Gain Staging Structure: Waste of time?

[padillac]

What's wrong with that?

Oh, ummm. Well, let's see.

2. When digital goes into the red, it sounds like shit

Yeaaaaah, so that would be the wrong part. With a 32bit float signal path, until you actually output that stream to a converter/analog output stage, there is essentially no such thing as "headroom". You can clip 0dbfs all day long on any individual track and as long as you bring the master gain back down to where the actual output to the card doesn't see red there was effectively never any red.

You know, I had always heard this (especially about ableton), which is why it surprised me when I recorded a new audio track in the red and got nasty digital distortion. That surprised me based on everything that I had heard. And it shouldn't surprise me, because that's the way distortion works.

So it's true that I can take a -1db signal, add 2db of gain to it, reduce the master and not have clipping. But if I record a new clip too loud, it distorts, with no way of getting the distortion out of it no matter how low I turn everything at that point.

Three exceptions (that I can think of) to the "You can clip 0dbfs all day long on any individual track and as long as you bring the master gain back down" statement:

1) going over the "invisible headroom" limit of your DAW (+60db in ableton, from what I've read)
2) clipping a plugin that doesn't process bits with the same rules as your DAW
3) recording over 0db into any individual track

Those are reason enough for me to not clip individual tracks...and hopefully reason enough to temper your advice to people a bit, as well as to stop accusing people of not knowing what they're talking about because they advise you to not clip your DAW.

[padillac]

3. Recording close to 0db in digital means that a bit of processing can easily send the signal into the red, making it sound like shit.

It's funny, because looking at this "point" again just on it's own exposes how little time you've probably ever taken to understand what actually happens when applying processes to an already digitized audio stream with most any modern DAW. Code is a funny thing my friend. Seriously, check into that stuff... it's very enlightening.

Record something near the red, process it to go over 0db, now you're clipping the master and it sounds like shit.

Just turn the master down? Sure. But that doesn't change the fact that you can avoid having to turn the master down in the first place by simply recording it lower.

[Xenobt]

Odd you can find a personal attack in everything I say, rifftrax, you must have been a very slippery baby.

I made no sleight of FL users, just that a lot of crushed tracks I get are from it. And you left out my blind hatred of Reason, you're slipping.

I'm not lording that I make a living doing what I do over you or anyone else, but there is a level of dues that comes with that territory that deserves some respect, however grudging. And I didn't present myself as an authority, you did. An experienced professional, yes, Big Brother, no.

And why trash an owner because they might not know a VERY little used mic in real studio world? Ask the chief engineer. A lot of people call them PZMs, and in twenty years I've never seen one used anywhere I've worked. I tested one on the wall near the kit since I read they made a great drum room mic, but it had pretty high self-noise, and the transients weren't great. So I know that too! (BTW, a 414B in omni mode with a 10db pad was perfect!)

As for the rest, you can work crazy hot or low, but why? The workflow issue isn't a small one when you're spending your client's money (or even your own time) for every extra step it takes. At the end of a session, I can drop everything on their hard drive and it can go anywhere 24 bit and up.

But I said it a week ago, and I'll say it again, I get how 32 bit float works, and if you stay ITB in your own set-up, do it however you like. But track levels over 0db don't work well with gain based insert plugs, OTB processors, NARAS recommended practices, or final mastering.

So to the OP's question, that's a yes!

"When I was a boy of 14, my father was so ignorant I could hardly stand to have the old man around. But when I got to be 21, I was astonished at how much the old man had learned in seven years." - Mark Twain

[rifftrax]

And why trash an owner because they might not know a VERY little used mic in real studio world? Ask the chief engineer. A lot of people call them PZMs, and in twenty years I've never seen one used anywhere I've worked. I tested one on the wall near the kit since I read they made a great drum room mic, but it had pretty high self-noise, and the transients weren't great. So I know that too! (BTW, a 414B in omni mode with a 10db pad was perfect!)

