• New Sponsor: ShipNerd, Ship Your Gear with Us... for less! Click Here.

Mixing is hard!

Teal_66

Member
Messages
3,316
Well, despite these tough audience posters I know exactly what you mean. It is voodoo in the sense that it is not intuitive. The better you get at it the more you will see it as art instead of science, just like voodoo.

I like to put on my Dave Grusin cd "Out of the Shadows" as what I consider one pinnacle of sound engineering. After all these years I still can't figure out how to make my stuff sound like that. It's just mind boggling to think I might have the tools for it at my thumb tips and I just haven't reasoned out the lucky recipe. But it still makes me better as I go.
I've never heard of Dave Grusin or "Out of the Shadows", so I checked it out. Yes, that sounds fantastic. You have to look at the time period (1982). No computers and extraordinary musicians. I also think a lot of this is the actual room they were in too. It's like some of those old Motown records have a sound nobody can duplicate. I wonder if the Grusin guys were all playing together with mic bleed and all? It's a lot of the little things that all add up. It had to be quite the console too.
 

Yr Blues

Silver Supporting Member
Messages
2,814
I've never heard of Dave Grusin or "Out of the Shadows", so I checked it out. Yes, that sounds fantastic. You have to look at the time period (1982). No computers and extraordinary musicians. I also think a lot of this is the actual room they were in too. It's like some of those old Motown records have a sound nobody can duplicate. I wonder if the Grusin guys were all playing together with mic bleed and all? It's a lot of the little things that all add up. It had to be quite the console too.
You can't discount the infinite factors that we can't control. I guess that's what makes it so fun.
 
Messages
1,988
I thought 4:1 was standard. I'll try it.
Note that on the SSL bus compressor, the knee changes with the ratio. The 2:1 is a very soft knee, which compresses lightly but constantly, whereas the 4:1 ratio has a harder knee that, to my ears, pumps a little more but only touches the transients. So, counterintuitively, the 2:1 seems to be doing MORE apparent gain reduction than the 4:1, even if the GR on the meter is in the same vicinity.

Others could probably explain this more eloquently, but that threw me for a loop for a while until I read about the knee characteristics of the SSL.
 

louderock

Member
Messages
5,062
Make everything louder than everything else.

Try to listen to the mix as a whole and not individual sounds. If you fixate on a single thing too long, you’ll lose perspective. Just write your name on a piece of paper and then read it out loud for 30 seconds. Sounds like the dumbest word ever, right?

You don’t have to process every single sound just because you can. Get your drums and bass sounding good then slowly start adding the other tracks with no processing. How do they relate to everything else that’s going on and do you need to do anything to them with processing to make them fit in the big picture better.

Take breaks where you’re just listening to the mix as background music. Walk around the room or surf the internetz while NOT staring at the DAW screen.
 
Messages
2,695
I like to approach it from the perspective of telling a story to the listener…I keep in mind what specifically I want them to pay attention to at any one point in time. Most listeners won’t hear more than a couple things at a time. One way to keep tabs on that is to turn your monitors way down low…is that element still audible? If not, you have a masking problem or a level problem. But you need to figure out which one it is, and make the right adjustment.

Listen to the track and make a note of what your ear is drawn to…but do this as a listener, not critiquing your own mix. Play up these elements. It’s too easy to get sucked into a mentality of “everything needs to be heard”.

Oh, and here’s one secret to a good mix…they are not balanced! A mix should create an emotion in the listener. Play into that, let it be your guiding light.
 

soundchaser59

Thank You Great Spirit!
Gold Supporting Member
Messages
13,028
Note that on the SSL bus compressor, the knee changes with the ratio. The 2:1 is a very soft knee, which compresses lightly but constantly, whereas the 4:1 ratio has a harder knee that, to my ears, pumps a little more but only touches the transients. So, counterintuitively, the 2:1 seems to be doing MORE apparent gain reduction than the 4:1, even if the GR on the meter is in the same vicinity.

Others could probably explain this more eloquently, but that threw me for a loop for a while until I read about the knee characteristics of the SSL.
And this is related to why I say compression should be seen and not heard. I know people love to say "use your ears" but in my experience compression is the one treatment that should not be heard.

What I mean is I used to make it so that it was enough when I could just barely start to hear it working. But then I walk away for a week and come back and all of a sudden I could hear it too much and I start thinking "what did I do that for? It's way too much!" Then I look at the meters on the comp and the lights were telling me I was close to squashing it when all I really wanted to do was prevent instantaneous spikes.

Hard to explain but I ended up concluding that with compression I was actually better off trusting the meters, because if I could hear it working it was almost always too much compressor.
 

Kronos147

Member
Messages
487
And this is related to why I say compression should be seen and not heard. I know people love to say "use your ears" but in my experience compression is the one treatment that should not be heard.
I agree with soundchaser, good compression is not (typically) 'heard', only the artifacts of a poorly set compressor are (typically) heard.

