K-System

Most of the stuff I mix gets listened to by a relatively small population and I know how they will listen. Some of my friends have great gear, and for them I mix for that. Others have cheap geat, and then I mix for that. I try to make sure it will sound ok on either cheap or great, but I target whatever I think the main listening environment and volume may be.

I actually monitor over a range of SPL levels. The reason why I like 85 dB is because my space is quite big and I feel it is a nice volume to work at over the long haul.  But yes I dont mind working at 80 dB and even lower at times especially when checking mixes on my small mono Auratone type speaker.  I also do some heavy hitting at 95 dB and 105 dB SPL for short periods.  I find it reveals things like the bass low end and the reverbs tend to show themselves up loud a bit too.
 
K system is NOT just about SPL monitoring levels, it requires you to monitor rms or VU levels right through your system as well.  Although tracking does not require it per se I see no reason why tracking levels cannot be set using the VU meter as well.  So I track everything at the ref level right at the start.  I find then it easy to get balances right on a buss and then the buss ends up at the ref level and also the final mix too.
 
I work at K -14 a lot for more general things and K-20 in situations where I want more headroom.  So when I do this I recal my VU meters and turn the monitor level up so it is back to 85 dB SPL.
 
They say 85 dB is where the ear is the most flat (if you could call it that)  It is lovely volume for sure.  Pink noise at 85 dB will sound quite louder than music at the same level so be aware of that too.
 
Loudness wars are about something else.  You need to get into Loudness meters and also Dynamic Range meters too.  But before you master if you work with the K system you are well on the way to a nice punchy and dynamic master before you even start.
 
The Klanghelm meters are excellent and also they have the closest ballistics to the real deal which is what I also have too.

That is it exactly. Set your incoming levels so that the VU just reaches up to and hovers around the 0 dB VU mark.  With very transient sounds such as drums etc the VU meters are less help because they respond more slowly.  In these situations use your peak metering.  That is what it is there for.  But by the time you start sending say 8 drum sounds to a drum buss the VU will work OK over the buss because there is usually enough rms component in the signal to move the meter nicely.
 
I have created some special test tones to test the ballistics of a VU meter. eg how the meter dances to the music.  I have very nice real ones and over time I have set the Klanghelm meters to get very close to the movement of the real ones.  I will need to check my settings again but I will let you know what they are.  I prefer to match te ballistics rather than use Bob's 600ms time.
 
For me I prefer the VU VST's to move as close as possible to the real meters.  The ballistics of the meter will tell you tons about the signal it is monitoring.  After a while you become very expert at observing them.  (Note VU meters that use a bar graph display are NOT as good in this mode either)  You can even set compressors right once you understand the movement of the VU needle.  Really well mastered material makes them move a certain way too.  If the meter swings wildly it means there is something wrong.  (often it is unwanted subsonic material that is making the meter swing wildly so you track it down and sort it out)
 
 
 
 

Paul,
 
I know I'm a hobbyist and probably under-qualified to contribute here, but I can't resist.
 
I use RME's digicheck all the time on the output master bus. It has a 'start' button on it and measures the average loudness (ITU 1770/EBU R128)...
http://www.rme-audio.de/forum/viewtopic.php?id=13925
 
I measure the main output bus with this. It has targets for 12, 14 and 20 for the K-System.As I mix, I target 0 and balance compression and limiting so that I fall right at 0 with only a touch of limiting (maybe a dB or so).
 
I have a kid in the house, so I mix with headphones sometimes and with monitors at other times. I always end up checking the mix using ear-buds with my phone and in my car on the highway (monitoring a mix with a good amount of white-noise is recommended by Craig Anderton and others).
 
I did go through all the work of measuring every single track's average loudness and setting them all to be roughly the same, but this goes wrong when you have one instrument much louder than the others and bleeding into the other mics. Then, when you bring up the clarinet, you are also bringing up the overloud(tm) electric guitar.
 
So, after going through the effort of setting all the tracks to the same level, I'd have to double back and fix it by ear.
 
