Mixing With Massive Bus Comp Gain Reduction

[jangoux]

I think the other main component you're forgetting (or simply maybe overlooking), is that guys like Randy who do this compress the shit out of everything before it even hits the compressor as well.

Nothing really pops out before he hits the buss comp, it's all squeezed within an inch of its life. That way when you look at his bus comp meter, it's more floating within a 1-2 db range then pumping with the music.

Exactly. After i read this post, I went to the studio to track vocals for a band. It's kinda Alice in Chains-esque, so i thought lots of compression on the master bus would work. I squeezed the fuck out of the drums (compression on each track - Waves SSL Channel for kick, snare, toms, The Glue for OHs, THe Glue for drum buss reducing around 1 to 2 db, + parallel comp for snare and drum buss ; Everything clipped around -2db and yet a little saturation on the drum buss. Besides, shitloads of compression on vocals: -8 db on Waves CLA3A and -10 on CLA76). When everything hit the 2buss, it was WAY less dynamic, and this allowed me to hit the GSSL hard, and I ended up getting up to 8db of GR. My settings were 2:1 ratio, attack 30ms and Release 0.3. I could get more, but the kick low end was distorting things a little bit.

It sounded very modern, the toms sounded AWESOME, but it wasnt really a done mix, so I think i can improve it further. Until then (the mix will only be done by mid August), unfortunately, I cannot post the results.

Btw, I always had a problem with the GSSL narrowing things, but with these settings, I got a super wide 'mix'.

So, If anyone is willing to try, as CHFH13 said, squeezing the fuck out of things MAY help achieve good results.

 

[Teddyboy]

Hey, Jangoux, it sounds really interesting. Please post some clips when you're able to. Have you tried Auto release on 2buss?

So,if I understood correctly, you try to squeeze each element of the mix, so it is pretty flat with little to no dynamics so you can achieve so much gr (8-12)?
What if the music is really dynamic?

 

[slashvanyoung]

I was under the impression that Ermin already smashes the living hell out of nearly all elements with staged compression before they hit the 2buss...

 

[jangoux]

Hey, Jangoux, it sounds really interesting. Please post some clips when you're able to. Have you tried Auto release on 2buss?

So,if I understood correctly, you try to squeeze each element of the mix, so it is pretty flat with little to no dynamics so you can achieve so much gr (8-12)?
What if the music is really dynamic?

Yes, I tried Autorelease. On 0.1ms, the max i got was around 4db of reduction. The more i tried bigger release times, the most gain reduction i got, but also lots of pumping. To be honest, I didn't try a lot of different settings because I was tracking vocals, so I needed to pay attention to what the guy was doing. But by crushing everything before the 2buss, I could get this result. It's a pity the GSSL we have there came without the sidechain feature. I feel I could be getting a bit more gain reduction if it wasn't for the kick low end.

Btw, I also believe Ermin squashes everything before the 2buss. What we may be experiencing, is the behavior of different builds of the GSSL. The one that's on the studio i work on is made by Fabio Baumann, a known guy on the Prodigy Pro forum (and creator of the 'Fetboy', a fet-based preamplifier). He has a company that has loads of nice hardware replicas, including APIS, Neves, 1176s, GSSLs and some other cool stuff.

 

[Ermz]

I was under the impression that Ermin already smashes the living hell out of nearly all elements with staged compression before they hit the 2buss...

I do indeed, and especially so when attempting a technique like this. Always make sure to mix into the stereo comp from the start too, so I can 'build' the mix into it.

Don't know really, I'm just not having any luck with more than about 4dB of reduction. My unit starts to pump very audibly. Agree with jangoux that the kick is the main instigator. I can't help but feel a side-chain filter built into the unit would really help.

Thanks for all your responses so far.

 

[AboutBlank]

Just a little question:
It seems that all of you do pretty much on the 2 bus at the mixing stage...

I totally get that it's part of the "mix", but I wonder a little what mastering dudes say. I've mixed into an comp for awhile, but then stopped till now and left those things to the "mastering".
question of faith?

Or is it hunting for a special style?

 

[jangoux]

Ermin, What if we do our own 'sidechain' ? I mean, create a bus for everything BUT the kick, compress the fuck out of this group, and mix the kick to taste, sending this mix to the 2buss and then apply another compressor, just to glue them together ? I can try this tomorrow and see what happens. I think i can squeeze at least a couple fore Dbs of gain reduction with this. Dunno if the toms will also make it pump

 

[crillemannen]

Just a little question:
It seems that all of you do pretty much on the 2 bus at the mixing stage...

I totally get that it's part of the "mix", but I wonder a little what mastering dudes say. I've mixed into an comp for awhile, but then stopped till now and left those things to the "mastering".
question of faith?

Or is it hunting for a special style?

Well i think that applying compression on the mix bus is a way to add character to your mixes, and i think almost every professional do.
It's definitely a part of my style for sure

I tend to eq and do the foundation on my mixes before i add master bus compression. But after that i put a chain of compressors on, and then mix into the bus compression. But then again i don't smash the master bus, i usually do 4db of GR.

