Chapter 2
Listening
Anyone with a computer, some software and a pair of speakers has access to a studio that’s more than adequate. In fact, some world-class productions were mixed with the exact same tools that are at the average home producer’s disposal. So theoretically, you could compete with the very best if you had the chance to work for top artists. It all comes down to the choices you make.
And this is where things get tricky, of course. You will need a combination of talent (which this book unfortunately can’t give you), know-how (which this book will hopefully contribute to), and reliable monitoring. The latter is the first topic to be discussed here, because every mixing decision you make is based on what you hear. Of course, this depends largely on the characteristics of your room and your speakers, but how you use them is the biggest factor. That’s why we will start with our finest tool, and all its peculiarities: the ear.
2.1 Perception
During mixing, you make decisions about the relationship between individual sounds, the use of the frequency spectrum, and the dynamics. Naturally, you can only rely on what you hear. But what you hear—and how you interpret this—is dependent on more factors than you might expect at first.
The human auditory system is like a set of microphones, hooked up to an impressive signal processor, which interprets the incoming sound. The main difference between a biological and an electronic ear is whether the sound is transmitted linearly or not. A microphone is very capable of linearly converting sounds within the human hearing range into electricity. In other words: the frequency spectrum and dynamics of the electric signal transmitted by a microphone are virtually the same as the frequency spectrum and dynamics of the source sound. Our ears, on the other hand, don’t respond to sounds in a linear way: the auditory response is dependent on the properties of the source sound (loudness, dynamics, frequency characteristics) and on the duration of the ear’s exposure to the source sound. And it just so happens that these are the very properties you manipulate during mixing (see Figure 2.1). So if you want to make reliable mixing decisions, you have to keep the behavior of your ears in mind.
Figure 2.1 Perhaps the hardest thing about working with sound: you manipulate properties of the source sound, while similar properties are simultaneously being influenced in your perception of that sound, depending on the monitoring level you use.
2.2 Your Ear Is a Compressor
Adaptation
Your auditory system is aimed at distilling as much relevant information (such as voices, locations of prey or signs of impending danger) from your surroundings as possible. Sometimes crucial information is quite hidden: it’s only barely audible above the background noise. Your ear automatically adjusts itself to these conditions by shifting its own dynamic range to the area where the information is located. This process, which is similar to the way pupils respond to the amount of light that’s coming in, is called adaptation. An extreme example of this is that people in a perfectly quiet environment will eventually start to perceive their own heartbeat.
Adaptation works for both faint and loud sounds: a faint sound can be perceived as louder, and a loud sound can appear more faint. So adaptation works just like a compressor, but a lot slower. You’ll probably know the feeling of walking into a venue during a loud concert: after five to ten minutes, the sound doesn’t seem as loud as when you just came in. And when you bike back home, it also takes a while before you hear the sound of the wind in your ears getting louder again. And that’s a good thing, because if the adaptation mechanism reacted quicker, it would constantly intervene in musical proportions, like a compressor. Adaptation can be an obstacle when making critical mixing decisions, but it doesn’t have to be if you adjust your working method to it.
During mixing it’s customary to vary the monitoring level, so you can evaluate the mix at different sound levels. The best mixes sound balanced and impressive whether they’re played softly at home or blasting at a club. But if you vary the level too fast, your ears don’t have time to adapt. A common mistake is returning to an average monitoring level after you’ve been listening at a high level for a while, and deciding that you need more compression. Because the details you heard before now seem to have vanished, it makes sense to bring them up by decreasing the overall dynamics. But if you had held off your judgment just a little longer, your ears would have adapted to the new listening conditions, and all the details would have automatically reappeared.
