gain structure and mixing aka THE MONEYSHOT THREAD

[subfect]

turn it off if it bothers you I tend to leave it off. I do my own "normalizing" hehe

[elyhess]

turn it off if it bothers you I tend to leave it off. I do my own "normalizing" hehe

How would you go about normalizing it yourself? Like i guess i just dont understand the purpose of using it! Maybe if its beneficial i can learn to work around it but i dont see a real reason!

EDIT: Read thru some forums some guys was just saying his preference "I trust my ears more than my eyes, so It's always off on my machine." kinda how i feel may just leave it off

[subfect]

I just put a limiter on the master and push it until it just begins to start clipping. It also gives me an easy way to automate master volume without actually touching the master volume (for whne you want to create clips of a tune or something).

[macc]

^ Nothing wrong with doing that of course. It's exactly what I mentioned on the first (?) page of the thread, or near the start anyway.

Just be careful not to start leaning on your limiter too much to hold your mix together... there be devils that way, arrrr lad.


BTW - possibility of an article coming from here. Sort of humourous, sort of helpful, sort of... well, taking the piss really Anyone interested enough for me to post it up? Won't be done for a while, mind.

[wub]

Anyone interested enough for me to post it up? Won't be done for a while, mind.

Very much so mate

[Undrig]

This thread is incredibly useful and I could easily see myself getting lost in some other less important area of the mix trying to overcompensate for what I was missing here. Thank you!

[BroPassFilter]

Reading through this thread made me join this forum. I have always been caught up with trying to achieve the maximum perceived loudness, because that's what everyone says to do. Never thought about approaching mixing in a way that allows the tune to breath a bit. I will definitely apply some of the technical/philosophical lessons learned in this thread, and I hope to learn more and improve my craft. Cheers!

[macc]

^ That might be the best user name I have ever seen

[BroPassFilter]

^ That might be the best user name I have ever seen

Thanks mate. People always seem to get a kick out of it. haha.

[Undrig]

Once again, this thread taught me that sometimes what seems obvious can be achieved by the exact opposite. I was doing exactly like macc said. Using compression and eq to tame shitty things that were happening in the mix by having zero headroom. This has really helped a huge amount. Really grateful for the time you've taken here to explain everything and keep things on topic.

[kstan8]

.

[AxeD]

Kind of confused by some comments in here.

When I'm programming my drums I don't think about headroom, I just have my interfaces volume high and the volumes within
reason quite low. If all the elements come through the mix how I want em to and there is no unwanted distortion.. what's the problem?
I'm probably going to college for audio engineering so I'll learn all the theory anyway but what am I missing here?

[eldoogle]

Axe

You don't want your tracks to be clipping.
You don't want inserts on your tracks to be clipping.
And you don't want the master channel to be clipping.

Clipping is when the audio goes over 0 and distortion occurs.

Gainstaging correctly helps stop clipping and helps you obtain a better mix down when all your sounds are jiving together.

[travis_baker]

Anyone interested enough for me to post it up? Won't be done for a while, mind.

Very much so mate

[Electric_Head]

Kind of confused by some comments in here.

When I'm programming my drums I don't think about headroom, I just have my interfaces volume high and the volumes within
reason quite low. If all the elements come through the mix how I want em to and there is no unwanted distortion.. what's the problem?
I'm probably going to college for audio engineering so I'll learn all the theory anyway but what am I missing here?

you are turning down your levels in reason which to an extent is gain staging
your interface can handle the volume boost

[macc]

Kind of confused by some comments in here.

When I'm programming my drums I don't think about headroom, I just have my interfaces volume high and the volumes within
reason quite low. If all the elements come through the mix how I want em to and there is no unwanted distortion.. what's the problem?
I'm probably going to college for audio engineering so I'll learn all the theory anyway but what am I missing here?

You're not missing anything - you're doing exactly what we're talking about

[laurend]

Kind of confused by some comments in here.

When I'm programming my drums I don't think about headroom, I just have my interfaces volume high and the volumes within
reason quite low. If all the elements come through the mix how I want em to and there is no unwanted distortion.. what's the problem?
I'm probably going to college for audio engineering so I'll learn all the theory anyway but what am I missing here?

That's the best strategy.

