Bristol_Jonesey
Your starting fader positions will also vary depending on the total number of tracks in your project.
 
As a rough rule of thumb, the formula is (in Excel) -20 x Log(total number of tracks)
 
So as an example, for a 15 track project, initial fader positions should be around (on average!) -23.5dB

 
This thread is good reading with helpful reminders to keep my old habits from creeping back in and stealing clarity from my mixes. !!
 
And yes Bristol, I created an excel sheet with the formula you posted.
 
Thank you to all who are taking the time to dig into this topic!

I do sometimes, or very often put a mastering limiter on the mixing bus, before taking it to mastering session. But that is only because I want the levels to be on spot. I always bounce out my tracks at -6 to -10 db in the mix, so hard to tell how the mastering chain then will react to all the sounds and glues. So, I use it as a level adjuster. For sure taking it back off when bouncing out the mix. For the workflow it is better for me, so I do not have to bounce out several mixes just to level the bass or something else.
 
Back to the thread. I have never heard of people who bounce out a mix with a headroom of -0.1. without any compressor or limiter, they are doomed to get clipping in the mix. Compressors and limiters on the mixbus is a no go. The mastering engineer (or if you do it yourself) always ask and need no compressor or limiter on the master bus.
 
Headroom is the essence for the mastering stage. Working with a -0,1 headroom is a nightmare. You also said you want the mix to be as high as possible. Why? You will get that after the mastering stage if done properly. Maybe I did misunderstand, but you aren`t going to upload the mix to any streaming services?
 
If someone send a -0.1 mix to a professional mastering engineer I am sure he or she will get back to the person and ask for a lower mix. The lower the mix are, the better for the mastering session. So as John said in the first reply "Lower the volume" . It is not only obvious, it is also a solution.
 
Nothing wrong with the Sonar Meters in my experience.
 
Good luck with whatever solution you end up with 

sharke
This is a very interesting and informative article on levels and gain staging, and I would highly recommend that you read it. 
 
https://www.soundonsound.com/sos/sep13/articles/level-headed.htm
 
Of interest:
 

.
.
.
 
In practice, many people, including me, believe that the summing engines in different DAWs don't always produce the same results when summing lots of very high-level signals, and that these differences can be audible. The probable reason for this, to put it simply, is down to differences in how the 32- or 64-bit floating-point calculations are rounded to create the 24-bit audio part of the floating-point data.

 

If we're talking about Sonar, the "many people, including me, believe that the summing engines in different DAWs don't always produce the same results when summing lots of very high-level signals, and that these differences can be audible" are quite wrong. It's complete and utter nonsense and easily proven so.
 
The probable reason for this, to put it simply, is that the people who believe this stuff haven't a clue how the math works regarding digital audio.
 
Having said that, there's absolutely no reason to try to mix to within a fraction a dB of 0dBFS. If you aim lower, there's less chance of messing things up for no reason, and it's just a matter of turning up the listening volume a little now and adjusting the gain for the output format during mastering.

drewfx1
 
The probable reason for this, to put it simply, is that the people who believe this stuff haven't a clue how the math works regarding digital audio.
 

That's fair enough. Count me in as one of those people who don't understand the math and tend to believe people who sound like they do. LOL! 
 
On the subject of people who sound like they know what they're talking about, The following epic Gearslutz thread is worth a read not just for the back and forth on the subject of levels and summing and clipping and whatnot, but also for some extremely interesting posts from SSL/Sony Oxford veteran Paul Frindle. 
 
https://www.gearslutz.com/board/so-much-gear-so-little-time/463010-reason-most-itb-mixes-dona-t-sound-good-analog-mixes-restored.html
 
I was especially intrigued by this post: 
 
https://www.gearslutz.com/board/5063007-post76.html
 

However - harking back to the headroom issue (or lack of it in digital), rolling of LF increases the signal peak level most of the time - because it changes the wave shape (by differentiation)! Changing waveforms can create larger peak levels - even if you have done so by 'losing something'! So doing this on a digital system with everything flat out - and avoiding overs - will actually reduce the level and presence of the instruments involved, and even the whole mix, if you're aiming at high volumes and modulation!
......................................
However to maintain the impact of this and avoid overs and clipping you will need to lower the levels throughout your mix to accommodate the extra peak level - and make sure your final mix maximiser/limiter 'does the right thing' in pushing the density and gain of the program without simply clipping off your tightened sound and turning it back into mush again.. The work I have been doing has been heavily involved in preserving these subtle effects - if they are actually still there in the program, while remaining compatible with current loudness trends - and the search for improvement goes on still..

So yet again operating at lower peak levels - and knowing what you are doing when mixing, rather than leaving it to little-understood side effects of your gear - can indeed give you a fantastic advantage and save you a fortune :-)


 

Some time ago I noticed that simply engaging the HPF of the Quadcurve EQ was sometimes enough to "overheat" a ProChannel signal (i.e. whereas before the ProChannel clipping light was not showing red at any point during playback, engaging the HPF would make it glow red in a few places). I know there has been a lot of talk in the past about what those ProChannel clip lights actually mean in reality, but what this did make me realize that as Frindle says, rolling off the lows with a filter can actually increase the peak level even though you're losing frequencies. I always found that interesting. 

drewfx1
If we're talking about Sonar, the "many people, including me, believe that the summing engines in different DAWs don't always produce the same results when summing lots of very high-level signals, and that these differences can be audible" are quite wrong. It's complete and utter nonsense and easily proven so.
 
The probable reason for this, to put it simply, is that the people who believe this stuff haven't a clue how the math works regarding digital audio.

I stopped buying Sound on Sound magazine when I saw one too many of these articles where someone 'believed' they could hear a difference between 2 pieces of gear or two approaches but could not or would not actually measure it.

 From Steven's Slate Audiophiles fb page - Steven Slate Channels going into the red is fine because they won't digitally clip because it's a floating point mixer.  The only thing that cannot clip is the final master since it's gotta go cleanly out the D/A's.
 
So in my estimation - he's saying busses can clip, too.  CLA levels from the Audio Legends course -

 

Regarding the summing bus controversy, what complicates matters is that people may ascribe the wrong cause for an effect. The math of a summing engine should be indisputable, it's just math. It doesn't care whether it's adding big numbers or little numbers. However, think about what comes afterward. There can be non-linearities in converters and/or inter-sample distortion, which would come into play only when summing high-level signals. So someone concludes the issue is summing high-level signals when it has nothing to do with the summing bus.

Absolutely, Craig. The SSL X-ISM for the final say so in these matters!

sharke

 

Some time ago I noticed that simply engaging the HPF of the Quadcurve EQ was sometimes enough to "overheat" a ProChannel signal (i.e. whereas before the ProChannel clipping light was not showing red at any point during playback, engaging the HPF would make it glow red in a few places). I know there has been a lot of talk in the past about what those ProChannel clip lights actually mean in reality, but what this did make me realize that as Frindle says, rolling off the lows with a filter can actually increase the peak level even though you're losing frequencies. I always found that interesting. 

That's a normal property of a low or high pass filter. Basically, the neat shapes you see in the graph are a bit of a lie, really. There's always some sort of resonant boost around the cut-off frequency.

Sorry, John. There are lots of filters with no resonant boosts. The most basic filter, Butterworth, is specifically designed for no resonance. However, the time-domain waveform can certainly have a larger peak-peak swing after going through a Butterworth filter. It is not always obvious to relate what's going on in the time-domain by mucking with it in the frequency domain.