Limiter - master

andrushkiwt

Well-known member
I have a feeling I already know the answer, but I'll toss it out there to see what you guys think. Maybe there's another way I haven't thought of yet. Here's the deal:

Current project is about ready to go, on the mixing end. Of course, wanting to learn to do most things by myself, I'm applying a little master bus touchups, like always. Just like my mixing abilities, I think this is also improving. I've been lucky enough to have some of best (imo) mastering folks around help me out privately over the last couple of years, and I'm very grateful for everyone's time and contributions. I know they get paid to do this, and time spent helping me is possibly money lost for them- so I'm extremely thankful for those folks.

Now, one issue I've realized lately, and sure enough is proving to be an obstacle in my current track, is squeezing out/inching out some db's on the limiter, for general folks benefit. No - I do not want it squashed or crushed beyond recognition, I just want to get those extra db's to match nicely following a commercially produced track.

Here's the exact problem, and your response to this post will tell me if you've continued reading or if you stopped above and took this as a general "raising volume/limiting" question. My fault, I suppose, for rambling on too long without getting to exact issue - the snare. Forget what volume and measurement we're dealing with for a moment, because we'll get caught up in which methods are best, applicable, and troublesome, and just assume there's a snare that's causing too many peaks.

Muting the snare will allow me to raise my limiter about 10db of input, with its threshold at -1. With the snare engaged, and the limiter at 8db input, same threshold, I will get about 4-5 db of gain reduction. The snare itself is rather heavily compressed - 1176 style with med A and R. Then, further, the drums are compressed within Superior D's plugin, then again in the DAW instrument (drum) channel, again with an 1176 at med settings. Then, the whole tune is running through two slight compression plugs, API first then Tokelnikov second, the first one acting quicker than the second, but really, they both have minimal gain reduction happening. I'm talking 1-2 db's at most. Don't freak out and think that because there's two compressors in tandem that insane compression is happening. It's not, it's very minimal. After those two comes the limiter.

When I examine the individual snare hits in the MIDI file, yes, I can adjust them, be it velocity or timing. Moving the snare blobs in small increments forward and backward can drastically change how the limiter reacts at that moment of the song - the its velocity. The goal, or train of thought, in doing it this way is that I don't have to mess with the snare's mixing properties, which I really like, and I can just move that midi blob until the master limiter doesn't go crazy. But it takes a long time, and I have to continually rewind the track and play it from the beginning since the limiter is responding to everything happening just before it, and what happens before that snare hit affects how that snare hit is triggering the limiter.

So, I've already applied some snare compression, drum channel compression, master compression, and limiting. But the snare is still poking out too much. This is MIDI, so I can adjust timing and velocity. But it's tedious and seems to be random, as to whether moving the snare blob this way or that will either increase the limiters reduction (bad) or decrease its reduction (good). I want the limiter doing BARELY any reduction, but at the same time, I want to be able to raise the overall volume another couple db's. The snare is making that difficult, and, like I said, moving the hits around seems to be random as to how it affects the limiter. Sometimes moving the snare slightly forward decreases reduction, sometimes it raises it. Same with moving it slightly backwards. The other property I can adjust (just realized) is the length of the hit - quick, snapping-like hit, or a slower, drawn out hit. Again, both seem to affect the limiter randomly.

My best guess is that simply lowering the velocity of the snare, and compressing it more will be the answer to my limiter issue. It's just weird that moving it milliseconds changes the limiters' reaction so drastically. It's like the exact combination of everything happening at that exact moment in the song is pushing it over the edge, even if the snare is literally quieter.

It's a pain. How are you guys approaching your snares and their effect on your limiters? If you use one, to begin with.
 
This is just a "what if" WAG here 'cause I really don't know what I'm talking about... what if you copy the snare, lower the volume on the
hit a little, then invert phase and add it back in with the original to sort of shave it off a bit.
 
