Most transparent/musical way to raise volume?

[Nola]

I'm sorry if this is beaten to death, but say my file is at like -25db (peaks at -6db) after mixing. say I want to get it to -17db or so.

What is the most transparent/musical way to do this?

To my ear, limiting makes things harsh when trying to gain that much volume. Should I push the master fader up (I can get 6db this way), export, then limit a little bit to get added volume? Limiting a little doesn't seem to do much damage, but if I try to get 6 to 10db in volume from the limiter it gets harsh.

I am kinda ignorant about volume b/c I like quiet mixes and feel they're more pleasing sounding (less harsh) and would rather maintain that and turn up the headphone preamp than crush the song, but for the mix I'm working on right now it does need volume.

Thanks!

 

[Steenamaroo]

Hi,

If your peaks are already at or near zero then yeah..limiting for volume is probably going to be noticeable.
Compressing or volume automating per track can make a big difference.

A mastering engineers hands may be tied but yours are not.


Just for the learning experience do a bounce of your song and import it as a stereo wav.
Line it up with your session and mute it.

Now you can see where the big peaks are, while you're working at the mix.
Find out which tracks are causing stray peaks or buildups and consider individual limiting/compression/volume automation or even rearranging, to compensate.

Just don't go mental. Ups and downs and dynamic range are good. There's always a middle ground.

 

[miroslav]

Sometimes transparent and musical are two different things....but anyway...

I would reduce the -6dB peaks first, and I generally like to do that manually, and not by slapping any limiter/compressor on the track.
I will highlight the start/end of a peak, an just lower its volume to a more appropriate level.

I focus first on the most obvious/offending peaks...and when they are more in line with the rest of the lower level peaks across the whole track, I can then raise the overall track level...and if at that point I still need something more, some lightly applied limiting can be done.

I think if you start off aggressively trying to do that with a -25dB track with -6dB peaks...you kinda run out of room much quicker, and you have to hit it that much harder...and there's where the nasty comes in.

 

[Nola]

Hi,

If your peaks are already at or near zero then yeah..limiting for volume is probably going to be noticeable.
Compressing or volume automating per track can make a big difference.

A mastering engineers hands may be tied but yours are not.


Just for the learning experience do a bounce of your song and import it as a stereo wav.
Line it up with your session and mute it.

Now you can see where the big peaks are, while you're working at the mix.
Find out which tracks are causing stray peaks or buildups and consider individual limiting/compression/volume automation or even rearranging, to compensate.

Just don't go mental. Ups and downs and dynamic range are good. There's always a middle ground.

Thanks, Steen. The peaks are at -6db not 0.

The biggest spikes seem to be in the drums and vocal. So compress them more, and then what, I'd be able to limit the entire track more without harshness?

 

[miroslav]

Just for the learning experience do a bounce of your song and import it as a stereo wav.
Line it up with your session and mute it.

Now you can see where the big peaks are, while you're working at the mix.

Yeah...I do that quite a bit...especially when I'm close to the mixdown. I can identify the root cause of a stereo mix peak, and trace it back to maybe a guitar track or a drum hit...etc...and then I lower it on the individual track, rather than doing it across the whole final mix.

 

[Nola]

I would reduce the -6dB peaks first, and I generally like to do that manually

I could probably do that on the vocal since it's only certain words. On drums it would take forever since there are so many hits (pretty much every 2 and 4 beat has a higher spike). Should I just compress the drum more heavily?

 

[Tadpui]

I've been having luck doing like Steen mentioned and doing a fair amount of bus compression in the mix, before trying anything on the master bus. For me, the kick and snare have always dominated the push on a master bus limiter. For the first time, I tried some compression on the drum bus, just a few dB of GR, and it made a noticeable difference in how the master bus limiter reacts.

Of course, now my kick and snare have gone missing because I'd been overcompensating in other ways...but I'm optimistic that this'll end up with good results once I've gotten a little more practice doing this.

*edit: although now that I re-read Steen's post, that's not what he was saying at all

 

[Steenamaroo]

I could probably do that on the vocal since it's certain words. On drums it would take forever since there are so many hits (pretty much every 2 and 4 beat has a higher spike). Should I just compress the drum more heavily?

