Too compressed

Hi, I've been trying to get better at mixing. Gonna cut a long story short and get to it...

My drums (EDIT: VST drums...EZD, AD2, Ujam etc) are sounding really compressed and weak when played back on laptops especially...but not so bad on phones. Uploading to soundcloud the drums sound just hugely clamped on and squashed and small and lifeless. In my monitors, headphones and VSX mixing headphones I'm not noticing this.

What am I doing wrong?

I've started recently getting more into clippers to control the drum bus...aiming for -23lufs (roughly) and peaking like -7 or 8dbfs. Which I thought sounds reasonable enough. The snare and kick are usually the 2 most peaky aspects of a mix. And the issue I get is that I want my song to be -12lufs integrated right...not slammed by any means at all. However to get the song up to that level unless I take care of the peaks in the drum bus I'll have a limiter or a limiter and compressor on the mastering bus potentially clamping down like 6 or 7db to keep the peaks below 0 while getting the lufs up to -12db. This seems wrong to me or really far from ideal if you've got a limiter chopping 6, 7, 8db peaks. So that's why I try to control the drum bus average level and peaks before mastering.

Anyway, that's been my current kind of practice. But I'm finding my mixes sound poor on soundcloud and poor on a laptop and a tablet...but not too bad on a phone. Naturally enough...as I'm sure you could imagine...my mixes sound good on my HD600's (this hasn't always been the case, oh no, far from it...so I've made some progress), they sound good on my Yamaha HS8's and decent on my Mixcubes and good in Slate VSX and good on my phone. But they are still wrong and bad because soundcloud chews them a new one and laptops and tablets sound pretty bad.

I don't even know fully what my problem is. Could be just bad songs and bad arrangements, bad recordings, bad everything. I thought maybe it is because I'm using VST drums...that are already mixed and compressed and that I'm compressing them again. But in the mix I really don't compress them very much at all. Nowhere near slamming or anything close to that. I tame the peaks...because like I said I don't want outrageous kick and snare peaks that'll just have to be squashed down at mastering. Any tips? I can upload a song to hear if you want. Second song 'Another Time' in particular has these thin weak drums. Part of it might be just my selection of drums in EZD. It's some singer songwriter acoustic kit with a wimpy snare that's maybe better suited to a different kind of song like a mellow acoustic number or something.

Thanks
 

Attachments

  • Caterpillar Man.mp3
    5.6 MB
  • Another Time.mp3
    8 MB
Last edited:
Err...having another look into all this...the 'Another Time' song just seems to be a combo of bad VST drum choice, ponderous song and just garden variety bad mixing. I've been on a mixing run lately doing about 5 songs or so and fell to that problem where you just pretend that everything is OK without being truly ruthless in your assessment. Just listening again on my solo Mixcube and even basic levels sound off to me and there's a veil of tubby low mid range in a lot of the guitars and the 2 key L and R rhythm guitars just sound lame.

I gave it some thought and realised that among many errors, I've actually been mixing on Slate VSX in one VSX room using only "Farfield" speakers and cross checking on HD600 headies turned up forgivingly thusly hiding many errors. I didn't bother to cross check in any other VSX rooms...OR VSX midfields or nearfields. And I gave only cursory listens on my Cubes and Yamahas. And I also had not bothered to check as I did mixes by checking on laptops etc.

Seems I just found that VSX Farfields and loud playback checks in HD600's sounded good and didn't want to get into the weeds of anything else.

So, my bad. As usual.
 
They do indeed sound compressed and dull. So why not just turn off the limiters and compressors and see what that sounds like? If some of the drums are ok - and other components need treatment - separate them and process what needs processing.

