Skip to main content

Hi all,

Do you compress your audio signals before going into your DAW? Or limit?

I have an Aphex Expressor and use it to compress signals going into my DAW. But it occurs to me that it might be better to use the outboard Expressor merely for limiting (defeating much of its features) and then rely on plugins for non-destructive compression in-DAW. Is this a good idea?

Any advice is greatly appreciated.


Topic Tags


jammster Wed, 03/31/2010 - 12:06

I believe our Miss Remy posted a year or so ago about using compression to obtain better signal to noise into your DAW.

I believe you would need to know what context she was making a reference too. I know that she is dead on target about this, but how it is done by a professional engineer is beyond me.

Personally, I wish I could comment more about how that is done because I usually do not use compression during tracking, although I may want to use a tiny bit.

jammster Wed, 03/31/2010 - 12:22

Yes, by using compression we are thereby decreasing the dynamic range, this is why you need to pay close attention in setting your analog compressor properly so you don't squash it. I have squashed tracks in the past by doing this, I do not recommend it for daily use. A small amount of compression in the right place can make the track come to life.

From the article:

"In simple terms, a compressor is an automatic volume control. Using downward compression, loud sounds over a certain threshold are reduced in level while quiet sounds remain untreated. Upward compression involves making sounds below the threshold louder while the louder passages remain unchanged. Both reduce the dynamic range of an audio signal. This may be done for aesthetic reasons or to deal with technical limitations of audio equipment, which is seldom able to cope with the dynamic range the human ear can tolerate."

[[url=http://[/URL]="http://en.wikipedia…"]Dynamic range compression - Wikipedia, the free encyclopedia[/]="http://en.wikipedia…"]Dynamic range compression - Wikipedia, the free encyclopedia[/]

Dave_D Thu, 04/01/2010 - 09:34

Thank you, all! I was thinking somebody might say that my Expressor was better than plugin compressors and wasteful to employ as a mere limiter. I feel better now, especially given the fact that this new approach has eliminated much of the noise I was getting. That said, the Expressor is proving a bit difficult to setup as a brick-wall limiter. It appears to be slow, and the threshold setting is too sensitive. At best, I've dialed in brick-wall limiting at -6 dB. Any higher threshold and it peaks. Go figure. For max s/n ratio, I really want a brick-wall at -0.3 dB. Am I asking for too much?

Thanks again, everyone!


Jeemy Thu, 04/22/2010 - 12:21

There's enough headroom in 24-bit digital to track at -10dBFS and leave yourself no room for overs and plenty for upwards expansion in mix without major volume and artifact headaches.

That said, I usually track at -6dB peaks, maybe even -3, just through force of habit - my JLM pres have soft limiters configured to my RME FF800 and they catch any brutal drum transients in a very non-destructive manner. And I don't have enough outboard to compress 16 channels at a time on recording, well, I do, but thats another story.

If I am recording only a few channels however, I use my Buzz compressors with a very light gain reduction, so I am seeing the needles hover from 0-3dB reduction. I know I'm gonna use them anyway and I'd rather not patch em in later, I like how they sound and it gets everything that bit more "mix-ready". When you are working to the clock I think thats a sensible decision.

bouldersound Thu, 04/22/2010 - 13:41

Todzilla, post: 346725 wrote: With 24 bit recording, I find inbound compression to be unnecessary and "limiting" (pun intended) your flexibility with a mix.

Record dry and process later for effect. The days of inbound compression are behind us, IMO.

You don't need dynamics for "protection" during tracking, but they are still useful for aesthetic purposes. I am fortunate to work in a studio with two channels of Drawmer compression which I use liberally into the DAW.

Dave_D Mon, 04/26/2010 - 11:05

Thanks again, everyone!

I'll play with this a bit. Perhaps a very small bit of compression will improve the compressor's (limiting) response time without affecting the signal too much; especially with the side-chain control disabled. If not, it sounds like working at -6 dB is alright after all. Problem solved!


bouldersound Mon, 04/26/2010 - 11:42

Dave_D, post: 347053 wrote: Perhaps a very small bit of compression will improve the compressor's (limiting) response time without affecting the signal too much; especially with the side-chain control disabled.

I can't make any sense out of that sentence.

Dave_D, post: 347053 wrote: If not, it sounds like working at -6 dB is alright after all. Problem solved!

Set your tracking levels to peak at -12dBFS or even a bit lower. You'll never overload a mix bus and your analog noise floor will always be above the digital noise floor. If your pan controls are set to -6dB center your mix will land right around -12dBFS when things sum (depending on channel count and nature of the signals). When you are happy with your mix you can finalize it up to your preferred RMS level.