Yeah, with vocal there shouldn't really be massive stray peaks.
I mean, I'm speaking very broadly here but personally, I don't mind squashing vocal pretty hard because most of the time I want it to be pretty static and consistent.
Hard Ts and percussive sounds jumping out isn't nice so I tame all of that stuff and yeah, I do generally compress quite hard.

Drums by their nature are going to be full of peaks. It's what they are...but you can bring the whole picture in a bit tighter a few ways.

With compression (for your reasons) I suppose it's kinda tricky because slow attack compression defeated the purpose and fast attack compression kills your transients.
Maybe consider a parallel compression bus on your kit?

That means you'd be adding a super-squashed copy of the drums to the real drums.
Usually what happens is the whole things sounds fuller and louder but the peaks don't increase proportionally.

 

[Nola]

Of course, now my kick and snare have gone missing

That's what I'm afraid of if I compress them to the point the spikes are gone...at that point you'd just have a boxy thud without any transient, right?

I feel like there's some secret to this that I'm not understanding...or maybe this is just really hard to get right.

 

[Steenamaroo]

*edit: although now that I re-read Steen's post, that's not what he was saying at all

I don't usually get what I'm saying....I forgive you.

This is one of the arguments against quantised music although, strangely enough, the activist never seem to mention it.
Quantised beats and instruments are, by their nature, totally accurate so all those peaks coincide perfectly (sample dependant).

If you asked a great drummer to hit kick, close hat, snare and one tom "simultaneously" it'd look like a cascade when zoomed in and, therefore, the total recorded peak would be lower.

Maybe you're not using quantised beats but I just thought I'd mention it.

That's what I'm afraid of if I compress them to the point the spikes are gone...at that point you'd just have a boxy thud without any transient, right?

I feel like there's some secret to this that I'm not understanding...or maybe this is just really hard to get right.

Try that parallel compression bus on the kit.

 

[Nola]

Hard Ts and percussive sounds jumping out isn't nice so I tame all of that stuff and yeah, I do generally compress quite hard.

How much GR do you do on vocals? I was at like -3db and that felt good. If I pushed it to -5db or more, which was needed to tame the peaks, the sibilance and those types of sounds became abrasive. So maybe for me manually ((I did use automation, too) doing it is better. But, I'm just curious how much compression you're using on them.

 

[Steenamaroo]

Yeah, that's a problem with heavy vocal compression.
Depending on your release the breaths and noises can really come to the surface.

If it's a problem I will automate those things but a long release can usually hit most of it.
The only problems left are usually sharp intakes of breath after relatively long breaks.

 

[Nola]

Yeah, that's a problem with heavy vocal compression.
Depending on your release the breaths and noises can really come to the surface.

If it's a problem I will automate those things but a long release can usually hit most of it.
The only problems left are usually sharp intakes of breath after relatively long breaks.

Okay cool. Thanks. I just wanted to make sure it wasn't me/doing something wrong with the compressor. I'm still working my way through compression and only recently even began to hear compression actually working. I couldn't hear it working for the longest time. Now that I hear it, I'm hearing harshness pretty quickly (though, on guitars I notice they become dull and lose their shimmer if I add too much, so I guess it depends). Man, so complex.

So how much gain reduction do you usually shoot for on vocals? Is -3db reasonable or is that light?

 

[Steenamaroo]

Okay cool. Thanks. I just wanted to make sure it wasn't me/doing something wrong with the compressor. I'm still working my way through compression and only recently even began to hear compression actually working. I couldn't hear it working for the longest time. Now that I hear it, I'm hearing harshness pretty quickly (though, on guitars I notice they become dull and lose their shimmer if I add too much, so I guess it depends). Man, so complex.

Take a track with plenty of dynamic range and put a compressor on it with the threshold pulled down pretty far, so you have heavy gain reduction.
Now play with the ratio, the attack and the release.
That should give you a much clearer picture of what every control is doing.
A good way to learn and understand what anything is doing is to observe it at its extremes.

So how much gain reduction do you usually shoot for on vocals? Is -3db reasonable or is that light?

