Read this article from Bob Katz. It lays it out plain for almost all to comprehend 
Yeah I know about K system meters. Thanks, they’re very useful, but I was just asking you for clarification as to what specifically you were referring to, as it wasn’t entirely clear how and why you were doing things. I would have thought an average level of -18dBFS RMS should be OK for stems.
I personally shoot for peaks at -12 for my main mix, not for every individual track. I build my mixes in context as I go.
In context, a pad may end up being significantly lower, and printing a stem may result in a file with barely a visible blip. That’s why I shoot for a higher target.
You can’t set the channel meters to the K system. This applies to the master meters only.
If, as sustudio suggests, you aim for channel meter activity between around -18 to -12 (for the main body of the metering, NOT the extreme peaks), then yes, you will be aproaching something close to k-20… in other words the average RMS signal level will probably be around -22dBRMS to -18dBRMS (this is also highly dependent on the dynamic characteristics of the signal).
Remember that you cannot see RMS level on the channel meters (they are strictly peak ONLY). These are showing the average peak activity in the main body of the meter display with the fast transient peaks in the upper part of the meter display (the narrow peak hold bars).
Ive never understood gain staging to get a consistent workflow from track to track so I’ve had to find my own way out of the puzzle. I’m not even joking, I just use my ears. And I clip almost every group/buss too + certain individual channels without exception.
I guess anyone with a degree would say that’s poop/amateur/crap but it’s worked for me. Granted, I’ve not won a Grammy. Yet.
Why are we talking RMS levels ? Peak levels should be the only thing one should focus on when we are talking gain staging. If we are talking delivering for broadcast then EBU should be the focus.
This is incorrect, and the manual is incorrect. Easy to verify: run a 1k sine wave test tone through the system [also good to calibrate & see what’s happening with the summing, groups etc generally, set at -20 as per Katz for that use]
In VST devices, create a mono group set to no output;
Insert the Steinberg Test Generator plug on this group & select a 1k sine wave (max it out at 0.00db for this purpose);
Create a mono audio track with its input set to the tone group & its output set to the stereo output bus.
Set the mixer meter positon to Input; put the Audio track in input monitoring or record.
Now try to pull that Pre gain down on the audio track and it is clear this has no effect on the input level (only the master fader output).
Similarly, record that tone on the same audio track, playback, same results.
The PRE function does not effect input levels, is only a gain stage first in the output chain.
Poor implementation of that feature & its terminology.
Meanwhile, the capacity to see and use audio input faders works very well (unlike many other DAWs); if recording for headroom here, no problem. However in ‘gain staging’, this takes on at least three different functions and meanings in this context: i) new audio recordings as above, fine; ii) stems from elsewhere as too hot (use the region audio handle first up to drop the level & before strip silence or similar; iii) VIs too hot, drop the level at their master output.
No the PRE section is pre fader, now you are confusing the input gain with the pre gain. It is like the manual describes.
No offence intended towards you personally, but I would highly recommend re-training your ears to discern digital distortion. Red-lining a fader in a digital realm is distortion, plain and simple. It’s not like an analog console where you could push a channel strip well into the red before it would distort, depending on the specs of that console.
I would suggest trying out a cut where the tracks are printed with peaks at -12, and mix it and makeup the gain at the final stage of your main stereo bus. If everything’s too quiet while you’re working, turn your speakers up louder.
I’d be curious to hear your feedback once you’ve done this.
You’re absolutely right re: managing peak levels for proper gain staging. I usually eyeball RMS levels when tracking a lot of instruments and when building a music mix, as I like to keep an eye on the overall dynamic range dynamic range.
I edited my post to avoid confusion for other readers. Peaks are all that matter when recording tracks to DAW.
It really isn’t so complicated! 
Record at healthy levels, allowing adequate headroom. -10dBFS peak is a good rule of thumb.
Maintain the peak level such that that with the channel’s fader at unity the peak remains as recorded (e.g. at -10dBFS). In other words, if an EQ boost adds some gain, then trim the EQ’s output to get the peaks back down to -10dBFS. If a compressor reduces the peak, then trim the compressor’s make-up gain to bring the peaks back up to -10dBFS.
Also do the same through any Group Channels.
Errr… That’s it! 
@ The Elf
yes but we must know the quantum theory behind all of this otherwise how will we hack the mainframe
AFAIK this is entirely correct. Not trying to score points here but you may be misunderstanding the precise nature of the signal flow through a channel.
No, the PRE gain still has an effect on the level going into the channel, it’s just that you cannot see this in the meters because the meters are set to “input”. If you listen you can still hear the gain change.
Not sure what you mean by “output chain”. What we have is a channel input signal which is passed through the various stages of the channel as outlined above, and then there is the channel output. So No, the PRE function does effect the input level going into the channel. And Yes, the PRE trim is indeed a gain stage which comes first in the signal chain of an audio channel.
Finally, in my description in the last paragraph above I am talking about metering and only metering. When you choose to switch the metering of all the channel meters to “input” you are metering the channel input signal BEFORE it gets to the PRE section. This is why when you move the input trim there is no change on the meters. It’s as simple as that.
If, after this description, you still think it does not work like this I’d suggest you indicate it directly to Steinberg.
When using 32bit floating files as the default Cubase internal format, which are virtually impossible to clip, gain-staging isn’t as critical, except for where the DAW interfaces to fixed integer sources and targets, which usually means with:
a) Inputs to the ADCs of your audio hardware, usually 24bit integer.
b) Outputs to the DACs of your audio hardware, usually 24bit integer.
c) Rendering to fixed bit-depth audio formats, like for 16bit WAV files.
d) Inputs to some level-dependent FX like compressors, maximisers and ditherers.
e) Output from ditherers and maximisers, which should not be changed at all.
f) Outputs feeding broadcast equipment.
For these, gain levels need to be optimised to use the most of the bandwidth without clipping. At any other stage, gain staging helps the meters give a good idea of what the signal is doing at that point, because meters slamming into max scale or hardly registering do not help you to make finer level adjustments when required, or track problems.
Note that all these calibration ‘standards’ are rules-of-thumb generic setups, and do NOT guarantee that any particular signal will not go full scale. They just give some predictability, but which one you use will depend upon whether its required (as for broadcast), or what suits the genre(s) you do best.
However, if using fixed integer internal formats, gain staging is CRITICAL to avoiding clipping, as every audio track is a potential overload point when it is being recorded.
Note that Cubase uses 32bit floating internally, so it is only at points where audio files are created (recording or in-place rendering) that levels are critical.
When doing pre-recording level checks, note that for acoustic sources, like vocals or instruments, the levels encountered once the artists get into the performance while recording can be substantially above those from the artists in the more subdued level-check phase. Make sure you get them to do the level checks with a bit more gusto.
+1
Your post, sir, was spot on. Can’t add or remove anything. Pure fact and well explained.
[quote=“roy_mattie”]
2) If you want a target goal for input levels, use the K-14 or K-20 options in the metering options.
I just cannot find these options 
can someone tell me where they are please…
thanks, Kevin
Read my post #33, as there’s a pic to show where to find it.
thanks…never used the this meter before…no wonder i couldn’t find it…
[emoji106] It’s actually quite useful when building a mix from scratch!
Exactly as expected because the channel meter is now monitoring the input to the track before the Pre Gain control!
You’ve been drawing the wrong conclusion.