Another pitfall is turning the sound up for just a moment, which will immediately make you impressed with the impact of your mix. But if you listen a bit longer at this high level, you will start hearing the same problems in your mix. So it’s very important to fully adjust to the new monitoring level. A rather nostalgic engineer said in an interview that what he missed most about using analog tape was the rewinding part, because this would create a necessary moment of rest before he could start evaluating the mix afresh. In a modern setting, it’s a good idea to create these moments of rest yourself, for example by having short breaks. After changing the monitoring level, it can also be very refreshing to first listen to a production front to back at the new level before you make a single adjustment. This will allow you to form a new frame of reference (to gain perspective on the project as a whole), which can then become your new starting point.
Figure 2.2 The dynamic range of your auditory system—and therefore the perceived loudness (the red line)—adapts to the ambient noise. However, it takes a while to get used to new conditions, which is why the concert seems much louder in the beginning than at the end, and the traffic noise on the way back seems much softer at first than when you’re almost home.
Protection
Exposure to high sound pressure levels can damage your ears. That’s why we have a built-in protective mechanism that’s activated above 80 to 85 dBSPL. This mechanism, which is called the acoustic reflex, causes contraction of the middle ear muscles when the threshold is exceeded. Or when you start talking, because the acoustic reflex also suppresses the sound of your own voice, so you can keep listening to your surroundings while you talk. These muscle contractions hamper the transmission of sound in the ear, causing compression with a ratio of about 3:1. However, this compression mostly affects low frequencies, so it also affects the perceived frequency spectrum.
Unlike the adaptation mechanism, the acoustic reflex reacts much faster (with an attack time of 50 to 100 ms), so it can considerably affect the perception of musical proportions. Therefore, it’s not a good idea to determine the amount and type of compression in a mix at a very high monitoring level, because then you’ll get the compression inside your ears as a bonus. If you listen to an overly dynamic vocal recording at an average monitoring level, it seems to pop in and out of the mix, but at a loud level it’s integrated into the music.
This effect can be very deceptive, and it’s one of the reasons why most mixing engineers balance their mixes at a medium or low monitoring level. However, a high monitoring level can be very useful if you want to determine whether you’ve gone too far with the use of compression and limiting. Combined with the compression of the ear, the low frequencies of an overly compressed mix will be substantially attenuated, and the sound image will appear flat and shrill. The way in which the ear’s protection mechanism responds to a sound image is one of the reasons why varying the monitoring level can give you a lot of insights into the characteristics of this sound image.
A protection mechanism is there for a reason, and it’s anything but infallible: the longer it’s activated and/or the louder the sounds, the higher the chance of hearing loss. So listen only briefly at a high monitoring level to gather the information you need, and then return to your reference level.
Fatigue
It’s better not to make the most critical mixing decisions at the end of a long day. Besides the fact that it’s hard to remain focused, there’s also a change that occurs in the ear after prolonged and/or heavy exposure to sound. The level of your hearing threshold (the lowest sound intensity you can perceive) gradually shifts upward. As a result, it will be more difficult to hear soft sounds, and details will disappear from the sound image. Because of this, making decisions about balance and dynamics will become increasingly hard with time. This effect is called a temporary threshold shift.
Initially, a temporary threshold shift won’t make working with sound impossible, but it can cause you to end up in a vicious cycle that will eventually make it impossible for you to work, and in the worst case it will give you permanent hearing loss. When a threshold shift causes the soft details in your mix to disappear, your will need to increase your monitoring level to hear the same sound image as before the threshold shift. But this will make the loud parts of your mix even louder, and it will put an extra strain on the protection mechanism of your ears, which complicates decision-making in mixing. After that, the threshold shift will only increase due to the higher monitoring level, making the problems gradually worse.
What Can You Do to Prevent This?
First of all, it’s important to take short breaks regularly, and to stick to a fixed reference monitoring level (which you can only deviate from for short periods of time). This is the only way to stay focused and reliable if you work with sound eight hours a day. As soon as you notice your ears getting tired (with some experience, you can learn to sense this pretty well) it’s time to stop and call it a day. If you can’t do this because of an approaching deadline, don’t be tempted to increase your monitoring level. Accept that you can no longer hear all the details, and don’t try to solve this by using more compression, for instance. Trust the balance you made before your ears got tired. Hold on to your rhythm of working at the reference level, and use lower and higher monitoring levels only to briefly check your decisions. And after an extremely long day of intense sound exposure you should try to give your ears some more time to recover.
