Adventures in compression (always something new to learn)

[Starship Krupa]

I've been working on my mix engineering chops pretty seriously for about 8 years now, and one of the things I love about it is how there's always something new to learn or insight to be gained, no matter how many hours I spend on my own mixes, watching and reading tutorials, taking online courses, whatever.

Today's insight:

Last night I came up with a nice little 3-note Krautrock-style arpeggio. This morning, I came up with a bass line to go with it. I also installed Acon Digital Dynamics BE, which I got with the latest issue of Beat. Nice dynamics processor, nothing that I didn't already have, but I wanted to try it out.

So I have my arp arping away in the top end, and my bassline going on down at the bottom, and I throw Dynamics BE on to see what it can do. I put it on the bass line, call up one of its "1176" presets and tune the threshold to try and put some bounce into it. Not much going on there, so I drag the plug-in up to the arp track.

I tune it to get it bouncing, and I momentarily forget which track I have it on, because suddenly the bass line is tickling my ears in a most pleasant way. So I'm wondering what I did to make it work on the bass line and then remember that I don't even have it on the bass line any more, that putting it on the arp is making the bass line jump up and down.

Makes complete sense, the high arp is now getting out of the way of the bass, but still, quite a moment of revelation. Compression is such a trippy thing.

[Tim Smith]

 

Using one track's compression to indirectly influence other tracks in a mix isn't something I had ever considered but I guess it makes sense. 

I once WAY overused compression in general. If I use compression on an individual track I'm most often looking only to 'tickle' any stray transients with a fairly fast attack. I have found  some compressors can easily goof with the track in bad ways, so I only go for track compression if it seems absolutely necessary. It's easy for me to ruin a piano track attempting to remove the hard attacks.

I suspect you might have also had some success with that mix if you had side chained into a compressor . You must have had it set perfectly to bring the bass line out.

[Starship Krupa]

1 hour ago, Tim Smith said:

Using one track's compression to indirectly influence other tracks in a mix isn't something I had ever considered but I guess it makes sense. 

It only made sense after I gave it some thought. One way of looking at a compressor is that it takes something away to bring out something else. It can help make a track "pop" and it can also push it back a bit. So if my compressor setting was knocking off the attack (I don't remember at the moment what the settings were), it would leave space in the mix for the bass track's attacks and then go back to sitting on top of the bass track. Or the other way around, the arp might have been stomping the bass track's attacks and letting the sustain bloom.

It's like EQ carving; dipping some frequencies out of the piano track to let the guitar track come through (something else I stumbled upon by accident and finally understood).

1 hour ago, Tim Smith said:

I suspect you might have also had some success with that mix if you had side chained into a compressor . You must have had it set perfectly to bring the bass line out.

Well, that's a challenge, trying to figure out just what I did to get that result. 😄It was serendipitous, to be sure, I was just turning knobs to see what I would get. A "mix with my ears, not my eyes or brain" moment.

One issue with sidechaining is that it can be a "blunt instrument." Even if I filter the sidechain input, it still ducks the entire full spectrum of the target track. Here is where I sing the praises of Trackspacer, but I still think that I wouldn't have gotten the same results with Trackspacer without maybe opening the panel and messing about with the attack and release.

Moreover, it was a good lesson and reminder that every single element of a mix affects the other elements.

Edited April 19, 2022 by Starship Krupa

[Tim Smith]

I'm mostly side chaining bass frequencies and setting those frequency limits when I side chain. I'm guessing you mean there isn't much control over the send to the side chain?

I guess you could post send an eq'd version of the track by putting an EQ post. Might be making things more complicated than they need be too. This is interesting because I had been putting my compression in a box so to speak. I only seen it as either a track plugin or a master plugin, IOW what does this TRACK need as opposed to how can I use it to affect the entire mix.

I had not heard of Trackspacer ? I'll need to check it out.

Thanks!

[Starship Krupa]

35 minutes ago, Tim Smith said:

I'm mostly side chaining bass frequencies and setting those frequency limits when I side chain. I'm guessing you mean there isn't much control over the send to the side chain?

What I mean is that whatever track you have the compressor on, you can filter the sidechain signal going into it all you want (freebie MCompressor has this built-in), but when the compressor is triggered, it affects the entire track, not just the bass.

I tend to use basses with a lot of triangle-wavy oscillator sync'd goodness in the lower (and even upper) mids. If I just do a simple sidechain on it with the trigger signal coming from the kick, yeah, works fine, but it ducks my entire bass track, funky blurpiness and all. When mostly it doesn't need to unless I'm trying for a Daft Punk effect. I just want the low end of the bass to get out of the way, not duck the entire spectrum.

