Saturday, April 20, 2019

How To Mix At The Right Volume Level

Back in 2018, we published an 

article about the R128 standard 

and how its adoption of Loudness Units as a way to measure perceived level has the potential to end, or at least mitigate, the “loudness wars.” This has already been realized to some degree. Trace this back far enough, and you’ll find mastering engineer Bob Katz’s work from the late ’90s to adapt standards that resulted in the K-System, which is used in the film industry for putting the focus back on dynamic range, instead of “loudness at all costs.”

If you leave enough headroom when mixing and mastering, and listen at a consistent, calibrated monitoring level, then the K-System is probably redundant. In fact, some professional engineers scoff at it (while other pros praise it — so what else is new in the ego-friendly world of pro audio?). And now with R128 and Loudness Units, we don’t really need the K-System any more, right?

Well, not so fast. With much music being made in home studios, people sometimes aren’t aware of the traditions “big studios” implement as a matter of course. The K-System can be helpful in promoting good mixing/mastering/monitoring practices. Treat it as another tool, not a religion, and you’ll probably find it can help you make better, more consistent, mixes.
How Loud Is Loud?

Unfortunately our ears aren’t perfect, as exemplified in the Fletcher-Munson curves. Simply stated, at lower listening levels, human hearing drops off at high and low frequencies — but as audio levels increase, there’s less of a difference compared with the midrange. What this means is this: No matter how carefully you mix or master a song, every listener will hear it differently.

Let that sink in.

They’ll be listening at different levels, over different systems, sometimes with younger ears that can still hear high frequencies, sometimes with older ears that can’t. The bottom line is that once your music goes out into the world, you have no idea whatsoever how it will be perceived.

This is why it’s so important to have a flat response as you mix or master. If you mix bass-heavy because of bad room acoustics, then the music will sound right only on systems that are bass-light — or when played in rooms with the same kind of bad acoustics! If the listener hears with a flat response, the odds are that they’ll perceive something closer to what you intended. But this also means you’re best off monitoring at levels where the response of your ears is as flat as possible, which unfortunately falls around 90dB — and that’s way too loud for repeated exposure. Fortunately, the range between 75dB and 85dB SPL (sound pressure level) is reasonably flat, and more likely to be the level at which your audience listens to music.

As a result, it’s best to try for a sound that’s equally “off” at the extremes. In other words, if at really low volumes the bass and treble seem a little shy, but at high volumes they’re overbearing, then you’ve probably hit the sweet spot for those people who listen at rational levels. If you tailor the frequency response so a hint of bass comes through over a smartphone speaker, then it will probably sound round and full over a car stereo (unless the bass is turned up all the way — which is a common affliction with rental cars until you dive into the menus and tweak the settings).

This is all well and good, but it’s not exactly a scientific method. The K-System provides a reference that takes some of the guesswork out of the process.
K-System Metering in the Home Studio

In my opinion, for a home studio, the most important aspect of K-System metering is that it couples meter readings to standardized monitoring levels. Suppose you monitor a mix at one volume one day, monitor a different mix at a different volume the next day, and then assemble both mixes into an album — you’ll wonder why they sound different. Well, it’s because your ears heard them differently while you were mixing. Had you monitored them at the same average level, you wouldn’t have that problem.

The K-System borrows from the film industry, which typically aims for monitoring around 83dB SPL in large control rooms. That’s pretty loud for small studios and also is considered close to the maximum safe level that won’t cause hearing damage with eight hours of exposure over a 24-hour period. The K-System matches this monitoring level to an average level of 0 on a K-System meter. Note, there’s a misconception that the K-System demands listening at a certain level, but that’s not the case. You can still achieve the benefits of consistent monitoring if you choose, for example, 75dB SPL as your monitoring “standard.”
About Those Meter Scales…

There are actually three different K-System meter scales, but unlike typical meters, these use a linear scale. In other words, each dB division occupies the same width (fig. 1). 

Figure 1: Cubase’s master meter includes the option to choose the three different K-System scales.

A conventional logarithmic scale increases the width of each dB as the level gets louder; although this corresponds more closely to human hearing, it’s more inclined to encourage mixing “hot” because visually (like it or not, people do look at meters when they mix), it looks like you’re using a reasonable amount of dynamic range. With a K-System meter, if you’re mixing so hot that the levels spend most of their time at the higher end of the K-meter scale, you’ll have visual confirmation that you aren’t using very much dynamic range.