Strokes for folks and every genre/mix has different requirements, but I wouldn't consider 3db reduction to be particularly heavy.
Cliché for sure, but it's all about how it sounds.

If you're hearing harshness keep in mind that the compressor probably isn't making anything sound harsh - it's probably pulling the extremes back and allowing you to hear the quiet parts more clearly.
They will sound harsh relative to the main sung parts under those circumstances.

With a vocal performance the quiet parts are lip smacks, breath intakes, clicks/pops etc so if there's a need to compress hard, you're going to want to keep on top of those noises too.
A gate before the compressor can help a lot with that, or you can manually edit.

 

[miroslav]

I could probably do that on the vocal since it's only certain words. On drums it would take forever since there are so many hits (pretty much every 2 and 4 beat has a higher spike). Should I just compress the drum more heavily?

Yeah...it would be tedious doing it for every single beat....but you don't really need to do that.
I usually look at the two main offenders....the snare and the kick...and I'll visually scan the track, and often there are some hits that peak substantially more than others. I mean...no one hits the snare the same way every time...etc...so I find that if I start by focusing on the odd very high peaks, and lowering them by a few dB...I instantly get that headroom back.
Sure...it's not as easy as slapping a plug on a track...but I've been doing it manually for a long time, and it's not as tedious as it may initially seem.

Heck...if you spend a few hours trying all kinds of plugs and approaches to fix the problem...you could probably run through your tracks in that same amount of time. I rarely have more than 30 tracks on any song...and of those, half are probably ruled out from the git-go as the possible peak culprits...so the rest can be gone through without a lot of work/time.
At first, it will feel tedious even doing a handful...but as you get better sat it, you will be surprised how fast you can highlight a peak, lower its level, and then move to another one. Left hand on the keyboard, right hand on the mouse...and then just assembly line the process.

That said...you have to pick something in your mix as the deciding loudness reference. Everyone has their approach...but I tend to use the drums as my level reference. IOW...I'll set the drum track level, and then I'll bring up the other tracks into balance WITHOUT touching the drum tracks. Once I get the balance...I then see how much more headroom I have left to play with, and then I can raise ALL the tracks equally/accordingly...that way I get more level, and I don't upset the balance.

After that...if I need more, I'll do the other trick...bounce out a mix, bring it back, line it up, and see what combined peaks are keeping me from getting more total mic level, and if it's a combination of say...a snare hit, bass note and guitar note that are linging up together and creating a high peak, I'll lower all their peaks individually at that spot by a little bit.
If I'm working on just the final mix...I can do the same thing...manually lower the very high peaks...and then bump up my overal mix level.

I've done that many times, and found that by doing it in stages...I can keep nudging the mix level higher and higher, and I still haven't put any kind of limiting/compression on anything. Then when I'm done, I can add some limiting or compression to squeeze out more level without need to do extreme limiting and compression.

Finally...how loud are you really shooting for?
I find there are many LOUD mixes that sound good...but then some people want LOUD...and some want to go even crazier, and get LOUD, but you can hear the mix going to shit for the sake of that.

So there really isn't any single LOUD standard out there...and if you get it in the ballpark, it's good enough, and you don't have to mangle the shit out of your tracks.

 

[Nola]

Yeah...it would be tedious doing it for every single beat....but you don't really need to do that.
I usually look at the two main offenders....the snare and the kick...and I'll visually scan the track, and often there are some hits that peak substantially more than others. I mean...no one hits the snare the same way every time...etc...so I find that if I start by focusing on the odd very high peaks, and lowering them by a few dB...I instantly get that headroom back.
Sure...it's not as easy as slapping a plug on a track...but I've been doing it manually for a long time, and it's not as tedious as it may initially seem.

Heck...if you spend a few hours trying all kinds of plugs and approaches to fix the problem...you could probably run through your tracks in that same amount of time. I rarely have more than 30 tracks on any song...and of those, half are probably ruled out from the git-go as the possible peak culprits...so the rest can be gone through without a lot of work/time.
At first, it will feel tedious even doing a handful...but as you get better sat it, you will be surprised how fast you can highlight a peak, lower its level, and then move to another one. Left hand on the keyboard, right hand on the mouse...and then just assembly line the process.