There are tables that show the length of time and level of exposure at which there’s a risk of permanent hearing loss. Each country has set its own standard, and internationally these standards differ by a couple of dB. It’s hard to say exactly at what level sound exposure becomes harmful, and this can be different from person to person. Both the duration and the amount of exposure determine the risk. For example, being exposed to 80 dBA for eight hours is just as damaging as two hours of 86 dBA (see Figure 2.3).
Figure 2.3 Permissible noise exposure. Rule of thumb: reducing the duration by half means you can raise the intensity of the exposure by 3 dB, without increasing the risk of hearing loss. Exposure in the area above the red line can cause permanent hearing loss. The 80 dBA point for eight hours of exposure is based on the European noise standards. Other standards have slightly different values: the American standard sets this point at 85 dBA, for instance. But all the standards fall within a pretty narrow range, and it’s clear that regularly exceeding the standard by more than a few dB is almost guaranteed to lead to hearing loss.
2.3 Your Ear Is an Equalizer
Frequency Response
If you listen to music and you keep turning the volume down, at the very end only a couple of things are still audible: the sounds in the upper midrange. At that level, the bass has vanished into thin air. This is because the ratio between the different frequencies you perceive is directly dependent on the sound intensity at which you perceive them. At a very low level, your ear is much less sensitive to low frequencies than to mid frequencies. Based on experiments, so-called equal-loudness contours have been created: graphs that show at what intensity test subjects perceive each frequency as equally loud. These curves are also known as Fletcher–Munson curves, named after the researchers who first conducted this study in 1933, though they have been slightly refined since then (see Figure 2.4).
Figure 2.4 Equal-loudness contours: all the frequency points in one line are perceived to be equally loud, even though their acoustic energy can differ up to 70 dB. The graph clearly shows that the highest and especially the lowest frequencies are perceptually louder at a higher sound intensity.
A Mix Is Won in the Midrange
The ear is most sensitive to midrange frequencies, not coincidentally the area where most of human speech is located. At higher loudness levels, the sensitivity to high and especially low frequencies increases, while it remains the same for mid frequencies. This continues until the point is reached where the acoustic reflex from the previous paragraph starts to temper the relative sensitivity to low frequencies again. And at the highest loudness levels, the hair cells are damaged (temporarily or permanently), which causes the relative sensitivity to high frequencies to decrease. The most important thing to learn from these curves is that the low and high ranges will always sound different, depending on the monitoring level. If your mix works in the midrange—if all the instruments are distinguishable and nicely balanced in this area—there’s a good chance your mix will sound acceptable at any monitoring level. That’s why many mixing engineers regularly switch to small speakers that only reproduce sounds in the midrange. To judge the right amount of high and low frequencies to complement this midrange balance, it helps to vary the monitoring level of your full-range system. This way, you will notice if the extremes are getting over the top.
Getting Used to Your Own Bad Mix
If you listen to the same sound long enough, your auditory system will become less sensitive to it. It’s interested in perceiving changes, so especially long, sustained bass notes or resonances will automatically attract less attention after a while. For example: you hear the hum of the refrigerator when you enter the kitchen, but after an hour you hardly notice it anymore. This phenomenon is different from adaptation in the sense that it’s just the refrigerator that becomes softer in your perception, not the entire sound image. That’s why it’s called frequency-dependent adaptation. Try to pay attention to it during your next mix: long, sustained sounds may seem loud when they make their entrance, but after a while they become less and less apparent.