Two suggestions: parallel compression for the snare and/or parallel saturation. Having essentially two snare tracks, one with the attack squashed allows you to reduce the volume of both tracks without losing the snare in the mix, thus not hitting your 2 bus limiter with high peaks. Saturation, on the other hand , will also decrease peaks while increasing overall loudness , again allowing you to turn down the snare tracks with the same reduction in limiting.
 
It's just weird that moving it milliseconds changes the limiters' reaction so drastically.

No...it's really not that weird, and I think this was kicked around in some threads not too long ago...I don't remember where I posted it...but it's about the transients, and how they can add up on a beat or on a note when you have 20-30 other tracks all doing something in the exact same spot...and even a few samples movement/separation of 1-2 elements will yield a noticeable drop in the transient energy, whereas just hitting it with more compression or whatever, seems to have little effect.
Pretty much the same principal as with frequency build-up when similar elements are stacked across multiple tracks.

The other thing...some compressors and limiters do a better job across a mix, while others are more suited for individual tracks...so it's not always a surprise if one choice isn't yielding the desired result, and the more you push, it just kills things without fixing them.
With the right tool for a given situation, you should be able to crank a lot of dBs without squashing the crap out of things.

I'm also not sure if many layers of compression don't contribute to issues. I mean...there's the notion that you are doing something good and "controlling" levels/peaks, but it's also "compressing" things and packing them in...so that energy is going to build up.
I've never liked the sound of substantial layers of compression...I mean, something's gotta give. It's not always a cumulative win-win thing...but it's also going to be source/mix dependent.
 
Good point, Miro. Once that layer of comp happens, its now packaged like that for whatever comes next in the chain. That's true. Didn't think of it like that before.

What would be your first approach for getting a dozen or so snare hits under control, for the benefit of raising the limiter quite a bit?

Yeah, there's gotts be other things on other tracks adding up right in those instances. Guess I could check those out as well.
 
Go in an manually reduce their transient peaks...or nudge them a few samples to break up the transient energy.

I do 90% of that type of level control rather than using a comp. I've done dozens of beats/notes on dozens of tracks like that. It's tedious, but it's focused control. Actually...it's not that tedious anymore, I can do it pretty fast...it's just boring as shit.

TBH, I'm not sure why a couple of dBs of level gain would be that impeded by a few snare hits, but then, it's all relative to what have and what you've done and/or are doing...kinda hard to have a clear, singular answer.
 
The loudest I can get it right now is with having 8 db input on limiter, and that causes roughly 4 db reduction. I know that isnt a value of loudness and it means nothing in regard to how loud it actually is, but I'm not home at the moment so I can't give an exact number.

Basically, it would be at a good level for me if I can get the limiter another 2db. At 10db input, same -1 threshold, with no more than 4 db reduction. The 4 is because anything more than that sounds very noticeable. But 3-4 db is pretty much negligible, as far as noticeably altering the sound.
 
The loudest I can get it right now is with having 8 db input on limiter, and that causes roughly 4 db reduction. I know that isnt a value of loudness and it means nothing in regard to how loud it actually is, but I'm not home at the moment so I can't give an exact number.

Basically, it would be at a good level for me if I can get the limiter another 2db. At 10db input, same -1 threshold, with no more than 4 db reduction. The 4 is because anything more than that sounds very noticeable. But 3-4 db is pretty much negligible, as far as noticeably altering the sound.

Bus the snare to a group channel (or whatever it called in your software). You likely already do...

On that snare group only channel, do the compression you do, but then add a good limiter at the end. Back to my love for the PRO-L. It stops those transients in a way that you just don't hear. Then send to the main drum bus.

Just the way I do it myself. It just works for me.


And no, numbers don't matter so don't worry about them. Well, I hope you know what I mean. :)
 
Bus the snare to a group channel (or whatever it called in your software). You likely already do...

