gain structure and mixing aka THE MONEYSHOT THREAD

[SunkLo]

Peak levels are only really important when recording and in relation to RMS levels.
You ear detects RMS as how load a sound is. The difference between that and the peak level is part of the dynamic range. Compression can be used to reduce that range. Upward compression can bring up the levels below the RMS. Basically you just need to be concerned with the RMS and the dynamic range, as that is going to be the most audible aspect.

Like I said, in a mix you're going to be leveling things by RMS. If the peak level is a lot greater than that, you'll have a very dynamic sound and may have to watch for clipping. But still the important value is the RMS, the perceived volume. This is what you'll set levels with, and this is what people's ears will perceive as the volume (hence perceived loudness)

The digital age has placed too much emphasis on the peak portion of a signal. There's a section in Mixing with your Mind I think that discusses it (look up the skyscraper analogy). Basically with tape, you go too quiet and you get noise, too loud and you get saturation. So your goal is to get the signal at an appropriate spot in between, and this middle of the dynamic range is where it sounds best. With digital the signal loses detail the quieter you go but there is no repercussions on the maximum end of the scale other than the clipping limit. Thus people are always trying to get things as hot as possible and are paying so much attention to the part of the signal at the maximum region of the dynamic range (the peaks)

Dynamic range is a balance point and how much a signal deviates above and below that point. The tape medium reflects this quite nicely as it puts its optimal level, where the signal is cleanest in line with the dynamic balance point (RMS) Digital focuses on an upper limit and how much the signal goes below, which is not inline with how the ear works. Unfortunately because of the peaks, the signal has to be lowered below the optimal signal level so that there is no clipping. The peaks dictate where the level sits, vs in analog, the RMS dictates where the level sits. Your ears are like analog. They don't care about the peak values to an extent, just the average. If you focus on that average and the range of deviation from it, you're thinking of dynamics in the manner that you are actually hearing.

For instance, if you're bypassing a compressor to A/B it and you match percieved levels it's easier to judge the difference in dynamic range rather than just the volume change which is a byproduct.

[macc]

RMS is practically just as irrelevant as peak levels as to how loud things sound, in my opinion and experience. The spectral balance of the track, via Fletcher Munson, has far more say on how loud something is perceived than the RMS (assuming they are at least comparable). Add 3dB to a track around 2-3kHz and turn it down until the RMS numbers, assuming they are unweighted, match. QED.

I honestly can't remember the last time I analysed a track's RMS. It must be at least 6-7 years. It's only meaningful when you assume the mix/spectral balance is perfect. Usually, it's not. I can show you a track with way too much bass that has high RMS but doesn't sound loud. I look at the RMS meters from time to time, and how they're working in comparison to the peak levels. And that's the real thing;

Because peak volume is irrelevant in most situations regarding percieving whether something sounds better.

I kind of disagree with this. Or in some situations I disagree with this. A lot of times, I have something that is peaking at a reasonable level, at a workable level relative to other elements of the mix, but the element doesn't sound loud enough. So I end up attempting to increase the perceived volume while keeping the peak at the same place. Is that making sense? Having a helluva time articulating the simplest shit right now.

You're touching on something extremely important and very profound, to be honest (sorry, am I being austere? ). It's infinitely more complicated than 'RMS matters more', it's a complicated and dynamic (no pun intended) thing. A gunshot sounds loud despite having relatively low RMS. Relatively low-peak level white noise can seem bloody loud, despite having no peaks. So just forget all that shit and just listen.

You're also touching on what I have said about in this thread a few times, about the inherent POWER in a mix. More impact, less voltage. Keeping the punch, but making it fat. That balance, and the way it ties into how the ear hears things, is what the art of (modern/non-classical) mixing is all about. Punchy and fat (and big and warm and tight and clean blah bah). 'Mutually exclusive' properties, all at once. Working with the ear's delicate perceptual balance. Peak/RMS/crest/spectral balance, all that... numbers will only get you so far.



