Hi,
I try to make a good gain staging now that I saw some youtube videos about this important setting. The problem is I’m not sure I understand everything.
I put a VU-meter plug (mvMeter2) on the bass track, I make the adjustments to have the signal not exceding -18db, then I look at the volume fader in the mix console and it says ‘-12’ (‘-12’ what? I’m not sure). On a synth track with a kind of clave percussive sound, I have the signal at -18db on the VU-meter but it says ‘-1,2’ on the volume fader.
I suspect there is something about the loudness, the peak, or something, but it’s confusing. Is there a simple way to explain that, in the cubase context? Keeping in mind that some plugs need a certain amount of volume (loudness?).
Also I tried with the pre gain on the console but if I have to gain +10 or more there is a lot of distortion and I’m not even at -18 on the VU-meter.
I’m not sure I will understand the answers, if there are any.
Very interesting, thank you. At the end it says “o dBVU is roughly the equivalent of -18 dBFS — a perfectly safe place to be” but 0db on the VU can be already too loud on the volume fader.
Maybe I should not adjust the volume with the VU. I saw some advice to keep the volume around -12 on the volume faders, maybe I will try to do that. Maybe my confusion is I don’t understand when people talk about keeping the tracks at -18 or -12, if it means with the DAWs exact value or with the VU-meter.
On the digital scale (track faders) you should not go above -10 dBFS when recording to keep some headroom. This is the peak signal.
We are no longer in the 90’s and you should really avoid setting your recording levels based on a VU meter. Back in the time we were using analog gear that had pleasant saturation and distortion when the VU meters were maxed out. This is not the case when working in digital and you should avoid digital clipping at all costs, and keep your eyes on the digital meters and not on the VUs.
VU meters are RMS and show you the average level for the last 400 ms or whatever duration you can configure.
That’s why when you record percussive material, the VU meter won’t show a high value because the transients are very short, but on the digital scale it could eventually go near or above 0.
Try using the TestGenerator plugin with a sine wave at -18 dBFS. The track meter will display -18 dBFS, but on the VU meter it will show 0 (it depends on the calibrated value). This is true because the signal is constant.
If you do the same experiment with a drum track and the signal is hot, the track meter will show let’s say -7 dBFS, but the VU meter could show -13 dBu. It’s perfectly normal since the RMS level of a snare track is much lower than guitars or bass.
To reinforce what Louis_R told you: the affirmation about the level equivalence is taking into account a test signal, not your real instrument signal.
For gain staging I recommend you to simply use the meters next to the faders. These are peak meters. If you try that they reach average “eye”wise -18 dBFS, peaking at -10 or -12 or near, that doesn’t mean the rms or LU value is -18 (only average with your eyes), but that should be enough for gain staging. It is not necessary that you do exact adjustments, it is only a rough starting point that will help you to be at the safe side, without burning the master, but you shouldn’t be NASA precise or use too much time setting this up.
“gain staging” became a popular word not too long ago and really all it means is just adjusting the level as you go - if needed - so that you have a good level at each stage of processing the signal. That’s really there is to it. People tend to overthink this in general.
Don’t clip the signal. Why? Because we typically have to deal with fixed point processing at some point and those numbers can clip. So best practice is to keep signal away from 0dBFS. If you use a True Peak meter then just don’t clip.
If you have a plugin that emulates analog including non-linear behaviors then read its manual to see what levels you should send into that plugin.
And then as you add instruments, or signals, the sum will increase. So in order to work more conveniently you can have your tracks at a lower level individually than what you would have if you only had the one track, but since you know you’ll be summing a bunch of them if you start lower there’s less likelihood of clipping the master.
See? Not so complicated.
Like Sibben wrote the track meters aren’t VU meters. You can set the ‘fallback’ time which is the time it takes for the meter to show a falling signal, but if memory serves me the ‘rise’ time is pretty much sample accurate, or very close to. In other words the meter is likely to register peak on the way up. VU is average. So within the time the VU is reading a higher level there are peaks that don’t show, but those peaks will show on the track meters.
Odd. I’m not sure it’s supposed to happen. Where in the signal chain did you see or hear that distortion?
Unfortunately a lot of people say a lot while skipping a lot of detail. The idea of “0dBVU = -18dBFS” isn’t as obvious or clear as some people make it sound. It actually depends.
Some plugins don’t have 0VU = -18dBFS
Some interfaces are not calibrated to 0VU = -18dBFS
So you’re right that it can be too loud for you if your master is at 0VU (using a VU meter) and you’re listening in your room. That’s going to be because you would have to figure out what the meter’s “0VU” is in “dBFS” in Cubase, and then you have to see in the interface manual what that amount of dBFS becomes once converted to analog: some interfaces convert -18dBFS to +4dBu which is 0VU, but others might have that reference be -16dBFS or -20dBFS.
In other words the electrical analog signal that goes to speakers often uses the same reference in different interfaces, but not necessarily.
And then you have to consider how loud you’ve turned up your speakers.
This is why that sentence you quoted isn’t really as obvious as it seems.
Thank you very much, all of you. I understand that I have to follow a simple rule, as Louis_R and Knopf said, not peaking more than -10 -12. I probably saw too much video on the topic, and not beeing a sound engineer I’m easily confused.
MattiasNYC, I tried to reproduce the distortion with the cubase pregain, but I can’t remember how I did that. I was probably wrong in a way. Concerning the synth clave that was too loud, I was talking about too loud on the volume fader, so the peak. But in fact if it’s at -18db (VU), it’s too loud in the mix. And I can see a lot of tracks that are at 0db -18 VU and peak at -10 - 12 on the fader. So maybe this numbers don’t come from nowhere.