[hookey]

Sorry if it isn't really answering your question, but -3 is way too high.

Remember that 6dB is half. So if you have one element at -6, that is half your headroom gone. Two elements at -6dB each = all your headroom gone. Having the drums at -3 will leave you fighting against clipping and struggling to keep everything down and under control.

Rather, set your drums for *around* -8 / -10 (ie, a bit less than half). The bass - if we are talking a pure sine sub - would probably sit best a dB or two below that, any distorted/fullband bass sounds should be effectively treated as different entities and mixed appropriately (due to Fletcher Munson).

This leaves you with a few dB headroom, and everything else is just parsley. No more fighting anything, you *will* get repeatable and consistent levels in your mixes, and better mixes as a result.

Hey sorry to reply to such old message but since u are still around I hope you can answer.

You are talking about your beat peaking at around -8/-10 db range, 6 TOPS. Im guessin this means you essentially leave all that headroom for the mastering engineer to work with and push it louder, doint it properly (we are on a dubstep forum and most people here like it loud after all, I do).

WHat took my attention was stumbling upon this the other day:

http://sofishfly.com/2011/02/skrillex-got-skillex/

Apparently the track "Cinema" that is around is an unmastered leak. I always find interesting when I find tracks of big name producers that have all those super expensive mastering companies behind and you wonder if the raw mixes are really different from the final product. So i downloaded the track and ran it throught my meter. You don't really need a meter to notice this is real loud before getting a proper master and is already peaking hard at 0db. It sounds really good and isn't even mastered.

So my questions are:

1) How is it possible to master a track that is already peaking at 0db
2) Is it really worth making such quiet mixes when you are aiming for really loud dubstep/dnb type of music? (club music, were the quietest parts are around -3.5db, and the rest is straight 0db peaks). I mean, what do you gain for it? Wouldn't you get a more polished track if you threated every part of your track separately and peaking at around -3db tops, using limiters and whatever is needed individually on the beats/whatever, then let the engineer do minor tweaking? Shouldn't a loud mix be already loud? I mean I don't see a track like "Scary Monsters and Nice Sprites" (to mention a really loud famous one) peaking at -8db before somehow the mastering engineer craking it to 0db retaining punchyness and nicely squashing it (squashing it in a way that sounds loud but nice and cool, like on SMANS or just any Skrillex track, not amateur-producer-waves-limiter-on-master-channel squashed, if you get what I mean). Wouldn't going from -8db peaking to real-club-loudness-0db banger somehow mud the mix and not even make it as loud as you wanted? I hope you get what im trying to explain. Maybe it's just feeling but I just don't see any those modern dubstep and dnb tracks being that quiet prior profesional mastering.

[skimpi]

Sorry if it isn't really answering your question, but -3 is way too high.

Remember that 6dB is half. So if you have one element at -6, that is half your headroom gone. Two elements at -6dB each = all your headroom gone. Having the drums at -3 will leave you fighting against clipping and struggling to keep everything down and under control.

Rather, set your drums for *around* -8 / -10 (ie, a bit less than half). The bass - if we are talking a pure sine sub - would probably sit best a dB or two below that, any distorted/fullband bass sounds should be effectively treated as different entities and mixed appropriately (due to Fletcher Munson).

This leaves you with a few dB headroom, and everything else is just parsley. No more fighting anything, you *will* get repeatable and consistent levels in your mixes, and better mixes as a result.

Hey sorry to reply to such old message but since u are still around I hope you can answer.

You are talking about your beat peaking at around -8/-10 db range, 6 TOPS. Im guessin this means you essentially leave all that headroom for the mastering engineer to work with and push it louder, doint it properly (we are on a dubstep forum and most people here like it loud after all, I do).

WHat took my attention was stumbling upon this the other day:

http://sofishfly.com/2011/02/skrillex-got-skillex/

Apparently the track "Cinema" that is around is an unmastered leak. I always find interesting when I find tracks of big name producers that have all those super expensive mastering companies behind and you wonder if the raw mixes are really different from the final product. So i downloaded the track and ran it throught my meter. You don't really need a meter to notice this is real loud before getting a proper master and is already peaking hard at 0db. It sounds really good and isn't even mastered.