The two plug-ins I know that can handle sidechain ducking of only a specific frequency range are Trackspacer (get it from Pluginboutique when it periodically goes on sale for $29), and MSpectralDynamics, which I also have but haven't entirely gotten a handle on.

[Lord Tim]

The other thing to think about is:

1. Does it *need* compression? A lot of times we throw a compressor on a track because... well, that's what we do for that instrument, right? Maybe carving out space with EQ and some clever automation might be best?

2. Don't forget you can also use compression as a transient shaper too if you play with your attack and ratio, so you could get things sounding even more spiky rather than doing the opposite.

A multiband compressor or dynamic EQ can be amazing on a track or bus too. I pretty much always have one on my guitar bus to tame the woofiness in my rhythm guitar tracks because that really builds up once you layer them, but if you just EQ'd it out, all of the guts would fall out of the sound. Doing a dynamic EQ sidechained to the bass to drop the lows on the guitars is another great trick because your ears trick you into thinking the guitars are super fat when it's actually the bass filling out the frequencies, but when it stops you still have the full thud of the guitars in the mix.

[Tim Smith]

10 hours ago, Starship Krupa said:

What I mean is that whatever track you have the compressor on, you can filter the sidechain signal going into it all you want (freebie MCompressor has this built-in), but when the compressor is triggered, it affects the entire track, not just the bass.

I tend to use basses with a lot of triangle-wavy oscillator sync'd goodness in the lower (and even upper) mids. If I just do a simple sidechain on it with the trigger signal coming from the kick, yeah, works fine, but it ducks my entire bass track, funky blurpiness and all. When mostly it doesn't need to unless I'm trying for a Daft Punk effect. I just want the low end of the bass to get out of the way, not duck the entire spectrum.

The two plug-ins I know that can handle sidechain ducking of only a specific frequency range are Trackspacer (get it from Pluginboutique when it periodically goes on sale for $29), and MSpectralDynamics, which I also have but haven't entirely gotten a handle on.

Oh I see.

One option I took note of so far as Bass control is concerned is to select different basses. One with a high end you really like, the other with the best low end playing the same part. If I had a studio recorded bass track I might make an audio to midi copy of it. At that point that sky is the limit with what could be done using sampled basses either alongside real bass or in addition to. I might roll off below a certain freq. on one and roll off above a certain freq. on the other. You could even treat the same bass differently using two ( or more) tracks with this technique.

What  I hear you saying ( and the reason to do it in the first place) would be we like what we hear on the track most of the time so we don't want to loose it. Other tracks could potentially step on it effectively masking a desirable sound. Two tracks, similar frequencies competing. Possible solutions:

- Old ducking side chain technique, which as you pointed out, can kill the dynamics in the whole track. If there wasn't anything to kill by side chaining, then it might work ok. Cakewalk has a decent master compressor with side chain capability for anyone who wants to experiment with it. I think this technique works best with a light touch. IOW low compression setting only knocking back a few db.

-Multiple tracks of same or similar instrument and parts cut up various ways using EQ with the offending freqs minimized through individual track treatment. The advantage over multi band compression is we have more control over individual elements using any plugin or method we desire. The disadvantage is it can get complicated all depending. Too much treatment and there could be phase issues, alignment issues etc. due to delays through plugins. A simple approach to this is probably best.

-Shifting frequency emphasis or de emphasis on offending tracks using EQ. Dissimilar tracks don't collide. I have found using this technique never completely wipes out the collisions because they almost always overlap slightly. Cutting them completely can sound artificial. It certainly helps, and might help enough to slide a mix into ok territory. This is usually the first thing I try.

6 hours ago, Lord Tim said:

The other thing to think about is:

1. Does it *need* compression? A lot of times we throw a compressor on a track because... well, that's what we do for that instrument, right? Maybe carving out space with EQ and some clever automation might be best?

2. Don't forget you can also use compression as a transient shaper too if you play with your attack and ratio, so you could get things sounding even more spiky rather than doing the opposite.

A multiband compressor or dynamic EQ can be amazing on a track or bus too. I pretty much always have one on my guitar bus to tame the woofiness in my rhythm guitar tracks because that really builds up once you layer them, but if you just EQ'd it out, all of the guts would fall out of the sound. Doing a dynamic EQ sidechained to the bass to drop the lows on the guitars is another great trick because your ears trick you into thinking the guitars are super fat when it's actually the bass filling out the frequencies, but when it stops you still have the full thud of the guitars in the mix.

