Skip to main content

Hi everyone, first post here, hopefully you can help me out a little. I have a fair bit of experience doing amature recording, and am pretty well versed in the use of audio gear, and the physics of audio (the old Yamaha Sound Reinforcement Handbook is my bible), so I don't think this is a total newbie question. I did check through the stickie'd thread on compression to see if this had already been answered, but I didn't see it there. I apologize in advance if this is something you hear once a month!

Q: What situations/desired results would actually benifit from the use of hard knee comp rather than soft?

I know the most correct answer is "whatever sounds best for that track in that song" but I've been doing a lot of experimenting lately, and even when testing myself by closing my eyes and pressing the soft knee button until I loose track of which setting I started on - I have a hard time telling the two apart unless I had my threshold really low and ratio really high (practically like a limiter). When I can tell them apart (but cannot tell which is which...) and pick the one I prefer, I always end up having picked the soft knee setting. And sometimes I just cannot hear any difference whatsoever.

In theory (meaning my guessing), I cannot think of a situation that wouldn't benifit from soft knee, it seems to always make things sound more natural and less obviously compressed.

So this really becomes two different questions for me.
1: Does anyone actually prefer hard knee in certain situations?
2: If I (or you or whoever) am compressing a track and I don't think I can hear the difference between the two - which do I pick as a default?

I've really tried to figure this out myself both through reasoning about it and listening carefully over and over, but I cannot figure it out.

Thanks for your help!


Codemonkey Fri, 04/24/2009 - 14:34

I recently read the manual for the Alesis 3630 Compressor (I think it was that). Some food for thought...

It has an automatic-knee mode. When the input signal is within 10dB of the threshold, it uses soft-knee mode. When it goes more than 10dB the threshold it switches to hard knee.

Can't say I understand this much myself, interesting to hear more opinions.

moonbaby Fri, 04/24/2009 - 15:30

OK. The first thing-the "knee".
This is where the input signal feeding the compressor goes above the threshold setting. In "soft" mode, as the signal crosses the threshold, the ratio of GR applied gradually increases to the maximum you've set. Let's say that you set the ratio to 10:1. Well, near the threshold, that ratio might only be 2:1, and as the signal gets stronger and goes further over the threshold setting, it will gradually go to 10:1. This is generally used for gentler compression applications.
In "Hard" mode, the set ratio will be almost instantaneous as the signal hits the threshold. This is more applicalble in "hard limiting" situations (like live sound protection), where you have to use a heavier hand in controlling dynamic peaks.
Both modes have their place, and the audible differences can be subtle.
Hope that helps...

AToE Fri, 04/24/2009 - 16:06

Thanks moonbaby, I did actually already understand the differences in what hard and soft knee do to the signal, but I've had trouble finding any situation where hard sounds better - especially in the more extreme compressor settings, where having that soft knee seems to help "hide" the compression, rather than make it blatant.

I can see what you're saying about live limiting, where speaker protection is important, benifiting from hard knee - but sound qualitywise would you recommend hard knee for any other scenarios?

Davedog Fri, 04/24/2009 - 16:08

I use the soft-knee application in the studio. Since you have much easier control in that environment theres no need for hard-knee other than as an effect.

Live, as Moon said, you have a much more difficult time (at times) discerning what is needing heavy handed control due to the many different signals and the relative levels at a live venue. As well as the babes walking around the FOH location. :shock:

AToE Fri, 04/24/2009 - 16:09

Guitarfreak wrote: Good info. Which one would be better to use for 'dynamic' compression? i.e. a singer that doesn't know how to stand still. Or more reasonably, one who doesn't know how to regulate volume.

The difference between hard and soft knee is pretty subtle, neither setting is going to drastically change the threshold and ratio settings (to my ears, I'mn sure others will chime in). I always use soft knee for vocals, I think it allows you to apply much more extreme compression without sounding obviously processed... but like I said earlier - I've been using soft knee for everything, so maybe take my advice with a bucket of salt. :wink:

soapfloats Fri, 04/24/2009 - 18:47

My understanding parallels some of the above: the hard knee would be better served in limiting situations - where you don't want the signal to go much above that threshold, if at all.

The only time I use it the box is at the end of the ST bus chain - on my master limiter to make sure an errant snare drum crack or cymbal crash doesn't pass that ceiling where I want the sound level to stop.

Most of the mastering engineers I work with want the peak db of a track to be in the -3 to -6 range. I'll have it safely in that range when SOMETHING (usually drums) punches that peak up to -1db or higher. Then comes the hard knee (and higher compression ratio.

I agree - soft knee tends to "mask" the compression better and sounds more natural. There is a reason almost all dbx comps either exclusively use soft knee or have it as an option.

AToE Fri, 04/24/2009 - 20:21

Ha, and here I was thinking I'd missed something massive! Thanks for the comments, this is pretty much how I'd been thinking about it already. I guess I made the mistake of assuming that because most comps I've seen have hard knee standard and soft as an option, that that meant that everyone was using hard knee all the time. I guess I should have just trusted my ears.

So the verdict seems to be: that there aren't really any/many situations where hard knee actually sounds better (or these are at least rare) than soft, it's just more usefull at preventing the odd clipping.

Thanks again.

IIRs Sun, 04/26/2009 - 17:54

Hard knees can be more precise and surgical than soft knees. If you are trying to set the compressor to 'catch' very specific elements of a sound without affecting others, a hard knee might be better. An obvious example would be if you had an filter in the side-chain for de-essing or de-popping. But also for controlling the finger clacks of an enthusiastic bass player, or for precise control of a drum's dynamics.

If you just want transparent gain reduction however, a soft knee will almost always do better.

anonymous Sun, 04/26/2009 - 19:38

I prefer the multi-knee approach. As in hard limiting and daisychaining light comrpession until I get rid of clipping and distortion (if I do indeed want to get rid of those things). I really don't pay attention to whether a compressor is in hard knee or soft knee mode, heck, I don't even think any of my VST or analog compressors have a soft or hard knee mode. :P