My stereo out is clipping even though all channels playing at -10dbfs!

What setting/metering do you mean when you say “-18”? The pre-gain? The peak of a single channel? The headroom of the master channel?
In general, I agree that it doesn’t hurt (in terms of audio quality) to lower the gain more then necessary, but because he wants to learn something here and has some mentor who told him what to do, I think it makes sense to work on some “good habits”.

-18 on the meter for each individual track. If you read what he wrote he’s trying to adjust the pre-gain to get that meter to read negative 10 which is too high. He should be adjusting the pre-gain to get the meter to read negative 18 average. This will give him the headroom he needs through the mixing process. It’s common practice, it’s not an opinion. It’s digital and in the digital world negative 18 is like using zero on an analog meter. It’s not going to sound any different if it’s lower or higher, it’s just going to change volume. The solution turn up the monitors or headphones. You lose sound under negative 36 and obviously when it clips, but the stage he’s at and a digital world negative 18 is the starting point.

Dbfs meter x

Wow I cant believe this is still a thing; mix engine > 32bit means you can do anything you want…the master buss WILL NOT CLIP while you are working in 32bit space (almost impossible)
-18 is a good general standard per track and whether you are Brauering or direct routing to mix buss, its only of significance in the analog world OR <32 bit where 0DBFS is critical

Cubase is of course 64bit since 9.5 so this is all just academic; just get sensible mix levels at the output and run a pre gain to the master as mentioned in some of the other posts…stop sweating the irrelevant stuff and focus on the art…which is so much easier these days. especially we now have 32 bit hardware like eg ur c and other 32 bit capable

Its just maths…not black art.
To prove the point, anyone here, post a clip of 32 bit file ‘gain staged’ for loudness and one without it and Ill prove it…64 bits is ridiculous headroom. You can see this all graphically by using 32 bit colour space and then compressing back to 8 bit output…it all makes sense immediately.

Constraining HDR per track is however altogether a different thing which actually does affect loudness…but thats another story…this is akin to shadows/highlights which is calibrated to intended output; ideally K-14 being a great foundation.

PROVISO: This assumes
All plugins in the path are 32bit
All Source track material is adhering to basic standard of -18 (a solid standard for good reason)

That’s not really entirely ‘correct’. If -18dBFS peak is what you’re talking about then it doesn’t really correspond. Looking at a VU meter that’s at zero would correspond to an average of -18dBFS, not peak.

So as far as I see it if nominal operating level in the analog domain is 0VU and that’s a comfortable listening level when monitoring (speakers) are set to calibrated reference then it makes more sense to leave a track so that average = -18dBFS, not peak. And through the process of actually mixing and applying effects the dynamic range of the signal going into the track’s path usually gets reduced by the end of it all using compression etc.

But still good practice to keep the master output below 0dBFS since any playback will take place using fixed point signals which do clip. I mean, at some point we’re all rendering out a file for someone to use. Granted, a mastering engineer can turn it down, but it just seems reasonable to not deliver things above zero… to me at least.

We’re not talking about peaks for talking about averages. Read the thread

Plus the fact that a vu meter is useless in the digital realm. Unless you want to sit there and watch it tick all day

Absolutely…i took that as understood as part of the standard ie calibration is using a test tone which is not peak

Agreed; thats what the pre gain is handy for. Im def not saying to ignore everything thats for sure. My hardware output is 32 bit so I dont particularly worry too much…it does creep the levels up though in the speakers…but usually pretty good unless its a huge project which I have turned minimal these days…If I cant do it in an 8x8 matrix of tracks…I have failed :slight_smile: Less is always more!

1 Like

Who’s “we”?

THE OP brought it up, talking about “clipping” implies it, bns23 mentioned it, and you implied it by responding to someone talking about “peak” and not making clear you weren’t.

So, it seems people were talking about setting levels so that peaks were at or below -10dBFS or -18dBFS, not average levels.

I wasn’t talking about VU meters in the digital realm.

What’s that matter You’re not adding anything useful to the conversation.

Then put me on ignore.

