Mix for Master

[starbuck26]

Hey gents,

I'm mixing my band's first full EP. It's baking my noodle. This is the first time I'm mixing anything to be mastered by another studio, and really the first time I'm really paying attention to my fader levels other than just adjusting them to what sounds right.

I tracked everything between -12 and -8 rms with peaks no higher than -6. The folks at the mastering studio tell me a mix between -10 and -8db is optimal to leave them with enough headroom. So I thought I was golden.

Now, I'm having problems. I'm adding overdubs all over the place, compressing the drums, and eqing everything. Problem is, when I'm eqing I wind up adding between 1 and 3 db per channel... then all of them combined, and im getting peaks around 0 to +3db on the master fader.

Basically, is there anything wrong with subtracting gain at the master fader to bring my levels down to where I want them overall, or should I go back and redo the eqs and mix the whole thing at a lower level?

Does this make any sense?

Please help me out before I get a nosebleed.

T

[jaybriggs84]

You don't want to just pull down the master fader cause there can still be problems before it gets to it

[Daveg62]

I typically cut more than boost while eq'ing. Might be something to be aware of in the future. I know it doesn't help your situation now. I think if it were me I would just take all the track levels all the way down and start from scratch. Bring the drums up first and then bass, vocals, and everything else. No need to re-adjust the eq if it's where you like it. If it's still too loud start over again till you end up where you want to be. Everyone works differently. I mix on an analog console so this technique works for me.

[starbuck26]

I mixed two tracks out of an eventual nine, so I'm going to start over and try to cut rather than boost.

I know I'm not supposed to add gain with the fader, but is there anything wrong with cutting gain at the fader? Boosting 3db in eq and then cutting to compensate at the fader, I mean?

[jndietz]

Why aren't you allowed to add gain with the fader?

[starbuck26]

Originally Posted by jndietz

Why aren't you allowed to add gain with the fader?

Well, I don't remember where I heard it, but I thought it better to record at a good target level by adding gain with the preamps, keeping the faders zeroed. Maybe that's all off the mark. I never really learned how to mix from a book or from someone who actually knew how to do it, just picked up pieces of advice from people all over the place.

[Massive Master]

*Generally* you're going to be okay reducing the master fader.

Personally, I like going at it at the track level - But that can affect dynamics and post-fader effects sends, etc.

(And not to be picky, but you're talking about dBFS - Not RMS - BIG difference).

[starbuck26]

Originally Posted by Massive Master

*Generally* you're going to be okay reducing the master fader.

Personally, I like going at it at the track level - But that can affect dynamics and post-fader effects sends, etc.

(And not to be picky, but you're talking about dBFS - Not RMS - BIG difference).

Right... so, basically, here's my approach:

I'm going to systematically redo each track, trying to cut rather than add track by track with my eqs... which *should* keep my tracks from jumping in volume before the master fader. But, in the end, I will correct my values with the master fader by cutting volume to keep my peaks at -6db and my levels at -8 to -12dbs RMS.

Right?

It still amazes me that these litte numbers next to all these faders actually mean something...

Sort of the antithesis of rock n' roll, no?

[Elton Bear]

Originally Posted by starbuck26

Sort of the antithesis of rock n' roll, no?

Completely off-topic, but a great engineer I know told me about this local punk who records his band live and mixes next to an egg-timer. He sets it to half a n hour or something, and when it goes off he leaves his mix as it is at that point to stop the mix getting stagnant or over-complicated. Now THAT us rock 'n' roll

[starbuck26]

Originally Posted by Elton Bear

Completely off-topic, but a great engineer I know told me about this local punk who records his band live and mixes next to an egg-timer. He sets it to half a n hour or something, and when it goes off he leaves his mix as it is at that point to stop the mix getting stagnant or over-complicated. Now THAT us rock 'n' roll

Im not sure I get it. WTF is an egg-timer? I searched the cubase help file, and then on zzounds.com. Nothing.

[JuliánFernández]

Originally Posted by starbuck26

But, in the end, I will correct my values with the master fader by cutting volume to keep my peaks at -6db and my levels at -8 to -12dbs RMS.

-12dBFS, not RMS...

[starbuck26]

Originally Posted by JuliánFernández

-12dBFS, not RMS...

-12 DBFS, indeed. I just looked it up.


sigh.

[Massive Master]

Originally Posted by starbuck26

Right... so, basically, here's my approach:

I'm going to systematically redo each track, trying to cut rather than add track by track with my eqs... which *should* keep my tracks from jumping in volume before the master fader. But, in the end, I will correct my values with the master fader by cutting volume to keep my peaks at -6db and my levels at -8 to -12dbs RMS.

Right?

It still amazes me that these litte numbers next to all these faders actually mean something...

Sort of the antithesis of rock n' roll, no?

Somebody already got the dBFS thing...