Sometimes I think the modern listener is so accustomed to leveled material, they think a level uninteresting track is what they need. Anything that jumps out scares them. Do we want dynamics, lack of dynamics or something in between?  This is where plugin automation can be so important IMO. Many 'set and forget' a plugin never using the potential the plugin has for automation. Maybe I want the acoustic guitar to really stick out until the synth comes in. Then I want the guitar and the synth to carry it equally. I might do better to automate a compressor or EQ setting than to automate volume. Admittedly for me it's a time saver to automate volume over finding a compressor I like and mapping it to my controller. I wouldn't go so far as to call it laziness but it's close. The nice thing about the Cakewalk Pro Channel is you can automate EVERYTHING, yet I think many most often only automate volume. Templates would help since we can use the same template over and over again. We can map a controller and save it to template. I don't tend to copy mix settings across songs but that's me.

I was using MDynamic EQ, the Ozone dynamic EQ's for awhile. Then I went to Fabfilter and often use the dynamic features in it. I go back and forth between several for this application and they really work well for what they do. Especially nice for having a visual of the mud, being able to monitor only that section ( fabfilter) and dynamically control it. 

With 20 guitar tracks on a metal mix ha ha....no it isn't funny. It's a huge mix within a mix and yes, those tracks ( or any other tracks) have a cumulative effect on the mix bus. I see it like the way a pencil artist makes a rough sketch with undefined lines as compared to a finely drawn picture. How do you take multiple guitar tracks and make that fine line? This was one of the challenges of mixing metal for me. We have lack of definition and we have mud buildup if it isn't tamed. I'm sure many engineers have spent hours lining up audio and removing phase issues. Not sure of your exact approach, I wanted some dissimilarity because it adds dimension and character. I didn't want to go too far and get a bunch of out of sync garbage either. It always seemed to be a balancing act. Fix one thing and something else pops up.

 

[Lord Tim]

For me, there's always 2 things I keep in mind when I do anything:

Consistent bass - if it's something that's really dynamic like a bass guitar or a really fat modulating bass synth, I'll usually always flatten the hell out of that. It's not the most interesting thing in isolation but the problems you get by the time it hits the mix bus is exponential. It can make the entire mix sound uneven and if you plan to really squish it at the end, unless you're taming everything with a multiband (which has its own issues because that'll also tame things like kick drums and sub hits, etc.), your mastering limiter is going to give you the stink eye as it absolutely ruins your project on the way out.

And guitars, yeah - I never clone anything, if you get stuff "perfect" it actually sounds smaller the more you layer onto it. But I'll always edit everything to be pretty tight, either cut out or automate silence into the gaps, and then bus that with a multiband on it, and then decide what I want to do in relation to the other stuff in the mix with low frequencies. I'll more often than not play into amps or amp sims with too much high or low end for what the mix needs because it feels good as I'm laying the track down, but then I'll do some serious carving after that - what ends up on the record is usually surprisingly thin when it's solo'd, but you'd never know in context. I never really worry about phase issues though, I've only found that's an issue if you're cloning anything or doing split sends from a single DI guitar track into multiple amps with different IRs or mic positions, which I'll never do, it just doesn't sound big or exciting enough.

For my main band, I don't think we have anything less than about 8 or so heavy guitar tracks (not necessarily playing at once all of the time, mind you) so it does start to pile up. Layer in bass, bass synth, low strings, fat polysynth... wow, yeah good luck with that! HAHA! Add in the 60 or 70 layers of vocals over the top and... yeah, check me into the asylum now, please. 🙄 On the other hand, my thrash/death band is a different kettle of evil because there's so much going on with kicks and guitar low end, that needs to be tamed or it rapidly spins off into the block of festival toilets.

Edited April 20, 2022 by Lord Tim

[Tim Smith]

Well you are quite talented there @Lord Tim and must be busy with those bands. 👍I checked out some of your material and I think it's mixed pretty good. I tried to comprehend the lyrics but as it's poetry it's pretty deep I guess.

[Lord Tim]

46 minutes ago, Tim Smith said:

Well you are quite talented there @Lord Tim and must be busy with those bands. 👍I checked out some of your material and I think it's mixed pretty good. I tried to comprehend the lyrics but as it's poetry it's pretty deep I guess.

Ha, thanks! You can blame the lyrics in the thrash band on the other guys - I was kind of added as a guest that somehow got promoted to vocalist when their last guy left. Hell of a tongue-twister band, I'll say that much!

LORD is fairly straightforward for the most part, with a few fun curveballs. If nothing else, we love to rip out a fun cover, and I have to say that stuff really makes you think about production in a very different way, and how you can apply it to other styles. There's been a lot of "but what the hell are they even doing?" moments that have turned into "a-haaaaa" learning moments for me over the years.