It’s begin to be clearer now, I only have to avoid the clipping and manage with the levels in relation with the plugs (like bx_masterdesk, not so easy, it seems it only works with high level volume). And try to use my ears, which I sometimes forget…
Keep in mind that if your mix has a medium level (because there’s not a lot of tracks or for whatever reason), you can still increase the pre-gain on the output channel so Masterdesk or other mastering plugins are fed with the appropriate level.
The thing is that these days with 32/64-bit processing the volume within a DAW can be all over the place without clipping. So it is necessary for producers and plugin preset designers to have a common frame of reference for signal level in order to help users hit the “Sweet spot” volume at which a plugin was designed to work. It just so happens that keeping audio at or just under 0VU is a pretty good way of ensuring that your gain is in the sweet spot for many dynamics processors (which is ~ 18dB RMS). This is just a continuation of gain staging best practices that have existed in recording studios for many decades. Arguably not needed any more, but still helpful in order to save time and sanity.
If your audio barely registers on a VU meter then this is a good sign that its way too low. If your audio is always maxing the VU meter then this is a good sign that it may be much louder than the plugin designer was expecting. A preset designer could specify an RMS value (for example, there are some presets in Ozone 9 which say “this preset works best with audio between -20 and -12RMS”). But this is rarely done. But it is safe to assume that the user and designer are both working with sensible signal levels. And 0VU is a pretty sensible level .
Adonde, the only thing I would add is that we probably shouldn’t talk about “VU” as if it is a set value when in Cubase/Nuendo. A LUFS or RMS meter should measure the absolute level of the signal but some VU meters allow you to set the reference level to where you want it.
So by avoiding talking about “0VU” as a likely “sweet spot” we also avoid situations where someone has used a VU meter plugin but set the reference point somewhere other than -18dBFS or where someone doesn’t have access to a VU meter plugin.
I think it’s generally better to just use dBFS RMS instead. I think it is less potentially confusing.
%100 agree. RMS would be simpler and just as effective. A lot of software and hardware still have VU meters though. And there’s just something about the design of those analog style gauges that sort of helps people to understand that maxing out the meter is not what you’re supposed to be doing.
Adonde and MattiasNYC, thank you for your answers. I read and read, I have to explore, I try some things. The truth is that I have mixes that are just the mixes when creating the song, more instinctive than really mixing. And the sound is not good, some things are too loud, probably clipping, the kick is often weak, but after weeks of trying to mix all that with some mix engineer optic, I listen to the first raw mix, and there is life in it that have been lost. Not always, but maybe some relation between instruments where interesting, or the snare where much more clear, a synth was crispy, etc. Something have been lost, there is a balance to find, and that’s not easy, and that’s exciting after all. I guess I should try to understand Supervision and Span, because I see RMS and LUFS, True Peak Clipping, even Max Crest Factor, and this one sounds mysterious.
Once you are relatively comfortable with what has been said above, I think this tutorial by Crazy Dom can also help. The most important aspect is that, at least in my studio, I want the peak signals next to the sliders always hitting somewhere in the upper third part of the slider. Do this by adjusting the pre-gain. This way, because it’s logarithmic, you have much more control over dB’s than in the lower third of the volume slider.
Then later in the video, clip-based editing. I sometimes feel users totally forget about this, and this is something I always do before considering compression. I want to feed my compressor a solid edited signal so it can better achieve the basic objectives.
Yes, I saw this one, he always have a ton of information. And you’re right I have to be comfortable with all this, I think what I need to understand is the vocabulary around the db, peak, clipping, because I can’t understand really all the advice if I don’t know what it is really about. I once saw something about electrical energy in sound, because at one point the sound is transform in electricity, go to a magnet that will move in a certain way, and we should mix for this magnets, somehow. It seems obvious but I can easily forget obvious things. I learned a lot this year having Michael Jakson’s ‘Wanna be startin’ somethin’’ as a reference track. I worked on some of my tracks that were much louder in the Span plugin, especially in the bass, but when I compared it with MJ suddenly the energy was impressive, very clear, everything very loud but in a way that you just want to listen to louder because it’s pleasant. And when I came back to my track, especially on headphones, the bass had an unpleasant pressure, the charley was harsh, it was not really louder but I wanted to turn down the volume. And I’m sure it’s all related to gain staging.
Wow, I’m sure I won’t be able to understand everything in this one but it looks very interesting. And full of wisdom. Thank you. And there is a “Decibels for Dummies” chapter!
Hi, I came across this older thread and as nobody mentioned it so far there is some misunderstanding and technically wrong information in the posts above:
VU and RMS meters are not the same. VUs do not average anything (… levels). How could they in the analog domain? They are just too slow to react on short peaks. Which is a good thing considering a general visualization of perceived loudness.
Use whatever helps your personal workflow but don‘t expect a VU needle to behave exactly the same as your RMS metering.
It’s really simple, but as usual people try making it out to be.
just use pre-gain in Cubase and try and hit around -18dbfs and your good.
and it isn’t that important if you are not using hardware devices, some plugins likes to see lower volumes, but just be in the ball park.
It is not that simple. Speaking in general, not specifically about VU meters, in analog domain you can measure average level and rms level too. Average (mean value) you can measure with moving coil instruments and rms with moving iron instruments. That says the operation principle of both kinds and it is mathematically accurate. That is the reason, for example, that moving coil ampermeters are used to measure DC current and moving coil ampermeters are used for CA currents.
How exactly a VU meter works I don’t know, as I am at this time not very informed about it. My point is that in analog domain average and rms can be measured very well.