Compression is one effect that can make things sound 'wrong'.

MSNBC made a change recently. They have some kind of brick wall limiter on the audio output. It has a slow attack and a really slow release, and the threshold is set wrong. Audio pops out until the slow attack makes it all quiet and then it stays like that until there is about a 1 second pause. Then the cycle renews. It makes their station unwatchable to me right now.
 
Messages
2,695
And this is related to why I say compression should be seen and not heard. I know people love to say "use your ears" but in my experience compression is the one treatment that should not be heard.
One thing to keep in mind if you want to really 'see compression' is that the sound of a compressor happens when the gain reduction meter is moving...mathematically, it's the derivative of the GR meter deflection.

If the meter is pinned at say 10 db GR and not moving, you're not really hearing compression until the release starts to occur. On compressors with soft knees many engineers use this to their advantage and can push some pretty high GR values without the signal sounding extremely compressed; they're really just trying to push into the compressor to find the slope they're looking for.

For SSL bus type compression/glue I use very little GR. Just get the needle to bounce a bit in time with the track. I'm probably around 1dB on the meter or so. Often I'm using the auto release, but sometimes I'll use one of the fixed settings if it works better with the track. 4:1 and slow attack is my go to here.
 

Stokely

Member
Messages
1,683
"Hearing" compression can be a production choice though. Sometimes people want a more squashed sound.

But in general yes I tend to go for more subtle. One trick is to combine several compressors in series with each one doing only a little bit (vs one doing the same amount of overall compression.)

I recommend setting up busses (groups). In Logic Pro, I've been using the summing stacks for this. Not only does it help with organization (you can visually collapse them) but any processing you might want to add to all the members of the group you can do to the buss (typical example is an effect on the drum buss). It also helps you when you have a good relative mix of say guitars, but you want them all to come up or down a bit--just adjust the buss level.

Try mixing mostly at lower levels, with occasional periods of louder volume and very soft volumes. If you turn things down *almost* all the way, what pops out? Ears will fatigue. Keep checking references from time to time. Run a rough mix out to your car or other listening environment (this is why studios tend(ed) to have multiple sets of monitors.)

Gain structure used to be more critical IMO, it was easy to paint yourself into a corner with instruments and busses too hot and no recourse but to turn down the master...bad! But noise used to be the main enemy. With digital in a DAW things seem more forgiving. But you still want to try to keep gains in their place. If everything is pegged up but you still need more, we can't just go to 11! :) That would tell you that something is out of whack, used to call it "fader creep".

One thing to try when you pull up tracks is to do a rough 5-10 minute mix just to see where the instruments come in at. You may want to adjust the gain at this point to level them out a bit (this would be pre-compression, pre-fader). On Logic you can do this to each region or apply a gain adjustment on the channel strip, same result. That way you don't get into your real mix and get surprised.

Mixing is great fun, it's like juggling dozens of balls (depending on track number) at once and sometimes it doesn't easily happen. But when it does come together it's a great feeling. I've also had it happen where I had a great mix but kept monkeying with it, and it fell apart on me :p That can be easily prevented these days with Daws, you can save versions and easily get back to your exact place. I learned on analog tape with no automation, my automation was pieces of tape stuck on the console by the faders with little notes on them :) You had one shot at a mix!
 

splatt

david torn / splattercell
Platinum Supporting Member
Messages
26,676
And this is related to why I say compression should be seen and not heard. I know people love to say "use your ears" but in my experience compression is the one treatment that should not be heard.
1) i like to hear them when it's an effect i want, and the more important side-note is that
2) a lot of folks, me included, like the feel/give that certain hardware compressors give --- even enjoying the different, often (but not always) subtle shades of distortion/clipping passing a signal through them can offer --- all in order to get a track to speak, or to blend, etc. indeed, there are some plug-ins that serve these functions well. eg: i might set an API-2500 for less than -1dB comp in old mode on the Mix-buss, just to thicken/juicify/glue-things-up; the fairchild comp is incredible for this when tracking, ime: unreal!)

my point is that some great compressors can offer clearly hearable character to your mix, even when the compressor circuit, itself, is not even engaged.

What I mean is I used to make it so that it was enough when I could just barely start to hear it working. But then I walk away for a week and come back and all of a sudden I could hear it too much and I start thinking "what did I do that for? It's way too much!" Then I look at the meters on the comp and the lights were telling me I was close to squashing it when all I really wanted to do was prevent instantaneous spikes.

Hard to explain but I ended up concluding that with compression I was actually better off trusting the meters, because if I could hear it working it was almost always too much compressor.
totally understandable, but..... the more you mix, the quicker you will become at getting what really works for you..... especially if you learn to mix on hard deadlines.
 