From what I've read Randy smashes the bus with up to 8db (as ermin said) and its part of he's sound, so whatever works for you as a producer/engineer.

But then again let's hope someone can jump in and give us a rack compressor that can handle massive GR

 

[Ermz]

It sucks that the GSSL doesn't seem capable of it. I was under the impression that it was almost identical in layout to the G series bus comps. Sort of hinders my aims back here a bit... and since I'm going to be blowing out on converters and distressors soon, I won't have money for a store-bought SSL clone any time soon.

Here's to hoping we find a way around it. Till then I'll be mixing my ass off to see if I can figure it out, and make sure the problem isn't with what I'm feeding the comp, rather than a limitation of its topology.

EDIT: Going back to the first reply... I find it hard to concede that it can only be done using the real console. Sure, maybe the meter ballistics are different, but the sounds should inherently be very similar, even if one is showing 8dB reduction and the other only 4, provided their internals are doing the same job. The GSSL, afaik, virtually is the bus comp from an E or G series desk just ripped right out. With AlexB's SSL saturation programs I'm getting that elusive 'width' and smearing that wasn't possible in ITB before, so the only options as far as I see it are:

1) My mixing is fucked and I'm not hitting the bus comp with enough control or whatever.
2) The meter ballistics are off, and when I'm getting the same results as 'the big guys' it meters more like 4dB rather than 8.
3) This particular model of SSL clone, any rackmount SSL clone etc. etc. isn't capable of achieving the same results as the ones in the desk for whatever reason.

To have someone rationalize this... or to have a way to scientifically show which is most likely the case would be immensely helpful in proceeding further. There is a lot of hearsay on the net, and really I'm not sure if anyone other than one of the aforementioned 'big guys' who do this regularly could elucidate. Perhaps in the mean time we could get someone to provide a source mix... something slow, strong and radio-rocky, and then we all hit it at extreme settings with our bus compressors and see which does the best job, and if so, whether it's enough of an improvement to be attributed to anything other than ballistics and or mixing skill?

 

[Aaron Smith]

I just have to say, this is one of the best threads I've read in like, a couple years.

 

[gabriel g.]

first of all: amazing thread!!!!!!

I think the architectour of the mixing desk plays a important role.
think about that, you can push high-quality mixing desk super hot so the individual channels lower the dynamic of the source. then there is commin phase distortion, harmonic content and unharmonic content in the chain. this and the heavy bus compression makes the sound you are after IMO.

As far as I know that nebula stuff is based on impulses (mathematicly rebuild of an dirac to get the system answer...)
this technic will never be able to reproduce UNharmonic content and compression, so maybe thats the reason it wont work Ermz.

I am so waiting for the slate console because it sounds like that they recreating a more complex copy of that sound, WITH not linear phase distortion and unharmonic content

 

[Ermz]

I think you may be oversimplifying Volterra Kernels, which is what Nebula is based on. If they were just simple IRs I would have noticed, and certainly would not be using them. The Nebula saturation and EQ programs are by far the best currently available ITB. I haven't heard anything algorithmic even touching them. Given that Slate's blind A/B samples with a real SSL desk were so close to his own algorithmic implementation, I have no doubt that these Nebula programs are bringing the goods. Honestly I doubt it's any of that 'the desk has magic mojo, blah blah' stuff. It has to do with the bus comp and the mix that's feeding it. Either the mix is unbalanced, or the bus comp just isn't built to handle that much GR.

We've heard this sort of thing with comparisons in the past. When Lasse recently did one you could hear the SSL comp getting more grabby and aggressive, whereas his Alan Smart was more gentle, and as a result more happy to take higher amounts of GR on a whole mix. The problem with the GSSL in particular is that it gets REALLY grabby on the kick, to the point where the mix just starts swinging up and down in time with it. Gives me a headache when mixing through it heavily!

 

[Morgan C]

Don't have much to add as my experience with outboard equipment is nil, but I remember Bendeth posted on the Gearslutz forum about his masterbus settings for the Underoath album. I remember 3ms attack, auto release and 8db GR but nothing else, I'm sure you could find it, his username is stereomixer IIRC.

 

[MortenDK]

This more of an anecdote, not sure it will bring anything to your search for an ansver but!, some years ago back in the day here in Denmark we had a big festival called "Midtfyns Festivalen" and my collegue Jakob was recording this band(I work for the Danish national Radio and TV station here in denmark as an assistent sound engineer) anyway, Daniel Lanois is going on after this band Jakob is recording for the radio, and Daniel comes into the bus(mobile recording) and asks if he can sit in with Jakob while hes recording/mixing this band.
After awhile Daniel than asks if he can be recorded too!, so Jakob goes sure!, than Daniel explains to jakob how he does his mixing for his band and shows him how to mix his band with 8-12 db GR on the mix buss, we have SSL 4000E/Gs in our mobile recording vehicles.
Now the GR on the channels was something between 6-12 db of GR.
The interresting thing was he than used the expantion in the console to get some life back into the intruments by using the expantion as "breather" and setting the expantion up with the attack and release to follow the instruments. This than went to the mix bus with 8-12db of GR. And it sounded really good!.
Now this is all from Jakob mouth I wasnt working at the Radio at that time!.
But it was cool to hear how much GR he used on the individual channels, and the expantion thing was really cool as well I thought!.
Just a story really, but try the expantion thing it might work!.