On that snare group only channel, do the compression you do, but then add a good limiter at the end. Back to my love for the PRO-L. It stops those transients in a way that you just don't hear. Then send to the main drum bus.

Just the way I do it myself. It just works for me.


And no, numbers don't matter so don't worry about them. Well, I hope you know what I mean. :)

I could do that. That's a good idea, actually. Glad I asked - everyone's got a different approach to this. Actually, Superior 3 has a limiter. I can try that, on just the snare channel.
 
It happens actually quite often. I mostly record live drums. Some drummers are consistent and some are not.

I tend to accent most with samples in different degrees. For the most part, it is the soft hits that get the manual edit. The hard hitting ones are sometimes all over the place as to how it hits a compressor. Sometimes you see the gain reduction, yet can't hear the attack of the snare. It weird...

So I also manually replace some of the hits that are whacky. The limiter at the end of the chain helps to keep things under control before the drum bus and therefore also the transients that do tend to hit the master out limiter.

Again, so many ways to skin the kitty. This way at least makes it less painful for Mr. Fluffy. :)
 
Well hello shit for brains
Ow you play the drums now do you or just remix other, like what you find on dudetube

looking at your second portfolio photo you look like you dont play anything infact you look a prick, but I guess that is why you run this site ayyyyyyyyyyy

Man, I really wish you would contact me directly again and have an honest conversation. Yes, I play 3 instruments and record/help/produce local bands and mix/master countless others of every genre. 35 years I have been doing various degrees of this. Never once have I disrespected nor done anything but by honest for the better of any situation. I just hope you will stop this thing you are trying to do as I have no anger towards you nor have I ever disrespected you. And by the way, I am merely a moderator here. I do not run the site. I just volunteer to help out because I care to do that for free.

I am one of those guys that will give you stuff for free because I wish to help those that show the desire to help themselves.

How about you follow suit and get over whatever anger you have. I surely never disrespected you.

PM me please. :)

JImmy
 
Slightly tediuou solution is to manually reduce the offending peaks. This leaves the majority of the snare hitting how you like it and the peaks NOT causing trouble.
 
Slightly tediuou solution is to manually reduce the offending peaks. This leaves the majority of the snare hitting how you like it and the peaks NOT causing trouble.

"Peaks", as in, coming from all tracks? Right? Because removing the loudest snare notes sometimes does not solve the problem. Taking the snare from a 97 velocity (my MIDI tops out at 100, not 120) down to a 94 doesn't affect the limiter reaction very much sometimes. And that's a big drop in perceived aggressiveness. Sure, dropping it to an 85 would work - but it'd be on par with the snares in the verse, not the heavier chorus where things have picked up.

So, like Miro said, there's things building up across all tracks at that exact point and causing the limiter to respond much more than it seems like it should. So, I think.
 
I sometimes call those "abberant peaks", and it really is just that everything in the mix just happens to be pushing hard in the same direction at the same time.

Have you tried inverting the polarity of the snare? That might get weird if it's bleeding into other drum mics, so you might have to invert those also, or maybe even just the whole drum buss. It will only accidentally help, though. It could as easily make things worse or more likely fix some of the peaks you've got now but cause others to pop up elsewhere.

A similar approach might be to use an allpass filter to rotate and "randomize" the relative phase of various frequencies in you snare track or drum buss. It's all the same uncertainty as above, though.

It is possible that there's just too much low frequency energy in that snare. Maybe if you highpass, you'll be able to keep the top end "voice" of the drum up front and present without contributing quite so much "meat" to the whole sandwich. It's likely that you could dial it in to where it doesn't sound or feel much different st all, but get rid of some subharmoic junk that will make a big difference. Using a "normal" minimum phase filter there will cause a bit of that phase rotation effect, too.

