Compressing/Limiting and Volume Gain Question

[mayan]

I have sort of a complicated question - I'll see if I can express short and quick...I appreciate your patience in advace:

I have a very complicated sonic piece with alot (I mean ALOT!) of tiny sampled elements - many of which have been sent through processors, sound manglers, etc. For the most part, I really like the mix that I've slaving over for the past month.

Here's the thing, it's mixed so that the peak comes in around -3.5db. Now, in whacking it with a limiter, Voxengo Elephant, at say 4.5db gain - I've noticed that some artifacts/transient sounds in a certain frequency rangeseem to be brought up to undesirable levels. Seems to be some sort of high-hat extra frequency in the upper treble that sticks out and pisses me off. I haven't really noticed this in my other work before this - but then again, I've never spent so much time developing/hand-crafting every individual quanta of sound before. What it really reveals is my real lack of understanding regarding limiting/compressing and gain.

Now...here comes the questions. Would it be better to just boost the mix -via gain change- to close to 0.00db? Would it then make sense to do a much lighter limiting (I do like some of the sound that are brought forward with limiting) of say 1, 2 or 3 db? Would it be better to use some sort of compressor? I have to admit that I know VERY little about compressing - enough to get me in trouble. I know that I don't want to get a pumping sound on this particular piece and I don't want something that's going to add something garish or steroidal to the sound. I really like the way it is right now, for the most part...just want to bring some sounds forward a little bit.

EQ wise, I could go through and fix the most little glitchy undesireables by hand...or perhaps even to a very narrow EQ frequency thing through the whole piece...although again, I'm worried about affecting the mix in an adverse fashion because it really works as long as those artifacts aren't brought forward. It would not be practical for me to redo the entire mix track by track...although in retrospect, although I compressed the "drum" track, I probably should have applied a limiter to it as well before mix down.

Any how...don't know if this makes sense or not. Any suggestions? Any further questions? I'd be happy to try to answer 'em...It's a VERY cool piece in my mind and I'm anxious to share it but it somehow brought out a very latent perfectionism in me and I don't want to release it until everything is just how I want to hear it. Thanks for your time in reading through this.

[l33troll]

/me thinks maybe you ask too much limiting? 4,5db is not very moderate, isn't 3db the max without arteafcts?
EQ wise, maybe it's better to "fix" on a track basis?

[mayan]

thanks l33troll...interesting point you raise - as I said, I know very little about the "art and science" of limiting. Is the idea to get the mix as close to the 0.0db point and then apply a lighter limiting then to -say, have a mix at -3.5 and then get it up to 0.0 through limiting?

As to the track EQ...I know your point - and that's the problem. Everything has been carefully EQ'd etc. for the mix. It's just that: 1) the mix is a bit soft; and 2) the limiter brings out stuff that improve the overall mix - but for those little sounds that develop. In other words, if I take my -3.5db peaked mix and increase the volume through a gain change, I don't get the little unwanted glitches. OTOH, I don't get some of the other things brought forward either.

It sounds like trying a "lighter" limit or mastering compression might be helpful. I guess it's a matter of continued experimentation. Too bad I have such tin ears. Maybe what I will try to is to do an upwards gain change and a much lighter limiting. Maybe also try a master compressor as well, although I do want to avoid pumping and steroidal changes to the overall mix.

[l33troll]

It seems you answer to your own questions:

Is the idea to get the mix as close to the 0.0db point and then apply a lighter limiting then to -say, have a mix at -3.5 and then get it up to 0.0 through limiting?

If you can do it you will get less artefacts. It's often said you get volume more by compression (in tracks) than by heavy limiting the mix. The limiter is just there for the "final touch". A -18 db set L2 works only with quite drunk audiences.

As to the track EQ...I know your point - and that's the problem. Everything has been carefully EQ'd etc. for the mix. It's just that: 1) the mix is a bit soft; and 2) the limiter brings out stuff that improve the overall mix - but for those little sounds that develop.

