Hey everyone,

so I've been reading and watching videos about gain staging a lot, but I still have lot's of questions and found no or conflicting information. So i decided to ask you guys.

Here is what I know:

- When mixing, you should keep your insert levels of vst instruments and recordings at around -18dbfs.

What I don't know is:

1. Do you have to look at the PEAK levels or at the RMS levels, when gain staging to -18dbfs? When I aim for the Peak levels, the sound is really really quiet, when I aim for the RMS, sometimes I get clipping at the peaks. So, which is t? If it is the RMS, how do I calculate this? My cubase pro just shows me the peak levels =(.

2. When recording guitars or vocals, do you have to aim for -18dbfs while recording, or is it wiser to record guitars louder (-6bdfs at peaks) and then reducing the signal at the mixing stage.

3. How do you go about gain staging drums in superior drummer? Should every single drum hit -18 dbfs? Should just the loudest Drum hit -18dbfs and the other pieces of the drum kit should be calmer? Should the entire Drumkit also hit -18dbfs?

4. Should and instrument Bus (e.g. 4 guitars, or all the vocals) also hit -18dbfs?

5. Should the Mix Bus hit -18 dbfs?


OK, I think this is all for now.

Thanks for your help!

Well, it's convenient to calibrate your system so that it interfaces well with any analog equipment you connect to it. Part of that can lead to you ending up around -18dBFS average levels. If plugins emulate analog and have a 'nominal operating level' of "0", chances are that that's going to be around -18dBFS, and that beyond that level the plugin starts to distort on purpose. It's easy enough to figure that out by reading the manual for the plugin or just pushing input levels.

Anyway, "should" is a strong word in this case. It won't hurt, and it will probably help, but if you end up louder or lower it's not the end of the world.

You should be looking at average levels, not peak. Maximum is 0dBFS, so from -18 to 0 you have 18 dB of headroom. That's where those peaks live.

I was taught to set levels so that I got the best sounding signal recorded, and then after those stages adjust levels to balance instruments. So to me this means that you set recording levels in a way where some instruments will be too loud if you're just monitoring everything as-is, meaning monitoring the input - but - you then adjust the levels afterwards and balance the mix, even during recording. It's how you set cue mixes etc.

I'd say that you 'should' set all your analog gear properly so that it sounds nice, which probably equals about -18dBFS average levels, and then after it's been recorded I think the track you record on will for sure have its fader post-record. I believe it goes like this in Cubase:

Analog source - Converter - Input Channel - Hard Drive/SSD - Audio track.

So this means that if you adjust your input channel fader it will record differently onto your drive, which you don't want (because it's unnecessary), and if you adjust the fader on your audio track it will be after your signal has been recorded.

(also see below for signal flow in Cubase / Nuendo's channels)

Well, here's where it gets a bit confusing for people I think. Again, this all really comes from using analog equipment which has clear limits to how far you can push it. It's not really the same with a mixer such as Cubase's. So, you could really set levels any way you want considering that this is a virtual instrument. You're not going to introduce more noise or distortion or anything unless it's built-in on purpose.

I would just balance the drum kit so it sounds good as a whole, and then I'd set the level of it all to where it's convenient. It makes no sense to target each component of a drum kit to some arbitrary number. Make it musical. Now, if you recorded analog sources like guitars and vocals, and those ended up averaging -18dBFS, then it's probably going to be a good ballpark to be around the same levels. Just use your ears.

It's just a matter of signal flow and processing. If you sum all your vocals in a group/bus and it exceeds -18dBFS average then that's not a problem by itself. Cubase can handle that just fine since it has a lot of headroom (way more than 18dB). If you have a device on an insert on that group though then you should make sure that the plugin doesn't distort. If it doesn't you're good. If it does you can either turn down the input level in the plugin if possible, or turn down your instruments before they hit the group.

Then you just adjust the output by moving the group fader. Really, there's not much more to it than that.

Depends on what you have going on after it. How loud is your monitor setup? Are you using Control Room to adjust levels? I mean, generally you probably don't want a signal that low on your mix bus, given that most people compress all (or most) instruments before hitting that bus. You're not damaging anything by the level being low, but there's also no automatic damage by the level being high. The only thing is convenience, meaning that if it's a high signal then your speakers might give your ears some work (which you can adjust if you're using a monitor controller of course).

PS:

I forgot:

Cubase and Nuendo has the ability to set the fallback time of your track meters to a different value. If you go to 'preferences' and 'metering', you should see "Meter's Fallback", and its measured in dB/second. Looking at meters is to a large part just a matter of convenience and a way for you to perceive whether or not something is loud. So you could set colors and fallback level to whatever works for you. So you could for example have a green meter up to say -20dBFS, then yellow -20dBFS to -16dBFS, then red above -16dBFS. That way, with a convenient fallback time you would be fine with your signal in yellow, and when you're in the red you're not clipping, but you're limiting your headroom more than you want. Those are just example numbers of course. The point is just that you don't have to just look at the number in the meter which as you say gives you a peak value, you can change the meter coloring / location to reflect something else.

To then check it, I would open just get a free RMS meter plugin somewhere and measure levels while looking at the channel meters and make sure you perceive a connection between the two so the values make sense.