As for the AB thing; one way to assess differences is to choose one element - ideally the most important element that sits in a relatively sensitive area. Say, vocal, snare drum, lead guitar, that sort of thing, depending on the music. Make sure that element sounds the same level in the AB, and assess how the rest of the track changes around it. Then you know what you're doing. Does it now get swamped by everything else? Does it sit back in the field? Does it sit better, stick out, blah blah. I tend to favour a more holistic thing rather than one element these days, but tbh I couldn't really describe what I'm listening for, just matching them so they sound the same level, and seeing what happens to the sonic image.

Ok, so this was a pretty austere post. Bollocks

[SunkLo]

Sorry meant to throw in at the end that I was using RMS loosely to mean perceived volume in some cases, just easier to type Your RMS meters are going to be showing different things depending on how you have them set up as well, so you have to take that into account. I know Reaper lets you calibrate your vu meters to different speeds, K-system etc. Not sure about other daws but I'm sure there's some good 3rd party meters available.

Right on macc about the mix power vs meter readouts. More impact, less voltage sums it up nicely. PPMs or Vus will only get you so far. I was trying to point out that a lot of people are paying too much attention to their PPMs. There's weekly threads on "what levels should my drums peak at?" etc. I think VUs provide more valuable information in a lot of cases but meters alone will only get you so far. Nothing really that will tell you how much punch or cohesion a mix has.

[nowaysj]

So just forget all that shit and just listen.

This is what it comes down to, and this is where I have my problem.

Just about every plugin I work with adds gain just by turning the fucker on. This should be illegal in my opinion. Swat teams should come crashing through skylights when a dev starts coding their shit to add gain like this.

So I instantiate a plugin, and now it's louder so it sounds better. Right here is where the trouble starts. First thing I have to do is turn down the plugin. So I've been using my peak meters to do this. Apparently wrong according to the benevolent (and austere) macc. Okay, so I'm supposed to reduce the gain, by ear, according to the perceived volume. But as was discussed:

Louder = Better
Better = Louder

where '=' means sounds like. If you take this to heart, there is trouble. When the plugin is turned on, it adds gain, but it also adds some better, how do you separate the louder from the better when they sound the same? That's it, I think that's what I'm trying to ask.

[Depone]

Bitch-slap me if im getting you wrong nowaysj, but...

The idea about matching the perceived volume levels is so that you can quickly a/b the sound to hear what positive or negative effects the said plugin is having on the audio.

For example, if your distorting a kick, I dial in the amount i want, then i turn down, or indeed up, the plugins output as to match the original perceived volume. This way you can hear whats happening clearly with a quick 'plugin bypass' click of a button or mouse.

Then do what you want, so long as your ears aren't deceiving you

[staticcast]

So just forget all that shit and just listen.

This is what it comes down to, and this is where I have my problem.

Just about every plugin I work with adds gain just by turning the fucker on. This should be illegal in my opinion. Swat teams should come crashing through skylights when a dev starts coding their shit to add gain like this.

So I instantiate a plugin, and now it's louder so it sounds better. Right here is where the trouble starts. First thing I have to do is turn down the plugin. So I've been using my peak meters to do this. Apparently wrong according to the benevolent (and austere) macc. Okay, so I'm supposed to reduce the gain, by ear, according to the perceived volume. But as was discussed:

Louder = Better
Better = Louder

where '=' means sounds like. If you take this to heart, there is trouble. When the plugin is turned on, it adds gain, but it also adds some better, how do you separate the louder from the better when they sound the same? That's it, I think that's what I'm trying to ask.

Lol, ok, I get you, but I don't really agree that louder=better works the other way around. But anyway. Like a lot of other related advice in this vein, it's meant as a guide, not an absolute - just make sure when you turn on your compressor it doesn't immediately jump out as being MUCH LOUDER or WAY QUIETER than when bypassed.

[HALP]

i dont really get that

if i put on a bit of saturation or something, or another effect, say chorus and change the dry wet
there's a difference in volume. I notice the difference in the overall mix, and adjust

end of story

[Phigure]

So just forget all that shit and just listen.

This is what it comes down to, and this is where I have my problem.