All this LUFS stuff in this section suggests you're mixing by numbers
I've started recently getting more into clippers to control the drum bus...aiming for -23lufs (roughly) and peaking like -7 or 8dbfs. Which I thought sounds reasonable enough. The snare and kick are usually the 2 most peaky aspects of a mix. And the issue I get is that I want my song to be -12lufs integrated right...not slammed by any means at all. However to get the song up to that level unless I take care of the peaks in the drum bus I'll have a limiter or a limiter and compressor on the mastering bus potentially clamping down like 6 or 7db to keep the peaks below 0 while getting the lufs up to -12db. This seems wrong to me or really far from ideal if you've got a limiter chopping 6, 7, 8db peaks. So that's why I try to control the drum bus average level and peaks before mastering.
I'd see me mixing and then maybe thinking the snares a bit too much, so I'd compress it maybe a tiny bit - but if I have it as a separate source, I just pull back the fader. I'd find a similar song on spotify to just check how loud that was. It's a quite traditional song - so plenty of examples and none of them are that dipped in maximum levels and compression.

Have you actually tried mixing it with no compression or limiting at all on the drums and just balancing the drums against the rest? I'd bet the snare perks up, the hats sizzle a bit and the who drum section brightens up. Once you have a mix, you can look at level indication for the final - but you have such a lot of dynamic range nowadays, it seems odd to me to even think of LUFS when you're mixing with EQ and levels being explored? Set the maximums and minimums once you're finished making the artistic changes, and then make only those needed to control things. Maybe you just got carried away with dynamic control way too early - using it unmusically? I can't remember using a limiter for years now. never actually needed one?
 
Thanks Rob. Yeah I can try mixing again turning off the limiters and compressors. Mixing by numbers...yes I am using that as a crutch or guideline or something. My main reason being that if I got a mix that was peaking digital scale at -3dbfs, say...and the average level was about -23...and I knew that ultimately I wanted to master it so the average level was -12...well then that's a further 11db I need to increase my mix...and if it's peaking at -3 and I can't go over 0/ zero...well that'd be 8db that the end of the mastering chain limiter would be clamping down on. And I thought that would not be ideal. So I was trying to get a mix peaking at say -6 with an average level of maybe -20. That way when doing the master I only have to raise that average level by 8db...and with the -6 peaks and the 6db buffer to 0/ zero the end of the chain mastering limiter would only clamp down -2db. If you know what I'm getting at.

Anyway, that was my crazy theorizing and planning.

I haven't tried mixing it with no compression or limiting on the drums yet. I will though.

If you haven't used limiters for years...how do you keep masters below 0 while achieving an overall/ average level you're satisfied with? Are you using clippers? Tape plugins? Saturation etc? When you say it seems odd to think of lufs when mixing...well I'm only thinking of it because I know I'll mix down the song and attempt to master it some days later. And in mastering I want a certain average level I suppose for the song. I mean if I didn't care about an average level/ lufs then I could mix the song at -25lufs and have this super quiet song. Alternatively I could master it so it's -6lufs trying to be the loudest on the block. So my -12lufs I feel is just a decent level to aim for. And like I said before, I am watching levels when I'm mixing...a mix at -25lufs with -2dbfs peaks...well I know in mastering that I'm gonna have issues with the limiter taming really amplified peaks as I push the track into the limiter to boost the volume.

I'm probably getting too old for all this. The arrangement of my songs is likely just quite average, tracking and recording probably just captures way too much mud in my 8x9ft room and my selection of VST drum sounds is probably often just off by enough to sabotage proceedings. And a million other things including bad singing and playing etc. It's not one thing in other words...it's everything. I'm not 20 anymore and I'm just not making decent progress. One thing that always seems to happen is that early on when mixing I'll get a drum and bass and rhythm guitar sound or a foundation and I'll just call it good and build the rest. I call it good without properly referencing proper music. In this way, I'm kind of pretending that my foundation is sounding good. When in reality it's far from good. You know I guess if you turn your headphones up enough or your speakers or if you squint while you listen anything can sound alright. Having said that though...it's probably not that...or any other specific thing...like I said before...it's just everything.

Anyway, thanks.
 
