Topic: making loud softer and soft louder in a mix using sound forge or other programs?

[SClassical]

hi i have a problem with my final stereo mix. my vocals is too loud and my accompliment is too soft. how do i reverse the vol of my track? is there a way to make the soft louder and loud softer? does sound force have this option? if not is there a program that allows me to reverse it so my loud is softer or soft is louder? thanks

[aaronji]

Most editors have a compression option; that's what you are looking to do.  There is a good thread started by gutbucket about parallel compression that may be useful.  page has also posted some very helpful information...

[achalsey]

An envelope tool might be what you're looking for.  I use one some what successfully for acoustic shows to boost the music levels but lower the applause.

[SClassical]

hi thanks for the replies. is the envelop tool in sound forge and work like a compression? I will look for it. thanks!!

[Brian Skalinder]

For reference, we have bookmarked some links to thread discussing compression, envelopes, and parallel compression in the Post Processing, etc. board's Readme1st thread.

[stevetoney]

Guys, please correct me if I'm wrong, but compression won't fix a situation where vocals are too loud RELATIVE to accompaniment, will it?   

[page]

Guys, please correct me if I'm wrong, but compression won't fix a situation where vocals are too loud RELATIVE to accompaniment, will it?   

Depends on how dramatic a difference it is. The bigger the difference the better it will work, I'm talking >5db. It's not end-all-be-all, but you can definitely improve things that way. It takes some practice to get used to and it's a case by case thing on what tactic you use (parallel, top down, etc).

[stevetoney]

Guys, please correct me if I'm wrong, but compression won't fix a situation where vocals are too loud RELATIVE to accompaniment, will it?   

Depends on how dramatic a difference it is. The bigger the difference the better it will work, I'm talking >5db. It's not end-all-be-all, but you can definitely improve things that way. It takes some practice to get used to and it's a case by case thing on what tactic you use (parallel, top down, etc).

Goddammitall.  I've said it before...I wish I could just go out and buy a book on this stuff.  Been using Audition for so many years and still don't know how in the hell it works.   

[page]

I read up on some terms and tried a crap ton of plugins (and then paid for the few I kept). After that, it was just practice. Dig out any onstage/stagelip tape you have, odds are it could use some improvement since it's not running through the sbd > compressor > PA stacks. If you need a sample or two to tinker with, I can dig up some of mine.

I still think it would make a great session for a TS.com convention.

[Matt Quinn]

Compression will probably not help a mismatch like that very much. Like page said, you can squash it all together if the difference is big enough, but in squashing it that much, you're probably not going to like the overall sonic results. Multiband compression would probably help a bit more, but only if the music is fairly sparse (like an acoustic guitar/singer). If it's something really dense frequency wise, you'll get too much overlap to make it sound good.   

And there are plenty of books out there.....many many many books, YouTube tutorials, etc about this stuff. Compression is not that complicated though. You set a volume threshold to determine when the compressor kicks on/off. Signals above that threshold will be compressed at a certain ratio....ie if you set it 2:1, for every 2db above the threshold going into the comp, it's only going to let 1 db through.  You set how fast the comp reacts to the signal passing the threshold with the Attack and release controls. Then you'll have a control for output/make up gain to raise the overall compressed signal. And that's pretty much it.

[page]

You set how fast the comp reacts to the signal passing the threshold with the Attack and release controls. Then you'll have a control for output/make up gain to raise the overall compressed signal. And that's pretty much it.

On this note in particular; its helpful to look at how fast sounds occur. Drum hits or hand clapping is a radically faster attack then electric guitar. Likewise, you can use a compressor to shape a sound (longer attacks to leave a punchy sound on the attack vs a super fast attack to just shave stuff off and compress drum hits at the same time).

[Gutbucket]

You can only reduce the degree of diffence between loud and quiet parts happening more or less simultaneously, you cannot reverse those differences.

The concepts behind compression is pretty straight forward.  The settings are interactive with each other and take a bit more time to figure out.  Learning to hear a change in adjustment in attack and relase can take time.  It's easy to convice yourself your settings are good but kill the life in the track while focusing on the details.  The subleties are important and challenging to get just right.  It's much more fidgety than EQ which is far more obvious and easy to hear.

Try a simple 'two knob' (threshold and ratio) compressor without too many other settings to start out with.

I think of compression in three ways and generally apply it in three seperate stages to address different things.  I don't always use them all, just as called for:
1) Top down- limiting wild peaks (high threshold, high ratio, fast attack, moderately fast release)
2) Straight full range- controlling the overall range of level like a volume knob adjustment (low threshold, low ratio, moderately fast attack, medium to slow release)
3) Bottom up (parallel comp)- bringing up the quiet parts and bringing out out details in the music  (low threshold, high ratio, paralleled with the uncompressed but limited audio) partly similar to full range, but works 'within' the loud parts too, unlike a 'volume' knob.

You can do 1 and 2 with use of the envelope tool and a lot of time- Peak limiting by zooming in and making lots of very small tweaks to bring down each individual errant peak, and full range by zooming out and making big, broad level changes.