Skip to main content

Can someone offer a strategy for mixing-down a multi-track project, wherein the lead voice tends to get overpowered by the increasing mass of audio signal as a result of accompanying instruments being gradually added?

My approach has been either to incrementally increase the gain on the lead track each time an accompanying instrument is added (until reaching zero dB), or to keep attenuating the combined accompaniment. I'm thinking that adding compression to the lead will degrade the voice by reducing its dynamic range.

Comments

paulears Sun, 08/16/2020 - 12:26

Compression keeps it loud, that's all. If the vocal gets overshadowed, then it's probably common to have some mild compression on it then pan the other tracks out of the way to create a little 'hole' and then balance and eq them to also prevent wiping the vocal out. It means sometimes chopping notches in the eq to make space. if everything has it's energy in the same band as the voice - you're bit stuck.

bouldersound Sun, 08/16/2020 - 13:01

First of all, I'd start the mix with a ton of headroom. I would probably incrementally reduce the instrumental tracks as I added more, and/or assign them to a submix bus with a compressor. Even then it might be worth raising the vocal track as instruments are added, then a bit of compression over the whole mix. Spread the level control measures across different inputs and multiple stages of compression to keep it from being too obvious.

Also, prioritize instruments. Allow some to take more of a back seat and give just one or two a feature role in the mix.

pcrecord Mon, 08/17/2020 - 06:29

Yes headroom is a must. Also using automation to keep the vocal from drowning deep.
You target shouldn't be zero db.. An healthy mix at its loudest will peak at around -6db or lower.
Don't stress about the loudness that much because the final touch will be done by the mastering engineer anyway
Also, don't put a compressor on the masterbuss that compress the mix even on quieter parts of the song. The problem with that is when adding more and more instrument the overall compression will be too much and kill the dynamics.
Why don't you post it here? we may have a better understanding of the problems by listening to it..

Boswell Mon, 08/17/2020 - 07:21

When you have a building instrumental sound though a song, I would try doing a purely instrumental mix without any vocal. Get it balanced (using compression as necessary) as you would like to hear it as an instrumental track, with a suitable build and climax. Then see what lessons you can learn from doing that to apply to re-mixing it as a backing track for the vocal. You need to aim for a climax that is only partly achieved through increasing mix volume.

Did you record the vocal and the instruments one-by-one, or possibly at the same time but in separate acoustic spaces? It's much more difficult to use compression in a mix when you have bleed from one instrument or vocal into the microphones of the other sources, as in a live performance, for example.

bouldersound Mon, 08/17/2020 - 07:31

pcrecord, post: 465246, member: 46460 wrote: Also, don't put a compressor on the masterbuss that compress the mix even on quieter parts of the song.

Right, I'd say it shouldn't start working until about two thirds of the build has happened. The ratio should be fairly modest. I suspect a medium to slow attack might work. I would set release to make things sound as natural as possible, but I would have to do it by ear.

Normally I resist using a compressor in the master bus, but this is kind of a special case.

pcrecord Mon, 08/17/2020 - 07:57

bouldersound, post: 465250, member: 38959 wrote: Right, I'd say it shouldn't start working until about two thirds of the build has happened. The ratio should be fairly modest. I suspect a medium to slow attack might work. I would set release to make things sound as natural as possible, but I would have to do it by ear.

Normally I resist using a compressor in the master bus, but this is kind of a special case.

I'd try to automate poste of the volumes first (on individual tracks) then if some things are out of place, maybe compression (on tracks again)
Once there is some track level control, then a buss compressor could glue things a bit further..

Is this one of those 128 tracks project ? ;)

KurtFoster Mon, 08/17/2020 - 08:28

how to mix rock and roll quickly: start with the kick snare and bass hitting the meter at about -18 or -3dB vU. try to keep them there through out the song. if you have to make moves don't worry, that's how it's done. now bring up the instruments to taste. you might need to make some moves on the instruments too, still ok. you can sub group drums /instruments at the expense of added processing / noise to make things easier to manage. last mix in the vocals. if the lead vocal gets buried at some points you can add a bit of compression to get it louder or ride the vocal fader.

kmetal Mon, 08/17/2020 - 12:07

Good advice so far. I think this is as likely an arrangement issue as a mix issue. There could be too much energy being added in the vocal range. Good arrangements leave space for this with instrument selection, and by keeping things in octaves other than vocal range. Also busy rythymic parts can interfere.

A mix trick you can do is when new elements enter, have then enter a bit loud, then fade them down some. You will hear this on Rick Rubin productions often.

Djard Mon, 08/17/2020 - 20:58

I find all of the responses here useful. Some of the suggestions I knew but forgot as most of my work is composing and arranging. My projects are typically 24 to 32 tracks, a combination of jazz-blues and lots of orchestral tracks. Only the guitars, keyboards and vocals are recorded live. Oh, and the didgeridoo! The rest of the tracks are created in Finale, using Garritan instruments.