Anyway - If everything is okay but you're clipping everywhere, just grab *all* the faders and pull them all down a few dB. Again, that might affect dynamics processing and post-fade sends somewhat - But IMO, it's usually for the better that way anyway... You can always add a little more verb if it dropped too much or what not.

[jndietz]

To be honest I never knew you were "supposed" to leave the faders on the mixer at zero and only change the preamp level.

I wonder what difference it makes?

[starbuck26]

Originally Posted by Massive Master

Anyway - If everything is okay but you're clipping everywhere

My interpretation of this... Everything sounds great, pristine, no audible clipping or distortion, but I'm pushing faders into the red. Right?

[starbuck26]

Originally Posted by jndietz

To be honest I never knew you were "supposed" to leave the faders on the mixer at zero and only change the preamp level.

I wonder what difference it makes?

I don't really know anything... but logically it seems to me that you want to get a proper level coming in when tracking, and leave yourself a bit of headroom in the mix. If you have a low signal coming in and jack the fader to its max... you'll only be able to lower the volume from thereon out.

But perhaps someone who knows what they're talking about can weigh in on that one. I've been trying to figure out where I heard that... still no luck.

[masteringhouse]

Originally Posted by starbuck26

Basically, is there anything wrong with subtracting gain at the master fader to bring my levels down to where I want them overall, or should I go back and redo the eqs and mix the whole thing at a lower level?

Does this make any sense?

Please help me out before I get a nosebleed.

T

It will depend on the bus structure of your DAW. For example in Pro tools TDM the bus operates at double the word length of the individual tracks to allow for additional headroom when summing tracks, so lowering levels here is absolutely fine. If OTOH the bus operates at the same bit depth as the tracks it would technically be better to reduce the level of the tracks.

The correct answer is to find out the details of how your DAW operates and work accordingly.

Doesn't help with the nosebleed though ...

[SouthSIDE Glen]

Originally Posted by starbuck26

I don't really know anything... but logically it seems to me that you want to get a proper level coming in when tracking, and leave yourself a bit of headroom in the mix.

From the way I look at it, this gets to the heart of it.

I look at mixing in the DAW as just another stage in the over gain structure chain; assume that you want the signal going out to be at roughly the same level as you have going in, more or less. Pretend that instead of going to digital master after mixdown, that it'll be heading back out to analog, as if you were recording to tape or sending it to an analog mastering chain. Considering most modern converters equate a 0VU analog line level to be equal to somewhere in the -22 to -16dBRMS range, this is a pretty good target range to look for coming out of the mix.

I've found that if you have good track levels coming into the mix - having a clean analog level going into the converter and just letting the converter do it's thing without boosting or cutting the digital track faders when recording - and by keeping those levels constant throughout the plug-in chain like you would at any analog gain stage, that the mix usually pretty much works itself out level-wise just fine.

Of course much depends on the music type and arrangement and all that, but I find for most rock/blues/pop/roots-related genres, with the master faders in my DAW set to unity gain, that the mixes seem to almost automatically come out right in the previously-mentioned RMS range (around -20 to -17 dBRMS) as if by magic.

But it's not magic; it's actually that the engineers that have designed the gear (hard and soft) that we work with to actually work that way. Somebody actually planned for all this. Kinda neat how that works, isn't it?

G.

[DavidK]

Originally Posted by starbuck26

Im not sure I get it. WTF is an egg-timer? I searched the cubase help file, and then on zzounds.com. Nothing.

Its a plug-in made by Betty Crocker.






























Thanks folks, I'll be here all week.

[SouthSIDE Glen]

Originally Posted by DavidK

Thanks folks, I'll be here all week.

No charge to get in, but it'll cost you a $10 exit charge if you leave before the show's over.


Worth every penny.




Sorry, Dave, you left that softball hanging there. I had to swing at it .

G.

[starbuck26]

Originally Posted by masteringhouse

It will depend on the bus structure of your DAW. For example in Pro tools TDM the bus operates at double the word length of the individual tracks to allow for additional headroom when summing tracks, so lowering levels here is absolutely fine. If OTOH the bus operates at the same bit depth as the tracks it would technically be better to reduce the level of the tracks.

The correct answer is to find out the details of how your DAW operates and work accordingly.

Doesn't help with the nosebleed though ...

I've been sifting through your reply and Glen's reply last night and this morning. Things are getting more complicated yet again.

So my DAW, a Zoom HD-16, records 44.1 khz at 16bit. I use that to track, and then move everything to Cubase LE, where I mix.

I assume that since Cubase handles a much higher bit depth that I am ok lowering fader level anywhere I please, either track by track or in the master fader.

Or do I not understand anything you're telling me?

[masteringhouse]

Originally Posted by starbuck26

I've been sifting through your reply and Glen's reply last night and this morning. Things are getting more complicated yet again.