Morten

 

[jangoux]

Well, I played with the GSSL a little more and the results were:

* On 0.1ms release, I couldnt get more than 3-4 db reduction
* On 0.3ms I could get up to around 7 or 8db on the loudest parts
* Autorelease didnt do it for me. Too jumpy.

I didnt really mix the song, just 'organized' it and got these results. I think on a song with a thinner kick, you could get a lot further. I also tried bussing everything but the kick to a buss with the GSSL, and the sending this buss and the kick to the stereo out. This didnt work for me..it was too weird not having the compressor holding the kick. It was like pizza without cheese, or a woman without breasts. It just isn't right! On this particular song, even though I didnt have the kick on the buss with the GSSL, i could get 3-4db MAX GR - and that surprised me a little bit. That lead me to think that you need a 'perfect scenario' to be able to achieve more than 4db GR, let alone 12db! Besides the massive experience those guys have, they probably get stuff to mix that is more 'ready' to mix than most of us get.

Btw, I won't be able to post anything until mid-august. The studio will not work for the next 20 days, and then I'll track this guy that supposedly played bass for Jimmy Hendrix on his youth, JJ Jackson. So around august 20, I will have time to work deeper on this.

 

[jangoux]

Btw..times > 0.3ms got me more gain reduction but also got fucked too much with the kick and caused distortion.

 

[Mikaël-ange]

@Ermin:

-First are your sure Randy use something like 8db GR?
I said that because (with my little experience with 4K E/G and G+) GR reading on SSL are often like alchemy You can have 4db GR with one and put 8db GR on an other 4k for get same result you get with 4db GR on first.

-Are you sure Randy use quad bus (aka SSL comp) for is 2bus?
I say this because most common mistake I hear is how hard CLA hit the quad bus on his 4k. But CLA use quad bus for drum not 2bus (and for all I know better than you guy flying around: learn the goddamn manual, rent a SSL 4k room and stop speaking about shit you didn't understand/know).

Because I've been running into the same problems (and I'm on an SSL!) I've actually started working with the Michael Bauer ABCD multi buss compression instead and I'm liking that a lot more! My setup is WAY simpler then his, but I can definitely see the benefits.

What did you run for Braunerize?

@mjlaudio: Nice to see more Geraslutz guys here

 

[C_F_H_13]

@Ermin:

-First are your sure Randy use something like 8db GR?
I said that because (with my little experience with 4K E/G and G+) GR reading on SSL are often like alchemy You can have 4db GR with one and put 8db GR on an other 4k for get same result you get with 4db GR on first.

-Are you sure Randy use quad bus (aka SSL comp) for is 2bus?
I say this because most common mistake I hear is how hard CLA hit the quad bus on his 4k. But CLA use quad bus for drum not 2bus (and for all I know better than you guy flying around: learn the goddamn manual, rent a SSL 4k room and stop speaking about shit you didn't understand/know).

YEs and Yes.

What did you run for Braunerize?

@mjlaudio: Nice to see more Geraslutz guys here

A- DISTRESSORS-NEVE 1084
B- SSL QUAD BUSS- Tube Tech EQs
C- 1176's
D- Focusrite Red 3-API 550a's

A-Gtrs
B-Drums/Bass
C-Extras
D-Vocals

 

[krokit]

I mix through a Smart C2 with the sidechain (which is set at 130hz) 4.1, 30ms attack and 100ms or auto release and after about 4db of gain reduction it starts to get a little mushy. Without the sidechain I couldnt use it on the master buss at all really.

The 'expansion' thing MortenDK mentions in his Daniel Lanois anecdote is interesting!

Also, I think that the 100hz, 300hz, (if I remember rightly) boosts that have been noted on Randy Staub mixes could be another clue as to how he achieves such high gain reduction. If the 100hz and 300hz boosts are pushing against heavy compression , wether sidechain or conventional, then maybe the low end is flattened and controlled enough to not pump the masterbuss comp??

Ive heard of some mix guys who mute the bass at the start of the mix and only when the mix is almost finished, unmute it and fit it into the finished mix! Mmm, might give that a shot myself.

I couldnt get anywhere near 12db gain reduction unless I was compressing in parallel.

Dunno, just thinking out loud really!




http://www.ascapestudios.com/

 

[Aaron Smith]

Ive heard of some mix guys who mute the bass at the start of the mix and only when the mix is almost finished, unmute it and fit it into the finished mix!

That's a really interesting idea. Makes perfect sense, but I've never tried such a thing. Definitely gonna see how it works out when I get some time!