Honestly, though, I'd probably start by limiting and/or saturating the snare as mentioned above, but I would also have a saturation on the master bus. Like you said, the limiter is reacting to some sort of average over time, and are not usually great at catching the kind of really fast peaks you're talking about here. A clipper/saturation really will shave those things off no matter how fast or sudden they are, and if it's really just for a sample or two, you'll probably never even notice.
 
I sometimes call those "abberant peaks", and it really is just that everything in the mix just happens to be pushing hard in the same direction at the same time.

Have you tried inverting the polarity of the snare? That might get weird if it's bleeding into other drum mics, so you might have to invert those also, or maybe even just the whole drum buss. It will only accidentally help, though. It could as easily make things worse or more likely fix some of the peaks you've got now but cause others to pop up elsewhere.

A similar approach might be to use an allpass filter to rotate and "randomize" the relative phase of various frequencies in you snare track or drum buss. It's all the same uncertainty as above, though.

It is possible that there's just too much low frequency energy in that snare. Maybe if you highpass, you'll be able to keep the top end "voice" of the drum up front and present without contributing quite so much "meat" to the whole sandwich. It's likely that you could dial it in to where it doesn't sound or feel much different st all, but get rid of some subharmoic junk that will make a big difference. Using a "normal" minimum phase filter there will cause a bit of that phase rotation effect, too.

Honestly, though, I'd probably start by limiting and/or saturating the snare as mentioned above, but I would also have a saturation on the master bus. Like you said, the limiter is reacting to some sort of average over time, and are not usually great at catching the kind of really fast peaks you're talking about here. A clipper/saturation really will shave those things off no matter how fast or sudden they are, and if it's really just for a sample or two, you'll probably never even notice.


Great post, thanks!

Few things. The snare is high passed at about 100hz. Theres a small 200hz cut, where the beef of it really is (and imagine all the bass, guitars, and vocal buildup there). And there is some saturation engaged on the master - its the Schepps Omni channel from Waves. The first slot is saturation. I'll orobably make a post on that sometime too, seeing as the saturation options are "odd, even, and heavy" - and though the sound is very different from each (very cool on both odd and even), I can't quite tell what exactly its doing differently.

So, ill look at throwing a limiter on the track. Think thats my best bet.
 
Because removing the loudest snare notes sometimes does not solve the problem. Taking the snare from a 97 velocity (my MIDI tops out at 100, not 120) down to a 94 doesn't affect the limiter reaction very much sometimes.

Going 97 to 94 in the MIDI velocity isn't really much...but the thing you have to realize is that it's not just the MIDI velocity setting.
When you're using Superior Drummer (and I'm sure it's the same in other drum apps)...if you have it set up normally, with all the randomizing/humanizing stuff turned on as it is by default (which should be, to avoid the "machinegun" drum sound effect)...
...even when you set the velocity, Superior Drummer is always randomizing those hits on it's own, and substituting samples, and even randomizing the velocity internally (or by sample selection)...on every pass (I'm like 90% sure, because I've heard the differences in the MIDI VSTi drum tracks).

So sometimes you adjust velocity in the MIDI editor and a hit sounds right....then on the next pass it seems a little louder, then on the next pass not...etc.
That's what happens when you use SD in "live" VSTi mode...it's always randomizing/humanizing. That's a good thing...BUT...it could also create odd situations when you go to mix later on with all your tracks, and you end up chasing these velocity ghosts.
IOW...the velocity setting in your MIDI editor comes after SD...so it's more like an overall setting, but before that the randomizing/humanizing is still going on.

That's one of the reasons I like to sort out my SD drums and then I record/bounce them out as audio tracks, and put away the SD VSTi. In doing so, the drums are "locked" in as an audio track and there is no more randomizing/humanizing going on with each pass.
After that, I'm dealing with pure audio tracks and processing them...rather than trying to process a "moving target" with the VSTi drums. So then I'm going lower/adjust the transient of an audio track if I need to tailor a drum hit...rather than relying on the velocity settings in the MIDI.
Maybe not a major issue for some folks, or in all cases, but it could be a problem...because I've seen/heard it where the MIDI hit velocity changes on some beats per pass.