I like the idea of creating an accompaniment mix then add the vocals. Just when I thought I knew a little, I learn here thst I hardly know anything. Sigh!

bouldersound Tue, 08/18/2020 - 23:30

Djard, post: 465266, member: 52026 wrote: That approach sounds like it requires a lot of skill acquired through experience. No doubt that mixing is an art form.

Having a structured process makes it easier. Getting kick/snare/bass/vocal core working together gives me a defined goal. Adding the rest of the tracks shouldn't obscure the core tracks. Once the added tracks start coming in, it might take a little fine tuning of eq and levels to keep the core from getting lost

audiokid Wed, 08/19/2020 - 11:27

All excellent suggestions. (y)

To further add something I do.... mix in mono and not too loud. Once you have it ... then switch to stereo and pan to taste. I rely on simple Auratone speakers in a treated room. Takes a bit to get used to those but they work great for me.
Don’t mix over too long durations as ears get fatigued.

Davedog Wed, 08/19/2020 - 12:27

Think about it like you are hearing a song from the next room. What do you hear clearly on every one of them?

Like Kurt, it starts with the percussion and the bass. I add vocals at this stage and the rest becomes icing on the cake. Once you establish the drums/bass/vocal levels and pan, don't touch them. Ever.

Bring everything to the level which still allows you to clearly hear the drums/bass/vocal ensemble. Pan for position of the additional tracks. Eq gets you front and back position. Some things may need dynamics help. Multiband compression on the more complicated instruments.

All instrumentation should also be bussed to alike subs. Guitars get a buss....percussive keys get a buss.....string arrangement gets a buss.... horns get a buss...etc etc. In this way you get a balance inside of their respective submixes and you have the ultimate control of what is hitting the 2 buss and from where.

Djard Wed, 08/19/2020 - 20:39

I do bus the groups but always save a raw wav file of each, just in case.... as wisely mentioned, failure to take a break can result in much waste of time. Yes, hearing the "last" final mix at low volume helps.

What do you more seasoned engineers do when you can't decide between three diffenent mixes, each with its own strength? Usually one ends up appealing more to me, but what if the luxury of time is not available: flip a coin?

kmetal Wed, 08/19/2020 - 21:13

Djard, post: 465270, member: 52026 wrote: What do you more seasoned engineers do when you can't decide between three diffenent mixes, each with its own strength? Usually one ends up appealing more to me, but what if the luxury of time is not available: flip a coin?

I take it out of the studio and listen on other systems i know well. If there is parts of different mixes i like i take notes and adjust ONLY those things and make a mix of that. If i make more changes than those i do a seperate one.

If there is no time i use the mix that has least flaws, and best highlights the key parts. ie mix 1 has better snare, mix 2 has vocals sitting better. Vocals would win. When in doubt just follow your gut, one will always feel right or one will have less stuff that's irksome. If the client is involed give them both and they'll pick.

Davedog Wed, 08/19/2020 - 22:41

Djard, post: 465270, member: 52026 wrote: I do bus the groups but always save a raw wav file of each, just in case.... as wisely mentioned, failure to take a break can result in much waste of time. Yes, hearing the "last" final mix at low volume helps.

What do you more seasoned engineers do when you can't decide between three diffenent mixes, each with its own strength? Usually one ends up appealing more to me, but what if the luxury of time is not available: flip a coin?

I have a three level process. When the tracks are raw, I make a balance mix that only has pan and volumes. Then a save and copy this mix to another session. I will usually name this session as the "edit" session. In this one I do all repairs, tuning, timing, arrangement changes etc. I never lose a track even if I don't use it or I chop it up till it's something else. Anything that gets chopped gets saved intact first and then muted and hidden. I may wind up with 25-45 or so working tracks but there will be a few more not being used and not on the page. When I'm done with the "edits" session I make another balance mix using recalls from the original sometimes or starting from scratch since during all the edits I've gotten to know the song quite well. At this time I will add a master and an aux master and print a copy to listen to. After a bit I'll save as and make a new session from this and build the final mix around these tracks. This usually is where I build all the stems/aux/submixes and assign all the busses. Then I print from this and live with it as long as possible before the clock starts ticking. By now I'm ready for the final mix....any small repairs may be done now...a note out of time...a bad guitar flub....mistake on the bass...etc.

When I mix I mix to a print track inside the daw. All my automation will be in place except I still like to ride a few faders due to my advanced age.......but these will be sections of a solo...an important emotional swell......all the volume manipulation will be done with the automation. I have Avid Artist controller and Artist mix so sometimes I'll automate live passes with the sliders and commit to memory...other times I'll use the pencil.