I guess that's not where it's being watched. Nor in the car or a plane, but probably in your own lounge, with not a sound to disturb you from elsewhere.

So, given that (IM(extremely)HO) the only reason for compression is to make up for the failings of the audio audit trail somewhere along the route, and yours is, presumably, digital, digital, digital all the way, what's the point?

And, as Steve says, why not nail this in the mix?

Or am I missing something "beyond my ken" being a complete sound ignoramus (apart from having spent a lifetime listening to good music on good systems - try Paul DuToid, Mussogsky's "Pictures at an Exhibition" @ the ConcertGeboew, Amsterdam circa 1987 - ish, DDD all the way - wow!)?

Cheers,


Chris

[Steve House]

Just adding a note to my original post, if compression is used on dialog to prevent clipping on loud passages while preventing soft ones from getting lost in the mix, it should be applied to that stem directly, before the final mix stage. Except for live-concert footage the music stem shouldn't need it since it will have already been mastered before ever getting cut into the program. Compressing after the mix can lead to pumping of the overall level. To control the final level at mastering to prevent accidental clipping, one might apply a brick-wall limiter set at -0.3 to -0.5 dBFS as the last step in the chain.

Originally Posted by Chris Soucy

....

Or am I missing something "beyond my ken" being a complete sound ignoramus (apart from having spent a lifetime listening to good music on good systems - try Paul DuToid, Mussogsky's "Pictures at an Exhibition" @ the ConcertGeboew, Amsterdam circa 1987 - ish, DDD all the way - wow!)?

...

Many years ago, back in the pre-digital age even, I learned what a good recording CAN sound like while attending a seminar on mic techniques put on jointly by Schoeps and DG where they demo'd using the original master tape of a single-point stereo recording (Blumline pair mic setup, Nagra recorder, 30ips half-track stereo) of the NY Philharmonic with the Metropolitan Opera chorus doing Beethoven's 9th in concert on location at St Patricks. The seminar was held in a proper dubbing stage theatre and the sound was nothing short of spectacular.

[Seth Bloombaum]

Well, I suppose we're into philosophy as much as sound engineering, and I'll say that there is no right or wrong when it comes to compression, but that won't stop me from sharing my opinion.

I frequently work with another engineer on field recording, mixing and live sound. I use compression and loudness maximizer all the time, as minimally as I see fit for the listening environment. I use it at the track/channel level and at the master bus level.

My good buddy would never think of doing that, but then his specialty is recording and mixing for film, especially dialog for theatrical release. He'll run a little limiter as needed on tracks or master.

Bottom line, his work sounds good, my work sounds good. If it sounds good, it is good.

I have a Ray Charles album that was quite a feat of engineering, completed just a couple years ago that combines some amazing older performances of Ray with a recent session of the Count Basie Orchestra, "Ray Charles Sings, Count Basie Swings". It has a dynamic range that is great at home, but unlistenable in the car. I'll probably rip and remaster it one of these days, so I can have something good for both.

Your average living room doesn't even come close to the listening environment of a theater. Every broadast and cable station out there uses compression and limiting across their master output for this reason.

I love stereo recording techniques for acoustic performance, and I've done a lot of it, with great dynamic range in the recordings. Who am I to tell a listener where they should listen to my recordings? If I'm doing it for them then I should be thinking of them and their likely listening environments first, not notions of the purity of wide dynamic range.

[Steve House]

Actually I won't disagree with you, Seth. The idea is to fit the dynamic range of the material into both the environment where it will be heard and also into the capabilities of the delivery medium. A mix that's right for a digital track on a DVD will be all wrong for the same material intended for distribution on a VHS tape. The reason I suggest not to apply compression to a mix of music and dialog at the final mastering stage is that the variations in the level of the dialog will drive the compressor action and can cause the music level to noticably and inappropriately pump up and down in sync with it.