You can always turn off the humanizing stuff is SD and try to do it all in the MIDI editor...but then it's harder to get a natural, drum sound without endless mind-numbing editing.
 
Have you ever tried abusing that channel strip to analayze its behavior? Like, if you send it a 150dbFS signal, is it actually going to clamp to whatever its simulated rail voltage is, or will it let things escalate without a good hard limit? And does it have a time component of any sort? If you hit it with a Single sample DC step, does it catch the razor edge? I don't fuck with Waves, but I'd be interested to know if it actually works.
 
I sometimes call those "abberant peaks", and it really is just that everything in the mix just happens to be pushing hard in the same direction at the same time.

Have you tried inverting the polarity of the snare? That might get weird if it's bleeding into other drum mics, so you might have to invert those also, or maybe even just the whole drum buss. It will only accidentally help, though. It could as easily make things worse or more likely fix some of the peaks you've got now but cause others to pop up elsewhere.

A similar approach might be to use an allpass filter to rotate and "randomize" the relative phase of various frequencies in you snare track or drum buss. It's all the same uncertainty as above, though.

It is possible that there's just too much low frequency energy in that snare. Maybe if you highpass, you'll be able to keep the top end "voice" of the drum up front and present without contributing quite so much "meat" to the whole sandwich. It's likely that you could dial it in to where it doesn't sound or feel much different st all, but get rid of some subharmoic junk that will make a big difference. Using a "normal" minimum phase filter there will cause a bit of that phase rotation effect, too.

Honestly, though, I'd probably start by limiting and/or saturating the snare as mentioned above, but I would also have a saturation on the master bus. Like you said, the limiter is reacting to some sort of average over time, and are not usually great at catching the kind of really fast peaks you're talking about here. A clipper/saturation really will shave those things off no matter how fast or sudden they are, and if it's really just for a sample or two, you'll probably never even notice.

Polarity flip is something I also used to do that seemed to work in some live recording situations. Weirdly, sometimes it works to reverse the polarity of the top head. Rare occasion that I record a bottom snare head for ghost note clarity, it could be both or one that gets flipped.

This thread being about programmed drums, I am not so sure how that works out...I don't do a bunch of drum programming anymore. Mostly enhancement of acoustic drums recorded in studio.

Now as to your comment about limiters 'reacting to some sort of average over time', that is again why I love the FabFilter Pro-L. CPU hog when oversampling on master, but it reads those and grabs them in advance before playback, not as an average. The difference between a simple limiter and a 3rd party expensive one are things like this...

That does not mean the FFPL is the answer and resolve. The program material should be controlled in advance of the master out if done right. How to get there is a tough one to put a finger on.

So many variables... But that is what we spend so much time trying to figure it out. Learning is half the fun right!?!

:)
 
Last edited:
Now as to your comment about limiters 'reacting to some sort of average over time', that is again why I love the FabFilter Pro-L. CPU hog when oversampling on master, but it reads those and grabs them in advance before playback, not as an average. The difference between a simple limiter and a 3rd party expensive one are things like this...

The Limiter is literally the ONE thing I've never shopped around for. I still use my stock limiter in Studio One. I use their graphic EQ too, about 90% of the time, since I'm pretty used to it. But I recently picked up the SSL-G and API 560 on a huge Waves sale. I've used those in my current project and it works really well - boosting the guitar bus at some upper mid freq really sounds different than using the stock. Much more pleasant and natural sounding.

Anyways, maybe I should look at some other limiters. I haven't been able to get an overall level that I'm completely satisfied with, but, then again, I'm still figuring shit out. Like the above post - there's peaks that I need to figure out how to tame, etc...
 
You haven't tried the L2 yet? I think that's on a cheap sale too, not sure though. Then there's the L3, but I've never tried it.
 
Back
Top