You can even adapt to the sound of your own less-than-great mix. The fact that the frequency spectrum is out of balance will automatically become less noticeable if you listen to it long enough. It doesn’t help that you know which individual components make up your mix: you can hear all of them, because after a couple of hours, your brain can sift them out from your mushy mix. Such an unbalanced perception of the frequency spectrum can be prevented by taking regular breaks to allow your ears to recover. Besides this, try not to focus on short sections of a production (don’t use the loop function), but listen to the project as a whole, so you always hear the onset of individual sounds, instead off ‘dropping in’ halfway. The onset of a sound determines a great deal of its perceived loudness, and only when you listen to a production in its entirety can you judge the balance of individual elements in their context. Section 15.4 will focus on ways to stay fresh while mixing.
Hearing Accurately
You can distinguish different frequencies thanks to a membrane in your ear that can move along with a gradually decreasing frequency from beginning to end: the basilar membrane. It’s shaped in such a way that each frequency causes the greatest movement in a different part of the membrane. The movements of the membrane are picked up by hair cells that are connected to a nerve, which transmits an electrical signal when its hair cell starts to move. This mechanism is supported by the brain, which significantly increases the resolution of the system through analysis of the nerve signals.
The mechanism is most effective at low sound intensity levels, because high-intensity sounds will not only trigger the hair cells corresponding to their own frequency, but the adjacent hair cells as well. As a result, the frequency-resolving power of the ear is diminished at high sound intensity levels. This explains why it’s harder to hear if an instrument is in tune at a high monitoring level. So make sure you don’t set your monitoring level too high when you judge the tuning of instruments, the tonal precision of vocal recordings, or the amount of pitch correction to apply.
2.4 Tuning In
So far I’ve discussed the mechanisms your ear automatically activates, outside of your control. In a very different category are the phantom images you create yourself. It happens to me at least once a month that I grab an equalizer knob, start turning it and hear the sound change as expected . . . until I notice that the equalizer is in bypass mode. Because I have an expectation of what’s going to happen, I hear it too. This mistake is symbolic of a bigger problem: when you focus on selective aspects of your mix, you will automatically hear them better. It’s actually possible to ‘tune in’ and focus your hearing on a single instrument within a complex whole.
This is what’s so hard about mixing: you have to listen to the instrument you’re manipulating, but you shouldn’t focus on it so much that you start to perceive it out of context. Therefore, I try to listen to the imaginary position of an instrument in the whole, instead of just listening if it’s loud enough. I try to place it at a certain distance, and less at a certain volume level. If you force yourself to listen to the relationships within the whole this way, it will help you to prevent your perception from drifting off track.
Objectivity
We can easily fool ourselves if it’s in our own interest. If you’ve just hooked up overly expensive speaker cables, you can bet your life that you’ll hear a difference, simply to justify the purchase. The influence of a cable is something you could test scientifically, if you want to have a definite answer. If only it were that easy for the ideas you try out during mixing, because these ideas constantly affect your perception: sometimes you come up with such a beautiful concept that you literally can’t hear that it doesn’t work. There is one powerful remedy for these mirages: time. Taking some distance and then listening again is usually all you need to make an objective judgment.
2.5 A Fixed Reference
When looking for a good mix balance, you want to work by gut feeling. In order to do so, you should be able to rely on a mental picture of what a good balance sounds like: your reference. This reference should be linked to a specific sound level: it’s no use comparing a mental picture that works at a high sound level to a mix that’s coming from the speakers at an average level. That’s why it’s a good idea to set a fixed monitoring level. This is called calibrating your system. Of course you don’t need to work at this level all the time (by now you know how useful it is to change your monitoring level every now and then), as long as you can always fall back on it to compare your mix to the reference in your head.
How high should your reference level be? In any case, it should be a level you can work comfortably with all day, and without tiring your ears. You will feel what’s the right level for you, and then it’s just a matter of putting a mark next to the volume knob or saving a preset. A good starting point many engineers will eventually reach is the fact that pink noise with an average (RMS) level between −20 and −14 dBFS at the listening position produces a sound intensity of about 83 dBC (reproduced by one speaker). Then they will mix at that level as well: the average acoustic loudness of their mix will fluctuate around that level.