Now, a couple of years down the road, I still sometimes start by measuring loudness and adjusting things. It doesn't hurt as a jumping off place, but after I've mixed 1 or 2 of the songs from the night's concert, I'm usually ignoring those meters.
 
I do keep an eye on the meter on the main output through the mixing process. I tend to make all the songs around the same volume. (14) except choirs (20). On occasion, I listen to some post from the forum or a link from a friend and switch my meter to 12 as I listen. Once, I forgot to change it back and I was like, What is going on?? as I pushed all my levels through the roof. This is to say, I really never use 12 personally.
 
The sound coming from the stereo tends to fall at my personal comfort levels. I adjust it around different playback volumes. It's always interesting to hear the volume my wife sets it too when she's talking to my daughter in the backseat. Wow, all I can hear is the snare!! or some other thing.
 
Songs are pretty short compared to movies. It's relatively easy to just measure the average loudness on the mains and bag it on all the individual tracks. As Jeff points out, average loudness on a tom track is kinda pointless, so then you're looking at the average loudness across all the drums. Me, I tweak the drums a lot. Consider a track that suddenly has a lot of rim work b/c it's a quiet song... how do you want to do that one? Do you want the song to be a dB softer on the CD?
 
I do find that most of my fine adjustments are in the range of 1dB. Increasing the bass 1dB while maintaining the same average loudness means reducing the level of all other things ever so slightly and the overall impact is pretty dramatic at a healthy volume.
 
When it comes to setting levels for tracking, I tend to use RME's auto-level feature. This takes the human pretty much out of the equation. It keeps peaks closer to -14 than -20 (and adjusts gain when peaks come within -6). When I manually set gains, I make sure I don't clip, but I don't bother measuring them with the DigiCheck EBU Meters. I could but why bother?
 
What's the point?
I just wanted to share what I learned experimenting with the K-System. I don't really take it too seriously. I do make a good effort to make good mixes and I did learn a lot measuring average loudness at different points in the chain. In the end, an average loudness meter on the master out does the trick for me. It measures crest factor and average loudness and helps guide my compression and limiting (which I then adjust track levels around as I mix). I kinda think of mixing with the track faders to compression to limiting as bottom up mixing, and using the EBU meter down to limiting down to compression down to faders as top down mixing.
 
**After thought: I also always have the RME DigiCheck Totalizer on the the master bus. This has a vector scope and spectral analyzer in one window. I use workspaces to make it easy to open both to match the sample-rate of my project. 
 

I think you can over complicate matters doing Loudness measurements during a production.  I don't bother with any of that at all until I am in the mastering phase.  So measuring Loudness on individual tracks or buses is a bit of a waste of time.  It won't tell you much. eg a track may be very loud but end up very soft in the mix so what is the point of monitoring its loudness way back there.  It is still very easy to get all your tracks sitting at the same rms level with a VU meter.  (the VU meter will cover about 90% of the work required!)
 
I work using the K system all the way through a production maintaining correct rms or VU levels on tracks buses and final mix.  eg If I am doing a project at K -14 then I ensure all those areas are at K -14.  I have found that the Loudness measurements often are very close to the K system ref levels.  So if I produce a mix which is sitting at K -14 then the Loudness figure is often very close to that eg -15 LU etc.  (Assuming of course the track is a reasonably consistent level over the course of the track) Loudness meters are measuring over the whole time of the track.
 
Once a track is premastered around a K ref level then I go into mastering mode and then I involve Loudness and Dynamic Range metering as well as rms metering.  (All 3 are very important!)
 
Monitor controllers are fine and all but also not really necessary either.  The purpose of a monitor controller is to ensure exact SPL levels in the room.  (Without an SPL meter present and also offer very high audio signal path quality to your monitors but for most purposes we dont need that level of quality) That can be achieved easily having an SPL meter permanently set up in a good spot.  That will do the same thing overall.  I just check the meter all the time to ensure I am in the 85 dB SPL ballpark.
 
What I find more interesting is how often I look down at the meter after say 4 or 5 hours or so and notice it is telling me 90 or even 95 dB. ie the SPL level has just crept up over time!  That is far more relevant.  Then I take a break for a while and reset the SPL level in the room back to 85 dB or so.