gain staging. need help

[Mustang Martigan]

I have an Apollo interface. UA recommends an input level of -20dB coming into the unit and -18 going into their plugs, well all of 'em except a few. A lot of brands recommend -18.
Should you always being going in, and coming out of a plug at the same level? And let the effect do the talking. Or are there times when a level boost is also nessacary? And if yes, should this be the last plug in the chain? And if not, do you then add a trim to get back to -18?

Then I've seen videos that say your mix shouldn't exceed -5, to leave headroom for mastering. So then when does -18 become -5?

I've tried using a limiter, and setting the ceiling at -0.2, but this hurts my ears and sucks the life out of the material.

I'm always arriving at the 2bus at -18, and then adding a channel strip, with a big boost as final effect.. to get the levels to where they aren't too low compared to the music I listen to. It seems like the wrong move. I know that I'm doing something wrong, but what?

[MoreSpaceEcho]

it's late-ish so i make no promises of coherence here. but....

1. an average level of -20 going in is great. or even lower. i personally try and never have anything peaking over -10 when tracking.
1b. note the distinction between AVERAGE and PEAK levels above. this is probably where your -18 becomes -5? -18 is an average level, whereas when us mastering engineers talk about -5 or 6, we mean peak.

2. i wouldn't worry about -20 or -18 going into the plugs. nor would i worry about keeping the output of the plugs the same level as what goes into them. sometimes it makes sense to boost the signal going in, for example you might have a guitar di going into an amp sim. maybe it doesn't sound as aggressive as you want....turning up the input level will give you a different result than turning up the gain on the amp.

or, i still use a certain boutique, vintage digital de-esser....with some tracks i don't get enough action even if it's all the way up. so i'll boost the gain going in and just lower it the same amount going out, so the track stays the same level in the mix.

3. there's no reason to have your mix peaking right up at 0. treat -6 as your 0 when mixing.

4. don't try and make your unmastered mix as loud as a mastered ref track. turn the ref down for a fair comparison.

5. if you set up a calibrated monitoring level, you can forget about most of this gain staging stuff entirely and simply mix so it sounds good to you, and you'll know that you still have some headroom left.

6. setting up a calibrated monitoring level is easy: play pink noise at -20dbfs through your monitors. you've got an spl meter right? set it to C, slow. turn up your monitors till the spl meter reads 82-ish dB. mark that spot on your volume knob. you're done.

7. i've never really thought about what dB level any of my tracks are at (aside from keeping an eye on peak levels when tracking). maybe everyone else does? i've no idea. i might just be a caveman. but to me shit is either too loud, not loud enough, or about right. i say it all the time but just do what your ears tell you.

[Mustang Martigan]

6. setting up a calibrated monitoring level is easy: play pink noise at -20dbfs through your monitors. you've got an spl meter right? set it to C, slow. turn up your monitors till the spl meter reads 82-ish dB. mark that spot on your volume knob. you're done.

Thanks.

Nah, no spl meter, and even worse, no monitors. I just have a pair of Sennheiser hd280 pro headphones. I can't figure out if some of the stuff I have trouble with, other than gain staging, is cuz I'm stuck doing everyrhing via headphones. I feel like I can't hear what I'm trying to hear sometimes. Even worse than that, I no longer have any real amps, and/or a place to play them, and my real drums, for the time being. Everything is fake, whether it be VI or DI. This makes my heart sink, as I'm playing rock, but anyways..

I'll never have an acoustically sound environment for monitors. But I'm guessing that just by grabbing a fairly cheap, yet seemingly unobtainable, pair of lsr 305s (and apparently a spl meter) would drastically improve how I'm hearing things over the current headphone setup.

I just love music, especially writing, playing n mixing. Besides occasionally recording a friend's band, I'm just playing around with my ideas, trying to get them sounding to we're I'm satisfied.. which is possibly impossible.

[losthighway]

It might be kind of a "would you rather" question, but to me trying to treat a less than ideal room and mixing on some monitors makes more sense than only mixing on headphones. Yeah, room modes are a beast but a 5-6 pieces of 703 can be the difference between total chaos and a general picture with some likely question marks in the bass frequencies and maybe a couple other places. Some of those question mark frequencies in a room can be learned through trial and error. While far from ideal, there's a level of inaccuracy in many rooms where after putting in the hours and trying things on different systems you might say, "yeah this sounds a little scooped in the low mids", or "it's got a little boomyness down in the lows", (or perhaps the exact opposite depending on the room) "but I know once it leaves this little room it's going to be close".

But first I say treatment. Unless your room is a perfect cube, you can get it up to 80% and reference back to your headphones, which to me is preferable overall compared to only working with cans on all the time.

[ashcat_lt]

If you happened to have some piece of analog gear (a preamp or a mixer) between your source (microphone or whatever) and your interface, and you set it so that it operates in its nominal range (it's VU hovering around its own 0db mark), and set your interface for unity, you would most likely be getting about the best S/N you could out of the chain without distortion, and you would find that the signal hits your DAW averaging right around that -18dbfs level. That doesn't make -18 any magic number, it's just the way it falls out. But it's the noise vs distortion thing that's actually important. If you're plugging straight into your interface, and adjusting gain with its own knobs, then you should ideal set it so that you get as far above the noise floor as you can without clipping and not even worry about the numbers.