[Douglas Spotted Eagle]

Interesting dialog, as I'll usually take an approach different than either of you, apparently.
~Dialog tracks-compression as needed, usually very light. Maximizers occasionally used, depends on the voice.
All dialog sent to bus, bus lightly compressed
~Music-depending on the mix, rarely compressed due to mastering prior to music delivery. Musical elements are all on same bus. Maximizer often used.
~FX-usually fairly heavily compressed. Sent to same bus as atmospheres.
~Sweeteners-usually on FX bus, but not always. Compression heavily used here, along with widening which is in itself a bandwidth compression with other elements.
~Atmospherica-never compressed, usually too quiet to matter.
FX/Atmosphericals bus is compressed

All busses mixed down to master

~master bus has overall compression, sometimes virtually none, sometimes quite a bit.

YMMV

[Pete Cofrancesco]

Most of my stuff goes to DVD so I use the standard Ac3 compression. Since audio takes such a small percentage of the video space, overly compressing it doesn't make much sense. If I was doing studio work I'd probably pay more attention to the quality. The initial audio quality of dance recitals, lectures, weddings or other live events seems to be more important than what happens in the editing process. You'd have to go out of your way to butcher it by over compressing it. Most either use DVD or DV to back up a completed project. I'm not really sure what the nature of your question is?

[Sam Rosado]

Wow, There are a lot of great answers here and for once, I do not feel like a newbie. Sound recording has been MY field since the eighties from 8 trak 1/2 inch tape, to the 90's, when I first went digital, and all I can say is that compression and limiting has been a must! I agree with the UAD card, however, if you have a strong enough mac and a good budget go to WAVES proffesional. It has been the closest thing to hardware that i've found. I also use my DAW(DP5.1) to mix and master my audio. I have the luxury to mix in 5.1. I also have the equipment to do so, but, Waves has been worth every dollar. Compression settings are like girls, You have to find the one YOU like! There is not a defined preset, just listen, if it sounds good, then its good!(whosaidthat?)

[Seth Bloombaum]

Originally Posted by Pete Cofran

...Since audio takes such a small percentage of the video space, overly compressing it doesn't make much sense...

I think what you're referring to is reducing bit depth, sample rate, or going to a lossy format to reduce the storage space and bandwidth requirements of audio. Most of the discussion above has to do with the use of compressors, limiters and volume maximizers to affect the dynamic range (range between softest sounds and loudest) to create a mix that sounds good in various playback environments.

Originally Posted by Sam Rosado

...Compression settings are like girls, You have to find the one YOU like!...

Wow. Laughing out loud on that one! But there is some essential truth in that.

Originally Posted by Sam Rosado

...if it sounds good, then its good!(whosaidthat?)

I can never remember if it was Duke Ellington or Count Basie. Today I think it was Ellington.

[Graham Hickling]

Originally Posted by Lars Johnson

Hi folks,

However, the only way I can think of doing so is by putting a compressor effect on the Master Fader in the Audio Mixer window thus bringing the quietest parts of the video closer to the loudest parts of the video in terms of dynamic range. In theory this method would work just fine, but unfortunately it's a completely different outcome in practice.

I'm going to completely dodge the issue of whether it's good to compress ... I'll assume that having read all of the above posts you do still want to.

The focus of your post seems to be that the method you chose didn't give you the result you wanted. In what way - can you clarify?

FWIW, I often take my audio out to Cubase for mastering ... since it provides a wide range of options for applying compression and other effects to the final mix and/or the individual tracks. Audition would be similarly helpful, I assume.

[Pete Cofrancesco]

Originally Posted by Seth Bloombaum

I think what you're referring to is reducing bit depth, sample rate, or going to a lossy format to reduce the storage space and bandwidth requirements of audio. Most of the discussion above has to do with the use of compressors, limiters and volume maximizers to affect the dynamic range (range between softest sounds and loudest) to create a mix that sounds good in various playback environments.

after reading the other replies, i realized this but i couldn't delete my post and save myself the embarrassment. :-(