Louder is better

In 1933, two researchers at Bell Labs—Harvey Fletcher and W.A. Munson—conducted one of the most significant experiments in psychoacoustics. Their experiment was based on a series of tests taken by a group of listeners. Each test involved playing a test frequency followed by a reference tone of 1 kHz. The listener simply had to choose which of the two was louder. Successive tests involved either a different test frequency or different levels. Essentially, what Fletcher and Munson tried to conclude is how louder or softer different frequencies had to be in order to be perceived as loud as 1 kHz. They compiled their results and devised a chart known as the Fletcher–Munson curves. A chart based on the original Fletcher–Munson study is shown in Figure 2.1. I am presenting it upside down, as it bears a resemblance to the familiar frequency-response graphs that we see on some equalizers, with peaks at the top. A similar experiment was conducted two decades later by Robinson and Dadson (resulting in the Robinson–Dadson contours), and today we use the ISO 226 standard (which is still subject to occasional revisions). The formal name for the outcome of these studies is equal-loudness contours.

Each curve in Figure 2.1 is known as a phon curve, labeled by the level of the 1 kHz reference. To give an example of how to read this chart, we can follow the 20-phon curve and see that, if 1 kHz is played at 20 dBSPL, 100 Hz would need to be played at 50 dBSPL in order to appear equally loud (a 30 dB difference, which is by no means marginal). The graph also teaches us that our frequency perception has a bump around 3.5 kHz—this is due to the resonant frequency of our ear canal. Interestingly, this is pretty much the center frequency of a baby’s cry.

One important thing that the equal-loudness contours teach us is that we are more sensitive to mid-frequencies—an outcome of the lows and highs roll-off that can be seen on the various curves. But more importantly, it is evident that at louder levels our frequency perception becomes more even—the 0-phon curve in Figure 2.1 is the least flat of all the curves; the 100-phon curve is the most even. Another way to look at this is that the louder music is played, the louder the lows and highs are perceived. In extremely general terms, we associate lows with power and highs with definition, clarity, and spark. So it is only natural that loud levels make music more appealing—louder is perceived as better.

Figure 2.1 The Fletcher–Munson curves (shown here upside down). Note that on the level axis, soft levels are at the top, loud at the bottom.

This phenomenon explains the ever-rising level syndrome that many experience while mixing: once levels go up, it is no fun bringing them down. The more experienced among us develop the discipline to defeat this syndrome by keeping levels constant.

The fact that our frequency perception alters in relation to levels is a fundamental mixing issue. How are we supposed to craft a balanced mix if the frequency content varies with level? At what level should we mix? And what will happen when the listener plays the track at different levels? The answer is: we check our mix at different levels, and try to make it as level-proof as possible. We know what to expect when we listen at softer levels—less highs and lows. It is possible to equalize the different instruments so that even when the highs and lows are softened, the overall balance between instruments hardly changes. For example, if the kick’s presence is based solely on low frequencies, it will be heard less at quiet levels, if at all. If we ensure that the kick is also present on the high-mids, it will be heard much better at quiet levels. Many believe that the mids, which vary little with level, are the key to a balanced mix, and if the lows and highs are crafted as an extension to the mids, a mix will exhibit more stable balance at different levels. Also, many agree that if a mix sounds good when at low levels, it is likely to sound good when played loud; the opposite is not always true. Another pointer is that we can sometimes guess the rough level at which the mix is likely to be played (e.g., dance music is likely to be played louder than ambient), and so we can use that level as the main reference while mixing (as reference means occasionally—mixing at nightclub levels throughout is next to certain to damage your ears).

There is another reason why louder is perceived as better. When listening at soft levels, we hear more of the direct sound coming from the speakers and less of the sound reflected from the walls (i.e., the room response). Sound energy is absorbed, mostly as it encounters a surface. The small amount of energy our speakers emit at quiet levels is absorbed by walls to a degree that only a fraction of it reflects back to our ears. At louder levels, more energy is reflected and we start hearing more of the room response. As a consequence, the louder music is played, the more we hear the reflections coming from around us, which provides us with the appealing sensation that the music surrounds us. There is an experiment you can do to demonstrate this effect, which might be more apparent with eyes shut—play a mix at quiet levels through speakers and try to define the spatial boundary of the sound image. Most people will imagine a line, or a very short rectangle between the two speakers. As the music is made louder, the sound image grows, and at some point the two-dimensional rectangle turns into a vague surrounding sense.

When making individual instruments louder in the mix, we perceive them better. The core reason for this is masking—the ability of one sound to cover up another. More specifically, the frequency ranges of one instrument mask those of another. One of the principal rules of masking is that louder sounds overpower quieter sounds. The higher the level of an instrument in the mix, the more it will tend to mask other instruments, and the more clearly it will be perceived. When focusing on a particular instrument, it is tempting to bring it up to hear it better, and once up it is likely to sound better. This may be misleading and lead to suboptimal balance.

Percussives weigh less

It is important to distinguish the different natures of the instruments we are mixing. An important mix resource is space; when different instruments are combined, they compete for that space (mostly due to masking). Percussive instruments come and go—a kick, for example, has little to no sound between various hits. Percussives fight for space in successive, time-limited periods. On the other hand, sustain instruments play over longer periods and thus constantly fight for space. To give one extreme example, think of a rich pad produced using sawtooths (the most harmonically rich orderly waveform) that involves unison (an effect that spreads copies across the stereo image), and played in a legato fashion (long notes). Such a pad would fill both the frequency spectrum and the stereo panorama in a way that is most likely to mask many other elements in the mix.