So my questions are:

1) How is it possible to master a track that is already peaking at 0db
2) Is it really worth making such quiet mixes when you are aiming for really loud dubstep/dnb type of music? (club music, were the quietest parts are around -3.5db, and the rest is straight 0db peaks). I mean, what do you gain for it? Wouldn't you get a more polished track if you threated every part of your track separately and peaking at around -3db tops, using limiters and whatever is needed individually on the beats/whatever, then let the engineer do minor tweaking? Shouldn't a loud mix be already loud? I mean I don't see a track like "Scary Monsters and Nice Sprites" (to mention a really loud famous one) peaking at -8db before somehow the mastering engineer craking it to 0db retaining punchyness and nicely squashing it (squashing it in a way that sounds loud but nice and cool, like on SMANS or just any Skrillex track, not amateur-producer-waves-limiter-on-master-channel squashed, if you get what I mean). Wouldn't going from -8db peaking to real-club-loudness-0db banger somehow mud the mix and not even make it as loud as you wanted? I hope you get what im trying to explain. Maybe it's just feeling but I just don't see any those modern dubstep and dnb tracks being that quiet prior profesional mastering.

that cinema track may not be mastered, but it is probably limited to squash the dynamics and help it hit in the club better. this would then bring the track up to 0db and not let it go over.

and the reason you need to leave headroom before mastering, is incase the mastering engineer needs to do some work to the track, like EQing. if your track is peaking at 0db already, then the mastering engineer can boost any frequencies with the EQ cos it would push it over 0db. its also risky attenuating any frequencies as then some phase problems from the EQ could also push it over 0db, even if you arent actually boosting any frequencies. Also things like saturation, the mastering engineer cant do any of that, as it would also push it over the maximum. If you have left headroom, then the engineer can do whatever they need to do to the track, and not worry about breaching the maximum level.

[macc]

Hey sorry to reply to such old message but since u are still around I hope you can answer.

You are talking about your beat peaking at around -8/-10 db range, 6 TOPS. Im guessin this means you essentially leave all that headroom for the mastering engineer to work with and push it louder, doint it properly (we are on a dubstep forum and most people here like it loud after all, I do).

....

So my questions are:

1) How is it possible to master a track that is already peaking at 0db
2) Is it really worth making such quiet mixes when you are aiming for really loud dubstep/dnb type of music? (club music, were the quietest parts are around -3.5db, and the rest is straight 0db peaks). I mean, what do you gain for it? Wouldn't you get a more polished track if you threated every part of your track separately and peaking at around -3db tops, using limiters and whatever is needed individually on the beats/whatever, then let the engineer do minor tweaking? Shouldn't a loud mix be already loud? I mean I don't see a track like "Scary Monsters and Nice Sprites" (to mention a really loud famous one) peaking at -8db before somehow the mastering engineer craking it to 0db retaining punchyness and nicely squashing it (squashing it in a way that sounds loud but nice and cool, like on SMANS or just any Skrillex track, not amateur-producer-waves-limiter-on-master-channel squashed, if you get what I mean). Wouldn't going from -8db peaking to real-club-loudness-0db banger somehow mud the mix and not even make it as loud as you wanted? I hope you get what im trying to explain. Maybe it's just feeling but I just don't see any those modern dubstep and dnb tracks being that quiet prior profesional mastering.

1 - You turn it down before starting work. See discussion later on.

2 - This is actually lots of questions I'll try to answer it all, as it is very important. This may ramble a bit, so I apologise in advance. Also note that I'll use fairly extreme examples to make the point, but I actually see mixes like mixdown B on a very regular basis. It also doesn't apply to perfect mixes, but I don't see too many of those, even from the more experienced/bigger producers I have worked with.

What you're asking is the key point behind the entire thread. It's absolutely imperative that you are able to distinguish between powerful mixes and loud mixes. The thing you're missing is that having room for sound to breathe actually allows a good mastering engineer to get a louder, cleaner result.

Key points: It is impossible to remove distortion, and it is extremely difficult to put back punch that has been already removed trying to make things loud.

So, say we have a mix. As you say, care has been taken on every individual channel, to balance the impact of a sound with the fullness, each sound has its own space, each sound has the right balance in itself and also in relation to everything else, the right amount of dynamic control without flattening, transients are well-controlled, but still there. Blah blah, you get the point. This mix peaks around -6dB. Let's call this Mixdown A.