That said...you have to pick something in your mix as the deciding loudness reference. Everyone has their approach...but I tend to use the drums as my level reference. IOW...I'll set the drum track level, and then I'll bring up the other tracks into balance WITHOUT touching the drum tracks. Once I get the balance...I then see how much more headroom I have left to play with, and then I can raise ALL the tracks equally/accordingly...that way I get more level, and I don't upset the balance.

After that...if I need more, I'll do the other trick...bounce out a mix, bring it back, line it up, and see what combined peaks are keeping me from getting more total mic level, and if it's a combination of say...a snare hit, bass note and guitar note that are linging up together and creating a high peak, I'll lower all their peaks individually at that spot by a little bit.
If I'm working on just the final mix...I can do the same thing...manually lower the very high peaks...and then bump up my overal mix level.

I've done that many times, and found that by doing it in stages...I can keep nudging the mix level higher and higher, and I still haven't put any kind of limiting/compression on anything. Then when I'm done, I can add some limiting or compression to squeeze out more level without need to do extreme limiting and compression.

Finally...how loud are you really shooting for?
I find there are many LOUD mixes that sound good...but then some people want LOUD...and some want to go even crazier, and get LOUD, but you can hear the mix going to shit for the sake of that.

So there really isn't any single LOUD standard out there...and if you get it in the ballpark, it's good enough, and you don't have to mangle the shit out of your tracks.

That was extremely useful, thanks for typing all that out. Thanks Steen, too, your post was also very helpful.

Regarding how loud...not too loud. I hate the loudness war stuff. I like to have my headphone amp volume at about 3 o'clock at the song be at a comfortable level where I can hear it well (without straining) and listen for long periods of time yet it's not fatiguing. I find that usually is between -19db and -16db. I'm sure that depends on each person's hearing and all that.

 

[Nola]

I did the drum compression, and then went in and manually reduced the most offending vocal peaks. Now my peaks are at -10.5, so I gained quite a bit of headroom doing that. They were -6db before doing that.

So now what would be the best/musical way to gain that in volume?

i.e. now that I fixed the peaks, what do I do to get volume while not getting harshness?

 

[Chili]

Just for the learning experience do a bounce of your song and import it as a stereo wav.
Line it up with your session and mute it.

Now you can see where the big peaks are, while you're working at the mix.

That's a great idea. I might almost use the method except I like to just smash my mixes with a limiter.

But for real, I will try that trick when I have enough songs together for another CD.

 

[Chili]

I did the drum compression, and then went in and manually reduced the most offending vocal peaks. Now my peaks are at -10.5, so I gained quite a bit of headroom doing that. They were -6db before doing that.

So now what would be the best/musical way to gain that in volume?

i.e. now that I fixed the peaks, what do I do to get volume while not getting harshness?

At -10.5, are you able to raise the fader more? Using a limiter doesn't always mean you're going to get harshness. In fact, it's only when you abuse the limiter; like me, for instance.

One way is to use mild compression before the limiter so it isn't working so hard.

 

[timvracer]

I have a follow up question. As was mentioned, the obvious spikes in my mix are the kick and snare. I actually did what was suggested previously, walked through the entire song (on all the songs I am mixing) and manually adjusted up and down the hits to normalize them reasonably. This particular drummer was pretty inconsistent, so it made my job a bit harder and necessary. However, my question is around compression. I am also trying to use compression to beef up the sounds of the drum hits just a bit, and protect from errant peaks. I had been using a super fast attack, as it seemed logical to me that if I don't compress the transient, I am basically not achieving the goal of controlling peaks.

Is this correct? Is it "bad" to compress drum transients like this. It sounds good... I think, but I am not sure I feel I am doing the right thing and missing out on some of the "life" of the kick drum "snap" and the snare crack. However, if I don't compress them, I end up having to turn them down so much that they are not coming through consistently. Manually getting this hit by hit correct is hard, because it's not just the amplitude, but certain snare cracks carry those piercing frequencies louder than others, regardless of the overall combined amplitude.

I know, I am not asking a real clear question, but more perspective on this would be helpful.