Just about every plugin I work with adds gain just by turning the fucker on. This should be illegal in my opinion. Swat teams should come crashing through skylights when a dev starts coding their shit to add gain like this.

So I instantiate a plugin, and now it's louder so it sounds better. Right here is where the trouble starts. First thing I have to do is turn down the plugin. So I've been using my peak meters to do this. Apparently wrong according to the benevolent (and austere) macc. Okay, so I'm supposed to reduce the gain, by ear, according to the perceived volume. But as was discussed:

Louder = Better
Better = Louder

where '=' means sounds like. If you take this to heart, there is trouble. When the plugin is turned on, it adds gain, but it also adds some better, how do you separate the louder from the better when they sound the same? That's it, I think that's what I'm trying to ask.

Lol, ok, I get you, but I don't really agree that louder=better works the other way around. But anyway. Like a lot of other related advice in this vein, it's meant as a guide, not an absolute - just make sure when you turn on your compressor it doesn't immediately jump out as being MUCH LOUDER or WAY QUIETER than when bypassed.

i think the better is psychological. a better sound will sound louder.

feel free to correct me if i'm wrong, i'm fucking clueless and really just saying what seems to make sense (to me)

[nowaysj]

just make sure when you turn on your compressor it doesn't immediately jump out as being MUCH LOUDER or WAY QUIETER than when bypassed.

Problem is, and I've read about this and experienced this myself, with subtle adjustments trained ears will still pick something that it is 0.5db louder as better! Basically our ears are sluts, they'll spread their cochlea for any old big sound that comes along.

When I check my levels in my meters, I feel that is a reliable way to separate the louder from the better.

[SunkLo]

Louder doesn't mean better, it just seems better.

[staticcast]

just make sure when you turn on your compressor it doesn't immediately jump out as being MUCH LOUDER or WAY QUIETER than when bypassed.

Problem is, and I've read about this and experienced this myself, with subtle adjustments trained ears will still pick something that it is 0.5db louder as better! Basically our ears are sluts, they'll spread their cochlea for any old big sound that comes along.

When I check my levels in my meters, I feel that is a reliable way to separate the louder from the better.

Ah man, firstly, just fuck the meters. Meters are useful for telling you:

- When you're clipping
- When your snare drum has a MASSIVE PEAK in it (even though it doesn't sound loud enough in the mix)
- When your sub is eating up your headroom
- etc

Big, high-level problems. That's what meters are useful for. Not comparing loudness, *especially* if you're using peak meters (as in most DAWs).

And secondly, if a 0.5dB difference in gain is more noticeable than the FX unit itself, it must be a pretty subtle effect. Crank dat shit up and stop being a pussy.

[Phigure]

And secondly, if a 0.5dB difference in gain is more noticeable than the FX unit itself, it must be a pretty subtle effect. Crank dat shit up and stop being a pussy.

best sounds are made no by turning one dial to 11, but by turning 11 dials to 1


or something like that

[SunkLo]

nono, it's by turning 11 dials to 11
Eddie Van Halen knows

[staticcast]

nono, it's by turning 11 dials to 11
Eddie Van Halen knows

I was about to post literally the same thing

[beatsbydrew]

Just registered to say that this board and this topic are golden. I've been reconsidering gain structure and my tracks are starting to sound waaay better. Cheers.

[xx773xx]

nono, it's by turning 11 dials to 11
Eddie Van Halen knows

Tap faster d00d.

[toot]

I'm about 8 pages into this thread and it's one of the most helpful things I've ever read.

Macc, I owe you a pint (actually I owe you a year's supply of pints but times are hard )

Seriously, thanks!!!

[bilsner]

holy fooooooooooook just the first page of this made shiznit sound far better, how the hell didn't I know this before /smacks head on desk

nice one DSF

[macc]

Macc, I owe you a pint (actually I owe you a year's supply of pints but times are hard )

I can wait

[bilsner]

Here's a thing, when sorting out the gains on my drums, starting with the kick and snare, I noticed that the snare needs considerably more db to sit in the mix than the kick, I think around 6 db which seems alot. For some reason I am assuming they should be closer than that altho my ears are telling me different. What d'ya reckon ?