Now we have the same mix, but pushed aggressively. Say clipped by 6-8dB, in an effort to make it as loud as [some other tune]. This is Mixdown B.

Both these two versions of the same mixdown go to mastering. Let's imagine that the engineer (for whatever reason, not important) likes to start with a mix sitting at about -6dB peak level.

Here's the problem: this otherwise good mix has 5dB too much sub at 40Hz.

Mixdown A: already peaks at about -6dB, doesn't have any massive out of control peaks but still has punch/snap/definition. No distortion overall, and no distortion or pumping when the bass hits. Because the bass isn't interacting directly with anything else, making the required 5dB cut in the bass gives you the same overall sound with less bass. Easy. Now a good limiter allows you to keep quite a lot of the punch, as it has already been well-controlled, and put the loudness where you need it. You will have some small amounts of distortion (predominantly intermodulation distortion, essentially inaudible with a good, correctly-set limiter). Overall the final sound remains extremely clean, with space and impact at a big loudness. Easy peasy.

Mixdown B: needs to come down by 6dB (probably a touch more to make it equal loudness). Once that is done, it is clear that compared to mixdown A there is significant distortion, with everything dropping out when the bass hits. It sounds flat, the drums don't snap/hit like drums do in mixdown A. Doing the same 5dB bass cut makes the overall spectral balance right, so far so good. The thing is, now when the bass plays the distortion is still there - you can't get rid of that. To make it worse, because there is less bass, the distortion is even more audible. Great

Now we have the problem of getting some punch into the thing, to stop the drums sounding like someone punching a wet pancake. We have to use some sort of dynamic processing (whatever, not important) to get some impact back into things. So what we have now is 'fake-transients-made-from-previously-squashed-and-distorted-transients'. But it's better than no impact at all, I suppose.

With that done, now we have to get the level back up to where it was before, or louder (cos mastering 'has to make everything louder') right?. So we have to try to keep the impact we just made while limiting all the 'fake-transients-made-from-previously-squashed-and-distorted-transients'. The small amount of intermodulation distortion you get from the good limiter is now being applied to an already-distorted signal, which increases the total amount of distortion even more (google intermodulation distortion and it should become clear why). You're distorting the distortion, you have absolutely no choice. You minimise that as much as you can, but there's no getting around it.

I don't think I need to describe the result. It's better, yes - the tone is correct and it has more impact. But it has higher distortion, sounds flatter, and has much less space than Mixdown A. All we needed was for the thing to be mixed down with headroom.


To try to summarise, this idea:

'Shouldn't a loud mix be already loud?'

is a BIG problem when people with no experience in mixing misunderstand it. In a desperate battle to make their tracks as loud as [whoever], they batter the master buss as hard as they can and end up with an unclear mess. Corrections made in mastering can only have limited success. Of course, it's different if you make perfect mixes all the time. Most people don't. The further from perfect your mix is, the more difficult you make things by spanking it.

So I'd suggest rephrasing it; 'A mix that is intended to go loud should have the ability to go loud'.

There are a lot of ways of checking that your mix can go loud. Use them. Spank the absolute arse off your tune (at the same subjective level!!!) to make sure it doesn't fall apart. I often use 20dB+ of compression - yes, 20dB+ - on my master buss to see how my mixes behave under extreme pressure. But I turn that stuff off before giving it to someone else for mastering. FWIW I don't master my own mixes. What if I made a mistake? I've heard the tune 50283093274098 times. Best I don't set stuff in stone, and allow some 'wiggle room' for the person who is going to quality check it. I know I have given the track every chance of going super-duper loud AND clean, I've done the best I possibly can. But I don't try to do the final step before the right time.

6dB headroom, 1dB headroom, 24dB headroom, it doesn't matter at 24-bit - just don't spank your mix. It's like asking someone to proof-read and edit your new book, then giving them a scanned copy of a printout instead of the Word document.


Let go of your ego. Ignore the feelings of inadequacy and have the courage to submit something that isn't spanked. Make a powerful mix that has the ability to go loud, and find someone you trust to check/correct/finalise things for you. Your sound will thank you for it.