It's not that the studio owner in my example was a bad guy. He was a really cool cat and super nice. Sold me a big mix console at the time. What I'm saying is that all too often I run into authority figures who present information in a way that's relatively very misleading (whether through some type of bias or feeling that knowledge in one field [say mix engineering] give you inordinate insight into a pseudo-related area [say, computers]) and often it bolsters the arguments of other people in a position of "authority" and pretty soon you have some crazy B.S. that's propagated and it just gets absolutely out of control. Most every time I've seen this happen it's got some root cause with an argument by authority, which is a logical fallacy.

It's just kind of annoying to see that go on so much. If you're going to make an argument then make it based on facts, not position and general community feeling.

I don't see anything that you've mentioned as a personal attack. I know it's a general "attitude" that exists against people who happen to admit to using a program that a bunch of 14 years old use. If I had said I used Pro Tools and then proceeded to make the exact same arguement, doubtless any professional who also used pro tools would respond very differently. This is an issue of what's called in-group bias.

[DuX]

I don't care about all of you using gain staging or not but in my case it sounds better and I'll stick to it.

Keep flat those peak meters guys, please. Your music will sound brilliant. I have no doubts about that.

Just my little 0.02c

Seriously speaking, you should all just use ears, and if it doesn't sound good - it is not. Then use gain staging as it will help with mixing immensely, I can guarantee you that. I come from the hardware background and mixing on consoles, and I stick to the same principle in a DAW, and it has never failed me. OTOH, if you like the sound of what you're doing now, regardless the gain staging, then don't worry about it as that's how you like it and want it to be. Why worry then?

The only thing is... if you're going to send it to a mastering engineer, please keep the track loudness to around -6dBFS and no limiters on the master channel, only if you really like the sound of it that way, but then there's not much a mastering engineer can do for you anyway.

Cheers!

[aciddose]

the thing about anyone who decides to have an argument about which levels they use, is this...

they're obviously complete know-nothing amateurs.

anyone who does actually know anything would know that you need to use the right level for the task at hand.

[Xenobt]

anyone who does actually know anything would know that you need to use the right level for the task at hand.

Not as true as it used to be, especially where drums and guitars are concerned. If you tried to print overheads or cymbals to analog tape hotter than -15 db, it got spitty, gritty and damn near unusable. Now I can take my overheads up to the snare hit level, no harm, no foul.

Just the opposite for electric guitars, basses, kick and snare. You took it right past 0db to break-up point, which varied with each formula of tape you used and the recorder speed. Since 15 ips moved slower than 30 ips, it saturated faster, and required lower levels. Now you saturate with a plug or on the front end with hardware.

And tape compression was a beautiful thing. But it took a board with CRAZY headroom to make it work well, and since tape bleed and mixer crosstalk came into play, that was tricky too. "Ghosts" of adjacent tracks could make trouble if you muted a part tracked too hot. I sure don't miss that at ALL!

So it's not that I'm a zealot about this issue, it's just for nearly a hundred years there used to be pretty set rules about what worked and didn't (and why), and many DAWs are based on a recorder/mixer model. Except for Live which uses a GUI and workflow from a mothership somewhere! I'm trying to take the best from the new tools and techniques, while keeping all that hard won (and still important) history in mind.

Can you see why the "it doesn't matter" stance might seem a little naive to folks who spent many hours seeking the grail of good gain at every stage of the process from mic to master?

KVR/eSoundz: Xenobt

[Compyfox]

a lot of mixes specially from the 60's 70's are squashed too.
there is mostly much ore headroom but i think the listener like the sound and that's the reason why producers often try to "replicate" the analog behave.

While these mixes where indeed squashed to a certain extend (for sound design purposes), they still had their transients - not with modern productions that are driven towards the 0dB finish line.

The transients are what's important in a production. Take them away, and your song/mix/production/final product sounds liveless.

soundwise,think about jimmy hendrix playing a clean set without any pumping,i guess it doesn't work, that's why "mastering" is an art.

Mastering is just the icing on the cake. You can't polish a turd in mastering, even though the Mythbusters "busted" that one in one of their shows (different topic!!!). The better the production pre-mastering, the more you can make out of it.

Again, take a look at . It clearly shows why a proper mixing process (and gain staging) is important on the long run.

the "record industry" is creating rules only to break them in in every single release?
look at the festivals,any electronic musician is breaking the rules.