Limiting or compression will bring up the lower sounds, that's why you'd better operate on a track basis. (I don't know your precise problem but a general solution is to use a gate with compression so that the gate kills the undesired sounds before they are brought up by the compressor, but maybe you still want them in the mix but not too loud? In that case you must be moderate in your compression)

In other words, if I take my -3.5db peaked mix and increase the volume through a gain change, I don't get the little unwanted glitches. OTOH, I don't get some of the other things brought forward either.

It sounds like trying a "lighter" limit or mastering compression might be helpful. I guess it's a matter of continued experimentation. Too bad I have such tin ears. Maybe what I will try to is to do an upwards gain change and a much lighter limiting. Maybe also try a master compressor as well, although I do want to avoid pumping and steroidal changes to the overall mix.

I see nothing wrong with this. In fact it's considered better to have to lift up the master than to bring it down. Good luck

[mayan]

Thanks,l33troll...that is very helpful and informative. Much appreciated.

[kritikon]

In general it's usually better to use compression rather than limiting. Limiting is only extreme compression anyway, but if it's a hard-knee limiter, it's basically a waveshaper. Soft knee compression can be your friend.

Hard-knee limiting = waveshaping = distortion

If you limit too heavily, even with a very good limiter, you WILL get distortion, and not a very nice type of distortion either.

It's very nice that there is so much of a choice of limiters etc nowadays etc, but really it's not a plugin that you should go to before trying everything else out first. If you use a compressor well, have a good mix etc, then all you'll need a limiter for is to stop clips. Of course, you can use a limiter to squeeze extra perceived volume out as well, but definitely not at the expense of a compressor.

Do a search for stuff on compression. there's alot of stuff out there. You'll likely get less nastiness in your mixes if you use softer compression rather than limiting.

www.digitalfishphones.com

I think that site is still up (could be wrong though). Sascha had some essential reading in his manuals for Endorphin and his other plugins. Pages and pages, with all the stuff you'd likely ever need to know to get you up and running with compression. The manual was 1.5MB or something - but every bit worth reading IMO.

A good general guide to mixing would be something like - channel compression -> channel Eq (if you like using Eq) -> subgroup compression -> master compression -> master Eq (again, if you like Eq) -> master limiting. And you don't have to use compression at all those stages - it's just to show you that there is alot you can be doing to improve a mix before resorting to limiting (which is rather like cracking a pistachio nut with a JCB digger)

The limiting at the end is optional, and really should mainly be to stop clips as an afterthought. You could also mix around compression->Eq to Eq->compression (it's a matter of taste - personally I much prefer compression->Eq).

[~BURNY~]

It seems like your making a confusion here. Limiting isn't a matter of knee but a matter of ratio (10:1 and more is considered like limiting).
Knee is the time that compression reach its maximum ratio once treshold is crossed and attak engaged.

[Vervil]

Mayan, maybe you could use multiband compressor if you have it just so you can control compression in those ranges where most of the "nastyness" occuress, to keep it under control. Then you can go to another compressor and then just a bit of limiting at the end.

[kritikon]

It seems like your making a confusion here. Limiting isn't a matter of knee but a matter of ratio (10:1 and more is considered like limiting).
Knee is the time that compression reach its maximum ratio once treshold is crossed and attak engaged.

Knee is extremely important in limiting. Many limiters that people seem to use as s/w are hard knee brickwall limiters. Doesn't matter much what ratio you're using above 10:1 if it's hard knee...a soft or hard knee will make more difference to the sound quality than a couple of dB difference on threshold etc.

Anyway - for some weird reason I posted in completely the wrong thread. I meant to post that in another one ...

However, it's still relevant.

Bugger...now I've got to go and look for the thread that I meant to answer to - it was by some chap who got clips on his master output and hadn't figured out it's easier to just pull down the channel faders...

But my answer was kinda related at least

[kritikon]

No, hang on....

This was the right answer to the right topic. I found the other one.

I'm confusing myself now....I think I've lost it tonight.


Anyways...compress it rather than limit it - it still applies.

...I think I need a rest...

[mayan]

Excellent advice...and precisely the direction that I've headed in. For the most part, I've learned (in the past) to try to get the mix just right and then LIGHTLY apply EQ and compression. I'd never really used a limiter before...but was somewhat seduced by the promise of "transparent" volume gain with Elephant. I've definately seen the dangers inherent in the path.