So my DAW, a Zoom HD-16, records 44.1 khz at 16bit. I use that to track, and then move everything to Cubase LE, where I mix.

I assume that since Cubase handles a much higher bit depth that I am ok lowering fader level anywhere I please, either track by track or in the master fader.

Or do I not understand anything you're telling me?

Well, yes and no.

Having a larger bit depth in Cubase is an advantage, but the issue that you mentioned in the inital post more has to do with summing tracks.

Let's use an analogy to demonstrate. In place of audio let's use water and in place of bit depth let's use the concept of a bucket. Furthermore a bucket can be made of metal or rubber. When you reach the volume of a metal bucket, it will start to overflow, however a rubber bucket (like a ballon) will automatically resize itself when it is being filled up to a certain limit.

A fixed bit system is like the metal bucket, once you've used all of the bits it overflows and distorts. A rubber bucket (floating point system) will expand and contract based on how much it contains.

Ok, now with that in mind, imagine each track is a bucket of water, and all of this water gets poured into a drain (bus) and into another bucket that contains all of the water (master bucket). In some fixed bit systems (like Pro Tools TDM) the master bucket is made large enough to contain all of the water from all of the tracks even if they were filed to the brim. However it must pour this out to a smaller bucket (final bit depth) and as a result you may need to reduce the amount of water by taking some out intentionally. HOWEVER, no water was accidentally lost by overflowing, so life is good, the system knows exactly how much water was removed. Distortion will happen when the system thinks it knows how much water was removed but actually some of it was spilled all over the floor, or you didn't shake the bucket (dither) after removing the water (this is a whole other topic).

In the case of floating point the rubber buckets or balloons will stretch to accommodate the extra water to a certain point. In these systems the master bucket is usually the same size as the individual buckets, but due to stretching will not allow the water to spill if it is not near enough to popping the balloon.

The objective is to know how much water your master bucket can hold before overflowing or popping (clipping).

Ok, now I have a headache.

Just do what Glen said ...

[starbuck26]

Originally Posted by masteringhouse

Well, yes and no.

Having a larger bit depth in Cubase is an advantage, but the issue that you mentioned in the inital post more has to do with summing tracks.

Let's use an analogy to demonstrate. In place of audio let's use water and in place of bit depth let's use the concept of a bucket. Furthermore a bucket can be made of metal or rubber. When you reach the volume of a metal bucket, it will start to overflow, however a rubber bucket (like a ballon) will automatically resize itself when it is being filled up to a certain limit.

A fixed bit system is like the metal bucket, once you've used all of the bits it overflows and distorts. A rubber bucket (floating point system) will expand and contract based on how much it contains.

Ok, now with that in mind, imagine each track is a bucket of water, and all of this water gets poured into a drain (bus) and into another bucket that contains all of the water (master bucket). In some fixed bit systems (like Pro Tools TDM) the master bucket is made large enough to contain all of the water from all of the tracks even if they were filed to the brim. However it must pour this out to a smaller bucket (final bit depth) and as a result you may need to reduce the amount of water by taking some out intentionally. HOWEVER, no water was accidentally lost by overflowing, so life is good, the system knows exactly how much water was removed. Distortion will happen when the system thinks it knows how much water was removed but actually some of it was spilled all over the floor, or you didn't shake the bucket (dither) after removing the water (this is a whole other topic).

In the case of floating point the rubber buckets or balloons will stretch to accommodate the extra water to a certain point. In these systems the master bucket is usually the same size as the individual buckets, but due to stretching will not allow the water to spill if it is not near enough to popping the balloon.

The objective is to know how much water your master bucket can hold before overflowing or popping (clipping).

Ok, now I have a headache.

Just do what Glen said ...

That actually cleared a lot up. I can't thank you enough for that analogy.

Now my question is... how do I found out more about my cubase buckets?

As for doing what Glen said... I know enough from all of his advice on this site that I should always just do what he says... but, unfortunately, in this specific case... I have no fucking clue what he's talking about in this specific case, but I do recognize that there is yet another gem of knowledge awaiting me.

The word dither scares me by the way. I've been reading a textbook on mixing audio... I actually felt portions of my brain shutting down while reading about bit depth conversions.

I don't want to overcomplicate things here, though. If either you or Glen can confirm that I can safely lower my track faders in cubase LE I will get back to mixing my Stevie Nicks cover in relative peace and stupidity...

[SouthSIDE Glen]

Originally Posted by starbuck26

As for doing what Glen said... I know enough from all of his advice on this site that I should always just do what he says...

While I appreciate the vote of confidence....GOOD LORD NO! If you had seen my completely wrong initial answer I gave a couple of weeks ago about the operation of a mixer of the type I was sitting right in front of at the time, you'd know I make some boneheaded mistakes just like the next guy. Of course, I was sick as a dog at the time I gave that answer, but still...a boneheaded answer is a boneheaded answer.

