ston, look at it this way...
 
Yes, you (and your listeners) ultimately control loudness with the volume control. But let's say you aren't consistent with the volume you mix at. Maybe during the day you like it cranked but have to do it quieter at night to avoid problems with the neighbors.
 
Because of the equal-loudness curves batsbrew alludes to above, you perceive sound differently at different volumes, a phenomenon that will unconsciously affect all your mixing decisions. You'll end up with inconsistent mixes: too much bass in some, not enough in others; a tendency to over-compress some things and under-compress others; lost vocals, oppressive kicks, harsh guitars; mixes that don't translate well from one system to the next. Mixes that sound good loud but fall apart at low volume.
 
These can only be avoided by mixing and mastering at a consistent volume, and by correlating meter readings to what things are supposed to sound like - at that volume. That's where the calibration comes in. Once you've established a known relationship between meter readings and speaker volume, you are then free to turn the speakers up or down, knowing that you can still rely on the meters as a sanity check. 
 
The inspiration for the K-system was a previous loudness standard that had been in place for decades for motion picture exhibition. It assures that a movie's sound will be consistent from one venue to the next. Just as you want your music to be in the same ballpark as commercial offerings heard on iPods and radio, as well as being internally consistent within a collection on a CD. 
 
You can certainly get there without the K-system or R128, and it was done that way for many years. These standards are just conveniences to get you there with less trial-and-error and greater consistency.

bitflipper:
That was as good an explanation as I've ever heard.  Thank you!!

Bob Katz has a book that explains it better and much of the plug-ins available now follow the K-System or using Bob Katz approved hardware emulations.
 
Or BK for short.
 
p.s. Also bitflippers description is very well explained.
 

Reading this sort of stuff makes me realise that I simply do not have a clue what I am doing...
 
I need to take note though, and look further into this, as I am sure I am listening to my music much too loud at times, even to the point that I am left with that ringing in the ears sensation.
 
Also, I only ever mix on headphones nowadays, on account that people around me don't want to listen to my crap any more and I do not have the luxury of a sound proof studio.
 
Thanks for all the posts above, this is all certainly food for thought and I'm sure that, in the long run, my mixes will improve (a pity the music won't) and my ears will be thankful 

I understand the principle and the K system provides a good guideline, but unless you have had your own ears accurately calibrated for equal loudness and are monitoring with fresh, un-stressed ears (i.e. they're in the same ideal state as when they were calibrated), it remains a guideline only.
 
Also, to produce the equal-loudness contours in the first place relied upon people's subjective report on what constitutes "equal loudness".  You may argue that such differences are minimised by taking a large enough sample, but this is not necessary the case if the underlying mechanism which is being measured is intrinsically broken.
 
People are, in fact, very bad judges of 'equality' (or, equally, inequality).
 
This doesn't involve judgement of noise levels, but provides a useful demonstration of the principle involved IMO:
 
Fast forward to 4:35 in the following BBC documentary, presented by mathematician Marcus du Sautoy:
 
https://www.youtube.com/watch?v=GqTB7mLpiQY
 
I like to monitor across a variety of levels, sometimes using headphones, very much according to how I'm feeling, a part of which is a reflection of the state of freshness or fatigue of my ears.  This to me makes much more sense than setting a fixed level dependent upon the subjective opinion of a sample of people who are naturally poor judges of what constitutes 'more' or 'less'.

Thanks for that link, ston. It was a very enjoyable watch. It does seem to reinforce the reason the K-system exists, by showing how unreliable our senses are at determining absolute values and the need for objective measurements that are independent of human limitations. It indirectly makes a good argument for visual aids in music, which the "use your ears" crowd dismiss as crutches.

Nice stuff ... my two cents...
 
When I am teaching the K-System and monitor position in my class I describe what Jef and Bit are saying with the following analogy.
 
I ask my students to imagine using PhotoShop in a completely dark room and being asked to adjust the brightness and color levels in a digital image - then go outside to a sunlit area and perform the same adjustments.  Establishing a consistent balance between environmental light and screen energy is a necessity in such work.
 
I work on multiple client projects over weeks, months and sometimes years.  When i return to a project regardless of what stage we are in (tracking, mix, edit, master, etc) I need everything to be exactly as it was when we last worked on the project.  I carry my SPL meter with me when I move to different studios to do work. I have Bob Katz's Pink Noise file on a USB drive I carry around as well.  I can always set a quick position anywhere, possibly not perfect but close enough for Jazz.
 
Monitor position is about establishing repeatable permanent baseline relationships with the energy coming from your system and how your environment delivers that energy consistently.  Regarding the level - I've done the 83dbfs for years and have found it a bit hot for the work that I do (mostly classical)
 
Nice thread
 
-D

Great analogy, Don. I've never put those two things together as an analogy, despite my own experience in photography.
 
I spent 5 years working in a photo lab (remember those?). It was a professional service, so every negative was previewed and color-corrected by a human before printing. A clever device made by Kodak called a Video Color Negative Analyzer was used, which employed rotating filter wheels in front of an interlaced monochrome CRT. It was the same color-encoding system used for space exploration because it guaranteed extremely accurate color representation. A VCNA was to photography what monitors and equalizers are to audio.
 
The VCNAs were calibrated daily, and the lighting in the room where they were operated was carefully controlled for brightness and color, just as you control the acoustical environment in your mixing or mastering room.
 
My job was making sure the printers subsequently matched the VCNA color corrections exactly. Part of that entailed color-correcting the entire plant's overhead lighting, so no matter where you stood in the (windowless) production space colors looked the same. It was all about calibration and consistency.
 
I didn't appreciate how well that system worked until I took a side job at an automated lab geared for volume rather than quality. There, no such efforts were bothered with, and you didn't know what colors you'd get from photo to photo. When I mentioned this to the lab's owner, he dismissed it as irrelevant. Much the same way some amateur audio engineers dismiss speaker calibration and acoustical treatments as irrelevant.

Back in terms of audio I find when I work with the K system for every project especially an album where there may be a collections of songs or tracks then the consistency it provides is very good in terms of signal flow and gain structure.
 
The track count and thus mix complexity can smaller or larger but the same follows.  Keep your track rms levels all at the K Ref level.  If there are effects/chains make sure the right rms level is going in and out of every plugin.  Setup as many buses as needed but maintain the buss overall level to also be at the ref level.  Then the buses all finally meet on the stereo buss and yes it too is sitting at the perfect ref rms level as well.  Very rare clip lights anywhere to be seen even in a large mix.  VU meters do the monitoring so well.
 
Then when all this is right you turn up your speakers to 83-85 dB.  I find the SPL meter permanently in front of me pretty handy too.  Stops you from creeping up your room monitoring level.  Some genres require you work at levels louder than the norm but only for a short time.
 
Working on an album and having all the pre masters at the same rms level certainly goes a long way to making mastering much easier.   K reference levels are very close to loudness levels too. (for consistent energy tracks all the way through)
 

Well all this talk about metering has got me reading and experimenting again. Being as I have been away from all this sort of thing for a long time and Sonar Platinum was my entry back into.
 
I was playing through the console options and notice you can specify Peak+RMS,Peak or RMS only in the different metering options. Providing you take any errors into account, you could use that in place of another meter until you installed a more detailed one.