Listening to the two tracks, I think the problem is more that the snare level is high relative to the other drums. It really pops out. When I look at the track, all you see are the snare peaks, which seem to be 3-5dB above all the rest of the music. I have the snare in my MTPDK pulled down compared to the toms. Plus, they are very precise, each hit it basically the same dynamic.

Another Time.jpg

Can EZD "humanize" like MTPDK? You might try to do that on the snare drum, in addition to pulling the level down some. Drummers will vary things to keep it interesting, otherwise it can sound too much like a metronome.
 
Last edited:
I just listened to both songs. They are both great songs, with character.
To me, the drums on Caterpillar sounded detached, until the guitar solo, when it all came together.
Maybe the composition of the drums coud be tweaked.
On 'Another Time' it is just a simple driving rhythm, same all the way, which does make it a bit dull.
Around 2:15, the music goes dum - dum - dum. I wanted to hear the drums pick out that same sequence, with a different cymbal for each dum.
You're using artificial drums, but you need to get your head into real drummer mode, and think what would a real hot drummer play to the music.
There's got to be variety to keep the drums interesting, but also you have to think about how the drummer's arms and legs would move, to keep it real sounding.
The toms don't get a look-in.
You might switch from hi-hats to ride cymbal for a section.
A snare hit and crash cymbal together sounds good.
Hitting two toms at once sounds good.
Chuck a flam in here and there, if it fits.
If you're not a drummer yourself, study what top drummers do on Youtube, to make performances interesting.
As far as compression, I'd say do it spareingly, while picking over the results.
 
Listening an EZD "humanize" like MTPDK? You might try to do that on the snare drum, in addition to pulling the level down some. Drummers will vary things to keep it interesting, otherwise it can sound too much like a metronome.
thats like reverse quantization. like have you ever recorded your beats and thought..im too 'on time'..lets ruin it. it never seems robotic.

unless you go 'vocoded roto toms' each with their own notes in key with each other.
 
On processed drums, the old SD2, I almost never compress. I either edit the volume in the mixer or the velocity in the MIDI.
 
thats like reverse quantization. like have you ever recorded your beats and thought..im too 'on time'..lets ruin it. it never seems robotic.

unless you go 'vocoded roto toms' each with their own notes in key with each other.
Man, it's a universal concept. Increased entropy is the way the universe is headed. Randomness is the future..... Order must be abolished!
(you of all people should understand this!)
 
I'm not directing what I'm going to say at anyone in particular. Just addressing something I've read/heard way too many times over the years....

I don't understand why non-drummers always claim that real drummers will play off the beat to sound "human". That's simply not true. If a drummer played right on the grid for every hit, it would still sound human, as long as he's doing what I'll mention below. I've never once heard a drummer play on time and sound "robotic". A drummer can play perfectly on beat and still sound human.

Do guitar players de-tune their guitars to sound more "human"? "Oh no! I'm perfectly in tune!!!!!"

The important thing isn't varying timing/hit placement. The important thing is varying VELOCITIES. For example, playing all your hi-hits at the same volume will sound robotic. Using accents on the upbeats/downbeats is what makes you sound human.

That's what makes funk drummers funk. It's a huge part of what makes John Bonham's playing so recognizable. It's what brings the Purdie shuffle to life.

That's also why I can tell when a non-drummer programmed a drum machine. The accents/dynamics aren't there....or they're all wrong.

I've read people say things like "Play your snare a little later than your hi-hat". That is like the stupidest thing I've ever heard. I don't know a drummer who can even do that if he tried. And I don't see any reason anyone would even want to.

Don't worry about trying to play "not perfectly on time". Work on your dynamics.
 
Last edited:
Similarly to what @rob aylestone says -- in my experience, VST drums generally sound best with minimal processing. I usually just put a touch of saturation on the master buss and leave it at that.
Ok, I'm going to try to be really minimal on the VST drum stuff.
VST drums and EZD are already processed - aside from a little reverb they don't need much.
Yep. Ok, ditto to this. Will try a very light hand.
Listening to the two tracks, I think the problem is more that the snare level is high relative to the other drums. It really pops out. When I look at the track, all you see are the snare peaks, which seem to be 3-5dB above all the rest of the music. I have the snare in my MTPDK pulled down compared to the toms. Plus, they are very precise, each hit it basically the same dynamic.