Messages
2,695
..... the more you mix, the quicker you will become at getting what really works for you..... especially if you learn to mix on hard deadlines.
Bingo! One thing that helps here is to make some broad strokes and to move quickly...try not to get tunnelled in on one instrument. Treat it a bit like making a sculpture...you wouldn't spend hours making a perfect nose for your sculpture before having the head reasonably shaped. Moving around keeps your perspective on the big picture.

Oh, another thing that can help is to view editing as a separate step...if decide that you need to make edits while mixing it will interrupt your flow. Think of it as shifting gears.
 

ripgtr

Member
Messages
10,775
I've been recording at home since the mid 80s, on daws since the mid/late 90s.
Mixing is freakin' hard.
I can spend a week on a mix (ask me what I've been doing this week, lol) and still not get it as good as a real mixer. I'm going into the studio Sunday to play guitar for a guy, I've worked with this engineer. He sits at that console every day, all day, and records and mixes. He can knock out a better mix in a couple hours than I can in a couple weeks. But I'm a guitar player by trade I haven't put in the time mixing as I did playing.

All that being said, my mixes are WAY better than they used to be. A lot of reading, experimenting, working. Like most stuff, you get better the more you do it. You have to put the work in.
 

Billinder33

Member
Messages
2,178
I like to approach it from the perspective of telling a story to the listener…I keep in mind what specifically I want them to pay attention to at any one point in time. Most listeners won’t hear more than a couple things at a time.
This is a great point.

In that vein, I try to think of my mixes in terms of 'bottomline' and 'topline'. Bottomline being bass, drums, and non-focus rhythm tones, which rarely change tonality (except maybe in EDM or other compositional type music). Topline being several things that will constantly move in and out of focus during the course of the song like vocals, guitars, solos, intros/breaks/outros, etc.

I try to make my topline as consistent in energy across the song as my bottomline, and then make sure those two elements play nice together. Focusing on a consistent topline all the way through the song helps me stay away from getting into that rut of randomly bouncing deep attentions from one mix element to another like in your excellent sculpture example.
 
Last edited:

AndyZ

Member
Messages
1,049
totally understandable, but..... the more you mix, the quicker you will become at getting what really works for you..... especially if you learn to mix on hard deadlines.
David, I'm going to expand on this an inch further because it is so true!
Even more so when you have to learn what works for the client under hard deadlines! Plenty of times those decisions are not what I would ultimately do for myself.

So related to your bottom line that if you can't deliver a track from concept to final mix in 24 hrs, they'll just move on. I've done hundreds of tracks for TV/Commercials and you have to learn what they mean and how that translates to what you do to deliver it in a mix usually on the first pass or maybe 1 revision.
 

Stokely

Member
Messages
1,683
Interesting, I'll try that topline way of thinking :) I sort of do that already with mixing like groups together (keys + guitar, add vocals.....then guitar + vocal etc etc).

I rarely EQ anything soloed, unless I'm looking for something that's bothering me that's very specific. Even then, what really counts is how it sounds with other instruments.


I only worked in a small jingle studio (or so they were called then at least!) for a couple years, but I grew to dread the whole "can you solo that up?" coming from say the guitar player, talking about the guitar. Invariably you'd hear "it's too thin, can you fatten it up?" Sure buddy, if we were doing a solo electric guitar song I'd fatten it right up. As it is, instruments have to make room for other instruments, with vocals taking priority spot 1.
 

VESmedic

Silver Supporting Member
Messages
260
I just wanna say, id stick to a super slow attack on the buss comp. the slowest possible specifically on an SSL style. You want those transients to come through clearly and not be smothered, in a lot of styles of music anyways: especially all forms of rock. Dance/EDM you may wanna smear the transients alittle, but not so much so with rock/metal/pop style music. Slowest attack, fastest release: this is the way :) this will provide the most natural/less pumping style compression, and make it appear that it’s not even there, this is often what people want. However As always, trust your ears.
 

AndyZ

Member
Messages
1,049
Another option is use a comp plugin that has a mix level. I've got instances where I'm slamming the comp but it's really only only passing maybe 10-20% of that back in thru the signal. Both of the UAD and SSL Native buss comp plugins I use a lot have that option on them. Some of the Slate comps have this in their VMR. Even on the 2Bus comp I may not have the mix level up 100% at times. Just listen for what you want to hear...
 
Messages
2,695
Another option is use a comp plugin that has a mix level. I've got instances where I'm slamming the comp but it's really only only passing maybe 10-20% of that back in thru the signal. Both of the UAD and SSL Native buss comp plugins I use a lot have that option on them. Some of the Slate comps have this in their VMR. Even on the 2Bus comp I may not have the mix level up 100% at times. Just listen for what you want to hear...
Especially if you're using a real 'character' type of compressor like a vari-mu style. NI's supercharger has this in spades; it can really thump, but the mix knob comes in real handy to just tuck that under the signal...a little bit like having a different ratio. I never really thought too much about it before but I suppose what this does is let you work the compressor at a cool spot in its knee, but then effectively tailor the ratio of the blend.
 




Trending Topics

Top Bottom