A-Z Of Pro Tools - K Is For K-System

02-07-2015 11:00 AM

One area of Pro Tools 11 HD which received welcome improvements was the metering. It is a pity that there were no metering options based on the new loudness standards of EBU R128 and ATSC A/85 even though Avid told us when Pro Tools 11 was launched they would be added at some point, but the available options are still a vast improvement over the sample peak metering previously available. Most of this was aimed at the post community with options such as PPM meters, the offering for music mixing was the K System.
Designed by renowned mastering engineer Bob Katz, the K system seeks to address the over-use of dynamic compression and the resulting loudness war in music mixing by standardising metering with known monitoring levels. For a thorough explanation of this subject, Bob himself explains the system thoroughly in his article “An Integrated Approach to Metering, Monitoring, and Leveling Practices”. Its important to realise that the K system has nothing to do with K weighting, the EQ curve which is used in loudness measurement to mimic the response of human hearing.

Peak Vs Averaged Meters

The principal meter used in DAW environments has historically been the sample peak meter. While this fulfils its role of monitoring headroom in a digital system it has never offered any meaningful information about the perceived loudness of the signal. In the days of analogue studios and tape machines, peak levels were less important and the VU meter was a commonly used device. While there are differences between perceived level and the indicated level as shown on a VU meter, it was still a useful device as it is essentially an averaging meter. In a similar way to Dorrough meters, the K system combines a peak and an averaged level meter but crucially it directly links programme dynamics with monitoring level.
Crest Factor And Percieved Loudness

Central to understanding the K system is understanding crest factor. This describes the difference in dB between the averaged level of the signal (the kind of level reported by a VU type meter) and the peak level (as reported by a sample peak meter). Even the most dynamic music typically has a crest factor of less than 20 dB, usually considerably less. In his excellent book Mastering Audio, the art and the science, Bob Katz observes an inherent link between monitor position (meaning the position of the volume knob, not the loudspeakers) and the use/over-use of dynamic compression. It seems that typically listeners respond most positively to average programme levels around 83dBSPL and if monitors are attenuated, a tendency exists for engineers to use dynamic compression to bring the average level of the music back up towards this 83dB “comfort zone”. Conversely, if monitor levels are excessive, the tendency to bring up the average level using compression is reduced.
Examples of (L-R) K-20, K-14 and K-12 meters passing identical signals to a master fader with a sample peak meter.
There are three options in the K system: K-20, K-14 and K-12. The Pro Tools Reference Guide suggests K-20 at the mix stage to encourage a mastering-friendly mix, K-14 is suggested for mastering with K-12 being a possible option for broadcast. It should be noted that although the K-system predates EBU R128 and ATSC A/85 and that it was always designed for music production rather than broadcast production.
Calibrated Monitors

It is because of the influence monitoring level has on the use of dynamics processing that the link between typical crest factor (i.e. dynamic range) and monitor calibration (i.e. precisely how loud you monitors are) is made. If your monitors are set to the appropriate level then most people will naturally mix towards this “comfort zone” of an average level of 83dB. The different K metering types are suggested metering/monitor levels which encourage mixing towards a target crest factor. Hopefully you can see that without the associated monitor calibration the metering with its variable 0db point is of little use. When mixing to K14 the 0dB point on the meter is 6dB louder but it should only be used with the accompanying -6dB attenuation of the calibrated monitor output. What you are trying to achieve is keeping the average level the same but encouraging use of different dynamic ranges through our perception of loudness. In fact, in a way the K System seeks to offer meters designed not to be looked at all but instead allowing your ears to be the judge of how loud something should be, which after all is what really matters to the listener.