View attachment 135407

Can EZD "humanize" like MTPDK? You might try to do that on the snare drum, in addition to pulling the level down some. Drummers will vary things to keep it interesting, otherwise it can sound too much like a metronome.
Ok yeah see this kind of waveform is one I seem to end up with on my mixes a lot. And this could be the reason why the mixes sound particularly bad on soundcloud...because Soundcloud sees those large peaks and stamps on them which I presume sucks the life out of everything else including the drums/ snare. I don't think I linked to my Soundcloud. But Soundcloud sounds worse than the mp3 versions I attached. As for the uniformity/ precision of the snare hits...I think this is either just like you say a humanize thing or it's the clipper that I got in the habit of using on the drum bus to handle the levels as described in my earlier posts. Getting back to the waveform...I know it's wrong to judge by waveforms. But I have always thought that something is wrong...I'm doing something wrong when my songs end up looking like that. It just doesn't look right.
I just listened to both songs. They are both great songs, with character.
To me, the drums on Caterpillar sounded detached, until the guitar solo, when it all came together.
Maybe the composition of the drums coud be tweaked.
On 'Another Time' it is just a simple driving rhythm, same all the way, which does make it a bit dull.
Around 2:15, the music goes dum - dum - dum. I wanted to hear the drums pick out that same sequence, with a different cymbal for each dum.
You're using artificial drums, but you need to get your head into real drummer mode, and think what would a real hot drummer play to the music.
There's got to be variety to keep the drums interesting, but also you have to think about how the drummer's arms and legs would move, to keep it real sounding.
The toms don't get a look-in.
You might switch from hi-hats to ride cymbal for a section.
A snare hit and crash cymbal together sounds good.
Hitting two toms at once sounds good.
Chuck a flam in here and there, if it fits.
If you're not a drummer yourself, study what top drummers do on Youtube, to make performances interesting.
As far as compression, I'd say do it spareingly, while picking over the results.

Right with you on this. Thanks.
"Another Time" sounded good to me, much better snare. Overall, the drums seem to suit the genre/style...
So, I guess the question is, what kind of sound are you shooting for?
Just the Holy Grail of home recording...mixes that translate on as many playback situations as possible. I appear to have made the mistake in mixing these recent songs of a) only using Farfield speakers in one Slate VSX room, b) only checking the mix on HD600's at generous volume, c) only checking the mix on my phone - where things didn't sound so bad to me. I failed to check on my laptop or anything else. The mixes when heard on Soundcloud on my laptop just sound lame. Thanks for the kind words. But I'm a long way off yet.
On processed drums, the old SD2, I almost never compress. I either edit the volume in the mixer or the velocity in the MIDI.
Right. Well I need to follow a similar game plan here. Instead of me trying to be Mr Big Mix King here on already mixed drums.

I tell you what fellas...it's a real hill to climb doing it all over again. I thought I'd done some good mixes this time. But back to the drawing board.

Thanks for the advice.

EDIT: Oh...I wanted to ask if anyone has any thoughts on lufs level of a drum bus? Should I care? I was thinking that if drums are so much like short, transient hits...yeah, do I care so much for the average level of the drums? When mixing drums, should my main concern be the sound (naturally) and where stuff is peaking rather than the rms/ average level of the drums?
 
On individual busses - the levels are unimportant as long as they are not very low or very high - if you are glued to the LUFS approach, then it's on the mix to apply this level monitoring. The reality is when you mix with lots of busses - you just need the right structure to give the faders decent range. I tend to use VCAs now so even things like reverbs and the minimal processing are done per track rather than on a group.
 
Back
Top