Yep sorry…to qualify, calibration is always average not peak: lost context with OP
Additionally the source can be crazy of course. eg Egyptian Dhoumbek into fast mic eg 12mm sdc and fast pre ie tranny…the peaks are so super wild

1 Like

Oh he was useful…, he mentioned the analog way of looking to gain stageing and level measurement. These things are still valid as so much people refer to the wrong terms and methods.

This is not really true, since it depends on what you’re trying to measure. It depends on your personal preference. Harrissons Mixbus has them in every bus by default.

To make some things clearer…

A digital meter reading -18dB FS means in the analog world nothing.
It depends on the maximum level that the analog device can handle and the reference level of the used converters.

Most of them can handle levels around 22 to 26dBu, but some are overloaded with 20dBu already.
And the 0dB point is different on many analog devices… it could be at -10dBV, 4dBu, 6dBu.
Even the headroom provided is not the same on all devices…
But overloading an analog device will lead to distortion
that isn’t that hard as a digital clipping
and to drive analog devices over their nominal level
is common practice to achieve specific sounding signals (Tape saturation, guitar overdrive…)

Ok, now back to the digital world…

Thanks All.

Well I adjusted the gain staging to -18dbfs on each channel and now the master dbfs peaks at -4dbfs… could maybe use a little more headroom but can always adjust.

So that’s that problem sorted!

Thanks all.

1 Like

Hi Mattias! I will never put you to ignore😄 You are valueble to have. So. Where can I read about Nuendos Trackmeters? There are two scales… the one to the left is dbfs? Bou the one to the right?

This one?:

The left is the offset of the incoming signal that the fader is creating measured in dB. I think technically it’s probably correct to not use “dBFS” since you’re offsetting a signal and “dB” is a ratio, plus floating point allows you to go over “Full Scale”… so just “dB”.

The right is the signal’s actual level shown as dBFS. If memory serves me correctly the right meter will show peak when the signal is rising but you can set the fallback time to something else, closer to VU behavior for example.

So in the image the fader is attenuating the signal by 6.8dB (left), and the output has peaked at -5.6dBFS (right).

If it is the exact same sound/sample on every track then it will go crazy and peak the masterbus and you will have to pull them all down to -30 or something.

If it’s all different sounds on every track then it should not do that so much but then I would not mix like that.

I do usually adjust my input gain, or direct offline process the gain down, so the track starts out between -18 and -12 (I will do more if it’s always going to be a quiet element of the mix). That gives me the lots of fader room to work with (it’s best to work near zero on the fader because it’s calibrated differently the further down you go so it becomes more difficult to make small moves lower down the fader) but then the tracks end up at all different levels after that, all below or on -12, because I’m going to balance the mix with the faders so your never going to have 25 tracks all at -10.

I dare say your trying to get to the same point I usually am with each track on or below -12 to -18 and the master bus finishing below -6. :grinning:

Hi, I knew you were the right person to ask!
But… I really do not understand the meaning or reading …
I just recently got into this thing about gainstaging, nothing I have had any issues whith and I , when recording allways had the zero ( 0 ) in left column as a target… now, when I noticed that there is a different scale in the right column I was a little confused… and now you are saying that the figure 0 to the left not correspond to -18 vu?.. just for the sake of it… if I, when mixing whant every track to hoover around -15 -18dB … where on these two scales should I read?

That’s right. The scale directly next to the fader - not the one ‘under’ the meter - indicates how you are changing the signal. Not what the absolute value of the signal is.

On the right one.

Look at the following picture which has a signal generator set to -18dBFS (sine wave) on track 1, which goes to track 2, which goes to track 3, which goes to stereo out:

(metering is post-panner)

The signal generator sends -18dBFS to the fader of the track.
The fader sits a “zero”, which is “unity gain”, i.e. no change.
Therefore the meter shows -18dBFS.

On track 2 the fader has been lowered to -6dB.
Therefore the meter shows -18 -6 = -24dBFS.

On track 3 the fader is again at unity, zero gain, “0dB”.
Therefore the meter shows -24dBFS since there has been no change.

On stereo out the fader is set to -6dB again.
Therefore the meter shows -24 -6 = -30dBFS.

Makes sense?

1 Like

Yes! Finally it makes sense :grinning:
This was the explaining I needed, I am very grateful.
Thanks for your effort.

1 Like