Once it's "in the box", the level is pretty much arbitrary. The headroom inside the mixer of any decent DAW is just completely absurd. You can't really pick up noise or distortion without deliberately adding them. Some plugins do add those things for one reason or another, and for those things obviously the level going into them matters. Most of those, though, have ways to adjust the input and output level right on them. What happens in between - before and after - again doesn't much matter. Things like basic EQs, delays, reverbs, should be completely linear (amplitude-wise) and will do exactly the same thing no matter how loud or soft the signal going through them. Things like compressors, limiters, saturators etc are non-linear on purpose, and of course depend on the level of the audio they're processing. Then there's "analog" plugins that might model the way that a piece of analog gear - even something that would otherwise be linear like EQ or delay - fails to actually be linear and adds noise or distortion. For those, it might not be immediately obvious that they react to the level in a non-linear way. But, you know, why would you be using it if you weren't looking for distortion? Crank it up, see what happens!

Ultimately, there is no "right way" to do it. Turn the knobs til it sounds good. Try not to clip your ADC because you can't easily remove that distortion later. Be careful not to clip your DAC because that will obscure your perception of what is actually happening in the mix. In between the two, you can do whatever the hell you want.


As for "leaving headroom" for the ME: In the digital realm, it's just not a thing. This is common "advice" based on a misunderstanding of the situation. Ideally, you would send a floating-point file to your ME and then it just literally would not matter where your peaks hit. Too quiet? They'll turn it up. Too loud? They'll turn it down. Even with a fixed bit file, as long as it's not way quiet or actually clipping, it's the same thing. If your file peaks at -1dbfs, and the ME really needs 6db of room to work, they just turn the thing down 5db. If they throw a fit over having to do that because you didn't, find a new ME.

What people actually mean when they say "leave headroom" is that you should leave for your ME some Dynamic Range to work with. Don't go smashing it into compression and limiting trying to get it to final/competitive level. That kind of thing is very difficult to undo effectively, and it's generally assumed that your ME will be able to do this better, cleaner, and more appropriately than you might.

[vvv]

So for me, I try to keep unity gain between my analog chains (pre into compressor, etc.) and I try to keep incoming signals peaking less than -6dB into the computer.

Thereafter, my practice is to use EQ, limiters and compressors on most individual tracks to get them with a steady, usable gain level such that I am raising or lowering the gain on a given tracks at a given point for mix/drama/emotional purposes, not to make-up levels or lower the too loud.(Quick example: in mixing I want the rhythm guitar under the chorus - where I want it loudest - peaking at say -6dB, but lower under the verse at say -9dB. I want to do that with automation on the DAW immediately, not by ear verse by verse ... at least not at first.)

As I mix, I try to keep the entire mix peaking at -3dB on the loudest parts - I do not mix into a compressor or limiter.

When I have a raw two-mix, then I apply whatever processing, including EQ (pass filters only, if I've done it right) compression (usually none) and finally hard limiting (to get a -.3dB peak level).

I'm both prolific and broke and I further ascribe to the DIY ethic (mostly because I am both prolific and broke) so I do what I call "masturding" at that point; if necessary I lower the level of the entire track to apply whatever mastering plugs I am using (I been liking Izotope OZone Elements for finalizing).

[MoreSpaceEcho]

As for "leaving headroom" for the ME: In the digital realm, it's just not a thing. This is common "advice" based on a misunderstanding of the situation. Ideally, you would send a floating-point file to your ME and then it just literally would not matter where your peaks hit. Too quiet? They'll turn it up. Too loud? They'll turn it down. Even with a fixed bit file, as long as it's not way quiet or actually clipping, it's the same thing. If your file peaks at -1dbfs, and the ME really needs 6db of room to work, they just turn the thing down 5db. If they throw a fit over having to do that because you didn't, find a new ME.

What people actually mean when they say "leave headroom" is that you should leave for your ME some Dynamic Range to work with. Don't go smashing it into compression and limiting trying to get it to final/competitive level. That kind of thing is very difficult to undo effectively, and it's generally assumed that your ME will be able to do this better, cleaner, and more appropriately than you might.

this is all correct.

i turn things up or down all the time, it's fine. at 24 bits, you could be peaking at -48dbfs and still have a 16 bit file. and you'd really have to work hard to get a mix peaking THAT low. although i occasionally get things peaking at like -30 and i'm scratching my head wondering how anyone ends up that low. but you turn them up and they're usually fine.

saving mixes at 32 bit float is nice because if there are any peaks over 0dbfs, simply turning down the file will remove any clipping completely, which isn't the case for 24 or 16 bit fixed.

[KennyLusk]

So much great advice has been given here. Great reading.

FWIW I feel there's a big issue with not being able to tweak your mix on monitors/speakers.

Headphones are a great way to get started on a mix and set up some of the more time consuming mix tricks, but ultimately nothing translates well straight from a mix done on phones.