So do sound designers, but they still stick to the 0dBFS limit. We're not in Hardcore Gabba or Industrial realm from the late 80ies/mid 90ies anymore.

as for the radio.
i think the radio stations are doing it wrong.

Do you know how a radio station actually works?
If not, here's a quick rundown.

You have several stages of "sound editing" until the stream hits the listener.

1) there is the actual studio with the DJ - his console is usually connected to a compressor to have a consistent level for both his voice and the material he plays

2) the studio is then connected to another section (sub-studio, equipment, whatever) that prepares the material for broadcast. This can be additional EQ, compression, stereo field limiting, etc.

3) lastly there is the broadcast station. To have a low noise floor and high output level to reach as many people as possible at far away areas, the signal is boosted (compressed) even more.


The endresult is what you can hear on the radio. Especially loud radio stations, or sometimes with radio stations "crosstalking" into others. If you're still on analog terestrial that is. DAB (digital audio broadcast, terestrial in ths case again) is better in terms of frequency splitting and loudness, but it still needs a certain worklevel so the same old rules apply - since nobody really improved on that end.

Digital broadcast (cable/satellite) lies on a whole different ballpark as well, since here the limitations are bound to the MPEG Layer 2 (audio) standards, or AC3/DTS.


It is not the fault of the radio station alone. The evolution can't go above and beyond, you have to be backwards compatible in this case as well. It's a two edged blade actually: the record industry wants their stuff loud, so do the radio stations. This can't work on the long run.

so gain staging comes into account in both cases,with smashed waveforms and without and it has nothing to do what the end result will be.

If we talk about pure sound design, yes. If we talk about proper mixing/mastering, then I can not agree.

So it's true that I can take a -1db signal, add 2db of gain to it, reduce the master and not have clipping. But if I record a new clip too loud, it distorts, with no way of getting the distortion out of it no matter how low I turn everything at that point.

The ADC/DAC has a limit of 0dBFS, some are even as low as -1dB (or lower). So your best bet is to not(!) record higher than -3dBFS regardless of what you might have learned at a school or something. "Record as hot as possible" is definitely not true.

If we talk "in the box", you can pretty much take a -1dB signal, add the desired gain and compensate it at a summing bus or sub group to your liking. This is the "internal floating point mathematics" we're talking about. It is however not wise to do so.

So it's not that I'm a zealot about this issue, it's just for nearly a hundred years there used to be pretty set rules about what worked and didn't (and why), and many DAWs are based on a recorder/mixer model. Except for Live which uses a GUI and workflow from a mothership somewhere! I'm trying to take the best from the new tools and techniques, while keeping all that hard won (and still important) history in mind.

Can you see why the "it doesn't matter" stance might seem a little naive to folks who spent many hours seeking the grail of good gain at every stage of the process from mic to master?

KVR/eSoundz: Xenobt

This is by far the best post in this thread, and I can wholeheartly agree with every written word. Thanks Xenobt - I'm no alone after all.

[Ah Xoc Kin]

However, the thing that has been really striking for me is the power of dogma. I completely believed that when something was in the red, it was clipping. This was to the point where I actually believed that I could hear clipping when something was in the red. It turned out that dogma had actually messed with my head. When there's digital clipping on a track, you know about it!

From Roey Izhaki's book, Mixing Audio:

"Only clipping on the master track indicates clipping distortion. All
other individual tracks cannot clip."

I highly recommend that book.

[zerocrossing]

Sorry, I have a philosophy degree so I can't resist saying this. Socrates should probably be our best friend when approaching the discipline of audio production. Although there have been plenty of people after Socrates that said the same thing (and probably people before him), what he said was very important! Always take time to challenge your own beliefs! Especially the most fundamental ones. Wars and the most unimaginable cruelty are almost always caused by people who don't take enough time to question whether or not their own fundamental beliefs are based on anything over and above dogma.

As a philosophy student you should know better than to confuse beliefs with facts. Just because you don't understand the importants of gain staging and when and where it's important, doesn't mean it's something that can or should be ignored. What you're saying is the equivalent of saying, "I've got airbags and crumple zones in my new car so why should I be concerned about collisions?"

As others have said, compression and other factors are important. Signal to noise ratios still come into play, especially when using mics and older analog electronics/instruments.