Follow those simple rules:

Individual channels: When you are recording, vst or live just stay 1/3 below the clip (0dbfs peaks). Very easy, now the important thing is not to clip the insert plugins. After the plugins just dont clip.

MasterBus: Example: you have all the channels don't clipping but the masterbus is +8 over the red clip: Just use the trim knob on Cubase, under the rack menu, and lower it to -6 or -3 peaks, it's not that important. Just leave some headroom.
End of all the saga on gain staging!

Stop depending on meters to tell you what the proper levels are, and use your ears. If you look to meters for guidance too much, your ears will never be properly trained, which ultimately, is what you want.

Cheers.

Might be hard to find the top third between zero and negative infinity....

Thanks so much for all the answers. I have tried several things now and I have more questions.

What I tried was: I got an RMS Meter to read the RMS levels and the peak levels of my tracks and tried too meter Superior Drummer.


1. I tried to get the peak levels of the loudest drum (the rack tom in my case) to -18dbfs and the other toms much quieter, so the original balance wouldn't get lost.

I then used a VU meter (calibrated to -18dbfs) to check the levels. The thing was that the needle on the VU meter didn't even move one bit, which can't be good. In thought it was supposed to hit 0.


2. I the tried to get the loudest tom to an RMS of -18dbfs and again setting all the other pieces of the drum set up accordingly.

The results:
- Now the needle of the VU meter is reching about -20 to -10, but still nowhere near 0.

- On the other hand, some drums clip with an RMS of -18dbfs (the Snare and rack tom).

- When I try to amplyfy superior drummer to the point that the VU meter needle reaches 0, everything sounds super distorted.


So in conclusion:

1. I don't know why the VU meter isn't hitting 0 when the RMS of my toms is -18dbfs.

2. I don't know how to work with -18dbfs, when someof the peaks cause clipping.

3. Should I just reduce the entire drum kit drastically and not worry that the VU meter needle doesn't move at all?


This is all very confusing to me...

Where was the VU meter placed?

I placed it as the first plugin on the inserts, bypassing all other plugins. Is that wrong?

As A sidenote, I also found an "internal" Cubase meter that I didn't find before. It seems to show me the "MAX RMS" and "MAX PEAK" of .. something. I don't know what, since it doesn't correspond with any other meter (see attachments).


EDIT: OK sorry, my bad. The VU meter reaches from -7 to -3 db, I don't know how I could misread that. So this is alright then, correct?

I actually agree with Jeff Hyatt above that it's typically better to just work than stare at meters. Just wanted to mention that.

But anyway....

You just have to consider in what order things happen. I know it's annoying to read a bunch of my ramblings and schematics etc, but if you place a plugin in an insert in Cubase it'll by default read the level before the fader. So whatever that VU meter reads it'll be before your panning and before your fader. Just wanted to make that clear.

Seems you're looking at the Control Room meter section which is the one I was mentioning before I think. If you send one source only, i.e. one track only to your master output, then that Control Room meter if it's set up properly should show you the output level. If it's different from other meters you again have to consider what type of meters those are and where they are in the signal chain. There isn't an easier way to answer I think I'm afraid. It's just engineering 101 - signal flow.

As long as you aren't clipping anything it's all good.

Yes that is clear, that's why I put it there. I just want to make sure that the signals coming from the source (Superior Drummer) aren't too hot for my plugins.



It is indeed the control room meter. Strange that it is there, because I have control room deactivated in Cubase. Thank's for pointing that out =).

I also agree, I like working intuitively a lot, but on the other hand, that's how I ended up clipping many many of my plugins on many of my mixes along the way. So looking at meters does give me some security. It's just the next step up in production quality for me, if you will =)

Hey guys
Amazing thread and please help me out understand some more, I'm getting crazy about this peak/rms/vu things
I'm still no audio engineer by no means even if I'm now mixing and mastering for friends (been making music since I was 7, teach basic music production at a local school) so sorry if I say something wrong
What I'm saying is generally about gain staging and this -18 dB thing so
1) when I import tracks or record, before even touching the fader, I have to check if the signal of the clip is at -18 to -12 (/-10) dB PEAK. But does the lowest (useful) part have to be in the -12/-18 range or THE MAXIMUM part in -12/-18 dB range?
2) Why do people use VU meters? Is it because VU meters put -18 dB as their 0 level?
3) If I use virtual VU meters, do I have to check them in RMS or PEAK mode?

Please help me out in this mess

Analog meters measure levels from the noise floor up, digital meters measure levels from the clipping point down.

Analog and digital are two totally different animals when it comes to levels.

Analog tends to have a sweet spot when it comes to recording levels. The signal needs to be above the noise floor to prevent hiss, below the maximum level where analog distortion is minimal and frequency response is maximized. With most preamps that range exists between 25% and 75% of the preamps maximum gain. Below 25% the noise floor masks the sound and above you start to get clipping.

In digital the peaks simply needs to be below 0db. There is no sweet spot in digital because there is no sound in digital. What you read as peaks are simply ones and zeroes triggering a graphic illusion. There are no sine waves when you mix, you're just mathematics/algorithms/ones and zeroes.