In a practical sense, sustained instruments require somewhat more attention. Whether we are setting levels, panning, or equalizing them, our actions will have an effect over a longer period. Raising the level of a dense pad is likely to cause more masking problems than raising the level of a kick. If the kick masks the pad, it would only do so for short periods—perhaps not such a big deal. But if the pad masks the kick, it would do so constantly—a big deal indeed.

Importance

Imagine yourself being on a Seinfeld set. In the scene being shot, Jerry and Kramer stand in a long line of people at a box office, engaged in conversation. Being the stars of the show, among all people, the production efforts would have been focused on the two stars. The makeup artist, for example, probably spent quite some time with them, perhaps little time with the extras standing next to them, and most likely no time with any other extras standing farther away in the line. In the camera shot, Jerry and Kramer are seen clearly in the center and extras are out of focus. The importance of the stars will also have been evident in the work of the gaffer, the grips, the boom operator, or any other crew member, even the chef.

Equally, different mix elements have varying importance within the mix. The importance of each instrument depends on many factors, like the nature of the production being mixed. In hip-hop, for example, the beat and vocals are generally the most important elements. In jazz, the snare is more important than the kick. Spatial effects are an important part of ambient music. A prominent kick is central to club music, but of far less importance in most folk music. Many more examples can be given. We also have to consider the nature of each instrument and its role in the overall musical context. Vocals, for example, are often of prime importance, but the actual lyrics also play a crucial role. The lyrics of Frank Sinatra’s “My Way” are vital to the song’s impact, and mixing a vocal part as such calls for more emphasis. Arguably, the lyrics to “Give It Away” by Red Hot Chili Peppers are of little importance to the overall song climate.

Importance affects how we mix different elements, be it levels, frequencies, panning, or depth we are working on. We will shortly look at how the order in which we mix different instruments and sections may also be affected. Identifying importance can make the mixing process all the more effective as it minimizes the likelihood of delving into unnecessary or less important tasks—for example, spending a fair amount of time on treating pads that only play for a short period of time at relatively low level. Those of us who mix under time constraints have to prioritize our tasks. In extreme circumstances, you might have as little as one hour to mix the drums, just half an hour for the vocals, and so on.

Natural vs. artificial

A specific event that took place back in 1947 changed the course of music production forever. Patti Page, then an unknown singer, arrived at a studio to record a song called “Confess.” The studio was set up in the standard way for that era, with all the performers in the same room, waiting to cut the song live. But there was a problem—”Confess” was a duet where two voices overlap, but for a reason yet to be found no second vocalist showed up. Jack Rael, Page’s manager, came up with the unthinkable: Patti could sing the second voice as well, provided the engineer could find a way to overdub her voice. Legend has it that at that point, the engineer cried in horror: in real life, no person can sing two voices at the very same time. It’s ridiculous. Unnatural! But to the A&R guy from Mercury Records, this seemed like a great gimmick that could secure a hit. To achieve this, the engineer did something that was never done before—cloning the track from one machine to another while adding the second voice on top. What then seemed so bizarre is today an integral part of music production.

For our purposes, a “natural” sound is one that emanates from an instrument that is played in our presence. If there are any deficiencies with the raw recordings (which capture the natural sound), various mixing tools can be employed to make instruments sound “ more natural.” A mix is considered more natural if it presents a realistic sound stage (among other natural characteristics). If natural is our goal, it would make no sense to position the kick up front and the rest of the drum kit behind it.

However, natural is not always best—natural can also be seen as very ordinary. Early on in photography, it occurred to people that shadows, despite being such a natural part of our daily life, impair visuals. Most advertisements have had tone and color enhancements in order to make them look “better than life.” The same goes for studio recording. It is not uncommon today to place the kick in front of the drum kit, despite the fact that this creates a very unnatural spatial arrangement.

One of the principal decisions we make when we began a mix is whether we want things to sound natural or artificial. This applies on both the mix and instrument levels. Some mixes call for a more natural approach. Jazz enthusiasts, for example, expect a natural sound stage and natural-sounding instruments, although in recent years more and more jazz mixes involve an unnatural approach—for instance, compressed drums with an emphasized kick and snare. This fresh, contemporary sound has attracted a new audience (and even some connoisseurs), and facilitated a wider market for record companies to exploit. Popular music nowadays tends to be all but natural—the use of heavy compression, distortions, aggressive filtering, artificial reverbs, delays, distorted spatial images, and the like is routine. These paradigms, while not natural, increase the potential for creativity and profoundly affect the overall sound. Mixes are sonic illusions. The same way that color enhancement improves visuals, our mixing tools allow us to craft illusions that sound better or just different from real life. People who buy live albums expect a natural sound. Those who buy studio albums expect, to some extent, a sonic illusion, even if they don’t always realize that.

Some inexperienced engineers are hesitant to process since they consider the raw recording a natural touchstone. Often they are cautious about even gentle processing, considering it to be harmful. Listening to a commercial track that was mixed with an artificial approach will reveal just how extreme mixing treatments can be. Take vocals, for example: their body might be removed, they might be compressed so that there are no dynamic variations, or they might be overtly distorted. We have to remember that radical mixing is generally unperceived by those without a trained ear—the majority of listeners, that is. Here are three sentences my mother has never said and will probably never say:

• Listen to her voice. It’s over-compressed.
• That guitar is missing body.
• The snare is too loud.

The common listener does not think or speak using these terms. For them, it is either exciting or boring; they either feel it or don’t; and most importantly, they either like it or they don’t. This leaves a lot of room for wild and adventurous mixing treatments—we can filter the hell out of a guitar’s bottom end; people will not notice. We can make a snare sound like a Bruce Lee punch; people will not notice. Just to prove a point here, the verse kick on Nirvana’s “Smells Like Teen Spirit” reminds me more of a bouncing basketball than any bass drum I have ever heard playing in front of me. People do not notice.