In the K-System, there are three scales. However, 0dB does not represent the maximum possible level in any of them: instead, the 0dB point is shifted “down” from the top of the scale to either -12dB, -14dB, or -20dB, depending on the scale. These numbers represent the amount of headroom above 0, and therefore, the available dynamic range.

For a reference when mixing or mastering, you choose a scale based on the music you’re engineering — like -12dB for music with less dynamic range (e.g., dance music), -14dB for typical pop music, and -20dB for acoustic ensembles and classical music. If these numbers sound familiar, they’re also the LUFS numbers people often try to hit when conforming to the R128 standard. (For pop, I usually aim for an average level of -11 to -12 LUFS because the music might end up on CD as well as streaming; while those levels won’t win any loudness wars, at least it won’t sound too out of place in a playlist.)

The goal is to have your average level hover around the 0dB point. If a transient needs more headroom, there’s 12dB, 14dB, or 20dB of additional headroom — no worries. The headroom also means that intersample distortion caused by the D/A conversion process probably won’t be an issue. As with conventional meters, the K-System’s meters have green, yellow, and red color-coding to indicate levels. Levels above 0dB trigger the red, but this doesn’t mean there’s clipping — look at the peak meter to see if that’s occurring.

The K-System is probably most often used for mastering, so you’ll sometimes find processors like limiters or metering suites that include K-System meters (fig. 2). 

Figure 2: The Project Page in PreSonus Studio One, which is dedicated to mastering, includes K-System metering. Studio One’s Limiter also includes K-System metering, as does its Level Meter plug-in.
Calibrating Your Monitors

To use the K-System properly, the basic idea is that you need to calibrate your monitors so that stereo pink noise reading 0 on the meter produces 83dB SPL, as measured by a sound level meter like the

 Galaxy Audio Check Mate CM-130 SPL Meter 

 (fig. 3). But again, note that this level isn’t a rule. You need to choose a level that’s comfortable for you, which will be lower in a smaller room, and that hopefully is a level where your ear’s response is relatively flat (but also loud enough that the effects of dynamic processing are obvious). 

Figure 3: The Check Mate CM-130 is a useful tool for measuring SPL.

To calibrate, you’ll need a pink noise source — you can generate pink noise with a program like

Magix Sound Forge or Steinberg WaveLab,

but you can also find downloadable pink noise samples on the web. Run it through your system, and turn it up until the K-System meter reads 0; set the sound level meter to a slow response with C weighting, and place the meter where your ears are as you listen to the monitors. Adjust the monitor level for your desired standard dB SPL reading (I use 76dB SPL with nearfield monitors).

Having a consistent monitoring level does not preclude using other volume levels. I always start mixes at low levels to keep my ears fresh. After the pieces seem to be in place, I’ll go for the standard monitoring level for more detailed work. Eventually I’ll listen at a variety of levels to experience the range over which listeners will hear it, and if necessary, return to the standard level for making tweaks.

Note that you’ll need to do individual calibrations for the three meter scales if you use all three of them. You can put labels next to the level control on the back of your speaker to show the settings that produce the desired output for each K-Scale. Also note that this is a rough calibration; doing a truly refined calibration may mean calibrating each channel separately, verifying that your pan pot’s center truly is “center,” and the like. For more information, surf the web — you’ll find everything from Bob Katz’s  

Digital Domain website,

which describes use and calibration of the K-System in great detail, to opposing viewpoints that argue it’s redundant in the age of LUFS.

However, all I can go by is my own experience, which is that monitoring at a consistent level helps produce more consistent mixes and masters — and the K-System simplifies the process of finding that consistent level. 

Craig Anderton

Craig Anderton leads a dual life as a musician and author. As a musician, he has played on, mixed, or produced over 20 major label releases, as well as mastered hundreds of tracks, and recently released the album Simplicity. As an author, he has written over a thousand articles for magazines like Guitar Player, Sound on Sound, and Pro Sound News. He has also lectured on technology and the arts in 38 states, 10 countries, and in three languages. His web site is  
craiganderton.comRead more articles by Craig »

No comments:

Post a Comment

Comments Are Moderated And Saved