What happens when you change volume levels is the samples saved or being saved change mathematically. Instead of a sample being recognized at say 1 volt its recognized as a half a volt when the volume is cut. Its a proportional fraction to the original that can be divided down as small as needed.

Any noise floor that exists comes from the preamplifiers built into your interface or whatever you have plugged into that interface. The computer itself doesn't/shouldnt generate a noise floor once the analog signal has been converted to binary. There can be other digital artifacts but that's really a different topic.

main point of all this is when you're recording digital, there's no reason to get anywhere near the red or yellow zones on DAW meters. Most DAW's use peak meters, not average or RMS.

When you mix within a DAW you want to get the mixdown to have an RMS level of around -16 to -18dB or an average level of -12 to -14 so you have enough headroom to run your mastering tools. In most DAW programs that means you can run the mains meters so they peak around -2 dB give or take a couple of dB's
An occasional peak above 0dB is unlikely to cause any issue but there really isn't much reason to ever get that high.

The mains meters are the sum total of all your other tracks. If you have 3 tracks reading 33% of max, you should see the mains meter running higher then the individual tracks. it wont be 99% of course because you have different notes, at different frequencies, occurring at different times, plus there is allot of silence between notes. If however you had three low frequency tracks all playing the same notes you could easily see the mains meter exponentially higher then the individual meters.

How important is the level of recorded tracks? Not as much as you might think. There is no difference in frequency content so it isn't anything like analog recording where you had to find ideal levels to get just the right tone and saturation. None of that occurs (and if you think it does you need to educate yourself better).

What can be an issue when track levels are too high or low is how your plugins react. Code writers often use analog parameters when designing plugins. Some are wise enough to include the greater dynamic ranges of digital when designing those plugins, but some do not. You may in fact have to get your tracks to a specific level tracking for those plugins to work their best or produce the lowest level of noise.

I recorded analog for about 30 years before switching to digital in the late 90's. I've done enough recordings since then to find optimal recording levels which work best for me.

I generally found tracking guitars so the tracks wind up being around -10dB is ideal for most plugins. Tracked levels and recorded levels do wind up being different however. I may be reading close to zero on the recording meters but the tracks themselves can easily be 8~10db lower. Likely due to the frequency content.

Bass I try and target the -5dB range, Drum samples/drum machine tracks the -4 range and vocals in the -3 range.

I can bring all of these down in level but rarely have them higher simply because certain plugins like compressors or EQ's run out of headroom with the ceiling being right there above. I also realize this might be a bit too hot for some types of music but again, I used to record analog where you often worked with lower levels of headroom so I'm conditioned to working in cramped quarters like that. For someone new you could easily take all those levels down by 2~3dB if you want. You got up to 120 or so dB before you get to the noise floor so I wouldn't let a few dB bother you.

If anything be consistent when tracking. Try to avoid throwing pizza and hoping it sticks. If you do several recordings at say -10 and then when you use a plugin and select a stock preset and then you wind up having to boost the levels then maybe that track was tracked a bit weak. Try bringing the recording levels up in 1dB increments until you're no longer having to add makeup volume when using plugins.

Likewise if you insert a plugin and select a stock preset to say a compressor and it makes the sound blast from the speakers and runs the meters up into the red, then maybe that's a good indication that track was recorded too hot based on the way the track is driving the stock plugin presets.

What you ideally want is a track level that's just right so the plugins do exactly what they are supposed to without boosts or cuts needed. Then you'll have a range to work within to bring things up or down as needed.

Like I said, for the plugins I use I have targets I try and hit. They even change a bit with the different instruments I play and the type of music I'm playing. I can usually tell when I'm playing if the instrument is pumping too hot and simply back it down to give me a safety margin. If its something like a bass I often have to be conservative in my tracking levels because I like to dig into the strings to get aggressive bass tones. Other times I can leave it a bit higher playing stuff that's smoother.

That's something that simply comes with experience as a performer. There are no meters on an instrument telling you how hard to pick strings for a given song so there is always a certain amount of guesswork when it comes to tracking levels, but for most players they usually have a fairly consistent dynamic level. If you don't know what your limits are then simply see how far you can make the meters move when you dig in. If you slam the strings as hard as possible and stay below zero you're safe. The rest is simply what works for you and the music.

Most people talk about average levels when they mention -18dBFS, not peak levels. And you don't have to do it that way. Just stay away from clipping and be sufficiently above the noise floor. If you record 24-bit it's easy to accomplish both.

But anyway, yes, if you're recording then you want to set your analog equipment before the converter so that you average -18dBFS on the converter after conversion, if you choose to work this way. Again, not talking about peak. Adjusting a fader inside the DAW does not change the level into the converter or onto disk (unless you're in Cubase in which case you may have a separate input channel with a fader which indeed is before saving to hard drive - but no need to lower that level).

VU meters have a specific type of response, so looking at them sort of corresponds to how we hear things, even though more modern metering is more closely related to how we hear. I'd just use the DAW meters to check peaks, not a VU meter. That to me wouldn't make sense.

Thank you for you answer man!
Everything clear