Originally Posted by starbuck26

If either you or Glen can confirm that I can safely lower my track faders in cubase LE I will get back to mixing my Stevie Nicks cover in relative peace and stupidity...

Yes you can. Think of it this way; if you couldn't mess with your track levels, you couldn't use volume level automation - level automation is basically dynamically changing track volume levels on the fly. So sure, adjust the track levels. In fact, I recommend setting the levels so that you're getting the "right" levels out of your master bus with the master faders set to unity.

I'd try to explain more about what I was saying about gain structure, but I don't want to bruise your brain any further here. Let me just leave you with a bookmark: head over to IRN at www.independentrecording.net an click on the "Metering and Gain Structure" icon on the lower right to pull up an on-line interactive applet that's chock full of basic generic info on what the different meters along the line mean, how VU, DBFS, and RMS and other measurements actually mean and how relate to each other on a real and useful level, and some basic methods for setting levels all along your recording and mixing chain from mic to plug-in to get the most out of your gear and your signal. Play around with that applet at your leisure, take small bites and swallow between them, and you might just come out the end with no head injuries and an understanding of the common thread of levels around which all these widgets and doo-dads have actually been designed to work.

But for now, let me say this about track levels. When I sit down to start a mix, the initial set-up stages are usually a pretty quick 3-step process for me. This process is not the only way to skin the cat; others may take slightly different approaches, and that's fine. This is just what works for me. Maybe it'll help you out just a bit for starters, without having to go into a bunch of gain structure theory:

1.) I give the mix a faders-up listen. This means setting all track faders and master faders to unity gain (the 0 mark about 2/3 - 3/4 of the way up the fader. This tells me two things: first, how I want to initially adjust the track levels in step 2, and second, how good the tracking was and how much work I have ahead of me (the better a raw faders-up sounds, the better the tracking and the less work I have ahead in mixing.) I may or may not set initial panning at this point; usually I get the initial/basic panning scheme out of the way right then and there based upon what I hear.

2.) Based upon what I hear in the faders-up mix, I set initial volume levels for each track relative to one another, just getting them each roughly in the ball park for what I imagine in my head that I want the final mix to sound like. IMPORTANT here is that I never raise any of the track faders. I just reduce each one by the amount necessary to get the relative volume roughly in line with where it belongs in the mix with the other tracks.

The faders virtually *never* (well, maybe one in 20,000 times only) go above unity gain. At least one track - sometimes a couple - usually stays there at unity; especially if they are quieter-recorded tracks that need to be loud in the mix. Those are the ones with the highest gain setting, at unity. Others might be 2-3dB down, still others may be dropped more than that; it all depends upon how the tracks were recorded and how the mix plays out.

3.) Once I have those rough starting mix levels set up, I switch automation on for all tracks, setting the initial flat line level for each track somewhere around 3 or 4dB down from the top (maximum) level. This gets me ready to individually start actual track mixing for the final mix, using automation heavily (see the current thread in this forum on texture and automation) to do my mixing. The extra 3 or 4dB of headroom I give myself allows me to move the fader/volume level for that track up by up to that much at any point, should the overall track level be a bit low and I need some extra oomph here or there.

Then I start the track detailing in earnest. The EQing, compression, automations, etc.; usually submixing the rhythm section (bass & percussion) first, then the accompaniments, the finally the leads and main vocals, in that order (usually, but there are always exceptions.) If I need to make fine tuning adjustments to the individual fader and automation levels at this point, I do. But I almost never send a fader or automation level above unity gain. The only real exception to that is if the tracking is soooo poor and uneven, and one track is so anemic and low-level that it needs the above-unity boost just to be heard. But even in that case, I'll try to get a re-track first if possible.

Usually for me, assuming that I have fairly decently gain staged recordings with behaved recording levels, the final mix winds up coming right out in the sweet spot fairly naturally, without having to touch the master faders, just leaving them at unity. There are exceptions, sometimes, but it far more often than not just works out.

HTH,

G.

[starbuck26]

We all make mistakes. Big fucking deal. The most recent post of yours I had in my memory, Glen, was the one you wrote on the digital clipping on the RHCP album. No one really responded to it, likely because they were dumbfounded.

We are indeed NOT making pancakes.

The explanation of your gain staging basically put it all together. I had already been looking at independentrecording.net, and been studying the metering and gain staging charts. More useful than that, though, was the interactive eq chart. With my lack of experience it was a big help to have some sort of template to use when getting started.

So cheers, man. I meant what I said.

Thanks to everyone who gave me some insight into all this science...
a few people actually messaged me and said that this was a good question. So hopefully it's pointed a few of us in the right direction, and not just me.

And it's just in the "nick" of time. Stevie ain't getting any younger, and her nose is about to fall off... we better finish this album quick.

T