The RMS value that you use for calibration depends on the intended dynamics. If you work on pop music with little dynamics and a high RMS level, you calibrate your monitors at this higher RMS level to avoid working at an acoustic level that’s tiring your ears because it’s too high. But if you work on classical music with a lot of dynamics but a much lower average RMS level, you calibrate the monitors at this low level. This method is similar to the calibration system that’s been the norm in the film industry for years, and it also resembles the K-system for music studios, which was later introduced by Bob Katz.
Figure 2.5 If you set your mix at a specific reference level on the loudness meter and match this level with a fixed acoustic level, you will always know what to expect. The loudness level can be measured with an RMS meter, but an integrated LUFS meter that measures the average loudness of your mix as a whole is even better. More information on these meters can be found in section 14.4.
From the perspective of a mastering engineer (which just so happens to be Bob’s profession) a calibrated system is ideal. It provides a solid base for assessing the dynamics and frequency spectrum of a mix, because mastering engineers know exactly what to expect of a certain type of production at their usual monitoring level. For mixing engineers a calibrated system is a useful aid, but it’s not the holy grail. Unlike mastering engineers, mixers don’t work with material in which the internal balances are already (mostly) fixed. You first have to go through an entire process before there can be a mix with an RMS level on a meter that actually means something. But when you get to the final stage of the mix, the system will start to mean something again. For example, it can be a good indicator when you notice that your mix is around the reference level on your meter, but you don’t feel a lot of acoustic impact. Then you’re probably wasting too much energy on sound you don’t hear.
Decibel
When you work with sound, you will come across decibels all the time, and there’s a good reason for this. Because in sound it’s all about ratios, and the easiest way to express these is with a unit of ratio like the decibel. For example, when talking about the strength of a sound signal, you can express this in absolute terms: ‘This signal causes a peak voltage of 2 volts.’ However, this absolute number only means something in relation to the reference of your system. If your system clips at 1 volt, you know you’ll need to attenuate the signal by a factor of two to prevent this. This can also be expressed in decibels: reducing the strength of the signal by half corresponds to a decrease of 6 dB. The advantage of the decibel is that it’s easy to express a huge dynamic range with it. The difference between the faintest and the strongest sound human hearing can perceive is enormous: the strongest sound causes an increase in air pressure that’s one million times greater than the faintest sound. Because the decibel is a logarithmic unit, even this huge ratio can be clearly expressed with it:
So the strongest sound is 120 dB louder than the faintest sound.
Another advantage of decibels is that they make calculations easy. If you first amplify a sound by 2 dB and later by an additional 5 dB, the total amplification is 7 dB. If you had to calculate this with a non-logarithmic unit, you would have to multiply it twice by large numbers.
2.6 Taking Professional Care of Your Ears
Your senses are built to perceive as much relevant information as possible. They adapt to the circumstances. Therefore, your auditory system has a variable character, plus it’s prone to fatigue and overload. If you want to depend on it, you have to give it a chance to do its job reliably:
• Make sure the sound level your ear has to adapt to is not so high that this adaptation permanently distorts your perception.
• Use a fixed reference for your monitoring level that suits the type of production (genre/application) and your activities (recording/mixing/mastering).
• You can change the monitoring level briefly to check specific aspects, but make sure you give your ears enough time to adjust to the new level.
• Take regular breaks, and then listen to a production front to back at the reference level to hear balance mistakes. Always create the final balance at the reference monitoring level, preferably with fresh ears at the beginning of a new day, or after a break.
• Avoid overloading your ears by limiting your working days to eight hours as much as possible, and stick to the reference level with discipline, even when you’re tired.
Besides this, any other way you can come up with to keep some perspective on a project is a bonus. I once read in an interview with engineer Chuck Ainlay that he always has a small TV on the Sports Channel during a mixing session. This would take his mind off the project every now and then, allowing him to keep a fresh perspective on it.