Dave_D Mon, 04/26/2010 - 15:32

bouldersound, post: 347058 wrote: I can't make any sense out of that sentence.

Thanks, bouldersound! Cripes, that is sorta confusing, isn't it.

What I meant to say is that, in order to preserve the signal going into the DAW, I've disabled side-chain compression.

Next, I was wondering if a wee bit of compression wouldn't improve the unit's attack time, being that a jump from 1:1 to 16:1, for example, would be slower than from 2:1 to 16:1. Maybe not noticeably. I'm just thinkin' out loud. I'll stop now. :-)

hueseph Mon, 04/26/2010 - 17:30

I'm confused still. What does sidechaining have to do with any of this? Are you ducking? Are you triggering something? What's going on? How does adding compression decrease the attack time of a limiter?

being that a jump from 1:1 to 16:1, for example, would be slower than from 2:1 to 16:1.

This whole sentence reads like gibberish to me.

Dave_D Mon, 04/26/2010 - 21:00

hueseph, post: 347079 wrote: I'm confused still. What does sidechaining have to do with any of this?

All I meant to say is that, in order to best preserve the signal going into the DAW, I will not use the Model 106's built-in sidechain section (which employs a built-in notch EQ). If, by contrast, I used the sidechain section to, for example, tame a nasal quality in a vocal, it couldn't be undone later. That's all.

This whole sentence reads like gibberish to me.

My logic is probably flawed, but I'll try to explain another way: If it takes 0.05ms for this compressor to reach 10 dB compression from 0 dB (no compression), then it should take only 0.025ms to reach 10 dB compression if the compressor is already compressing by 5 dB. Right? Now, I was extrapolating from this that more compression should effectively improve the compressor's attack time, but thinking more on it, perhaps not by any reliable measure. I suppose it may help if levels are consistent and you're driving the compressor hard, but a more dynamic performance is still likely to jump straight from 0 dB to 10 dB compression and overload before the compressor can react. Does that make sense? If it's still confusing, please don't bother - I assure you, it's not worth it. I could've tested (and rejected) my theory in the time it took to compose this. :-)

I apologize for any unnecessary confusion.


hueseph Tue, 04/27/2010 - 04:33

Sidechain won't affect your output anyway UNLESS you intentionally plug something into the trigger input. So, there's really no reason to even mention it. It's good to know you read you manual though. As fas as nanoseconds go though it's insignificant. Is a blink faster than a wink? We may never know.

Adding another process however will sure add more latency than a single compressor on it's own. As insignificant as that latency may be.

planet10 Tue, 04/27/2010 - 08:35

here is my thoughts......
Vocals i always hot just a bit of comp to the input no more than -2 db just to settle things down and even then i tel the singer to use the mic and their position to their advantage........
Bass, same thing, hit it with just a bit to take care of the transients enough to not over load your input to the converter and such........
Drums, NEVER
Acoustic Guitar, it all depends on the performance and what signal chain i have, but mostly its the performance that will dictate the use of a comp.

Big K Fri, 04/30/2010 - 05:08

Nothing against using a little compressor on a dynamic source to tame it a bit before going into the A/D stage. After all, we used to do this everytime when recording to tape. Just keep an eye on the settings and don't let it warp the original sound too much. That might cause a sub optimal, maybe even unusable recording.
I have stopped using any comps or limiters before the converter quite a while ago. Recording in digital domain doesn't need to be as hot as when recording to tape. There is hardly any noise introduced and the recording can be pushed up later to any level.

planet10 Fri, 04/30/2010 - 07:42

thanks for the comment on digital recording doesnt have to be that hot. time and time again when i get something to mix the levels are at or above 0....they're explanation is "thats what the guy at the store told me to do, make sure im at or close to 0 db so i get all 24 bits recorded".............GOD HELP US ALL

Big K Fri, 04/30/2010 - 10:09

The wise ruler up above should give those shop guys a crash course in audio engineering.
Those poor little lost

If they could only manage to get their recordings in the box without distortion, but ( closing my eyes..) they often don't even hear their little catastophies, themselves. But they learn.... I hope.
Same problem occurs from time to time when I get extern mixes for mastering. Compressed and limited ad nauseam. We all know, how hard it is to uncompress ;-)
It would be so nice if they asked how to do it before killing all dynamic in the music.
After all, we, too, have human ears, only, and they suffer a lot..