I hate to break it to you, but your ears lie. In many instances, the sound we hear isn’t actually what’s happening. There is a whole area of science dedicated to the anomalies in our hearing called ‘psychoacoustics’, and that’s what we’re looking at today. This isn’t essential knowledge for running your studio or your business, but I find it fascinating, and as sound is our medium it definitely comes under the banner of ‘the more you know…’
In fact, it’s not really our ears that are lying to us. It’s the way our brains interpret the data presented to it. Some of the ‘misinterpretations’ are necessary – our perception of direction depends on our ability to alter the data presented and merge the signals from 2 ears. It’s fascinating how that works, but I’ll let you google that for yourselves. I want to think about a few things that may have implications for your recordings/productions here.
Selective hearing
This first one may not actually be in the realms of psychoacoustics, but it’s a simple to understand illustration of a way our ears can’t always be trusted.
Here’s the scenario. I need to do a location recording somewhere and I ask someone from the venue to find somewhere we can do it. We go to the chosen room, it’s at the far end of the building and no one goes there. There are no noisy things in the room, but the second we sit down their face drops as they hadn’t previously noticed how much noise gets into the room from outside. This has happened to me more than once.
Our brains are very good at ‘selective hearing’ (there are many jokes here about my elderly hard-of-hearing Grandma’s ability to not hear things she didn’t want to be told, but her ability to hear faint whispers about cake was legendary). We can simply ignore sounds that don’t interest us (and here there are jokes about many husbands’ ability to nod in the right places whilst not having a clue what she’s bleating on about). In the situation above, the room is quiet, if all you need is a place you can work without being disturbed, it’s not until you need silence that the noise that is actually there becomes apparent. This is why I recommend one of 2 ways to assess whether a new space is actually quiet enough for your studio.
Method one is to set up your recording equipment and record about an hour of room noise and listen back to that (or look at the waveform and see how flat it is). Our brains don’t ignore sound when we listen to recording in the same way as they do when we’re in situ.
The other method is to sit in the new space for an hour or so with your eyes closed. Doing this makes your ears your main sensory receptor and the noises reveal themselves.
Human hearing range
This is also a good example that most people already know about how our ears defy reality. We often quote the human hearing range is 20Hz to 20KHz, but we know that sound occurs outside those parameters.
Bats and dolphins use ultrasonic sound as echolocation to navigate around. The volume bats have to ‘shout’ at is quite phenomenal. They have been recorded regularly shouting at 135dB SPL (sound pressure levels – a measurement taken in air-bourne sound rather than digital/in system sound. This is a crucial measurement when thinking about psychoacoustics and how we hear things). For reference, a jet taking off is around 140dB SPL. Bats could quite easily deafen themselves when they shout, but they have a clever little adaptation to their ears where they dislocate the bones in their ears as they shriek so that they don’t deafen themselves, and reattach the bones to hear the echoes. But we’re getting off topic.
Sound also happens below the range we can hear (elephants and crocodiles use very low frequencies to communicate since you asked). It’s called infra-sound and it can have very strange effects on us. 17Hz is a particularly interesting frequency. I haven’t space here, but buy me a drink and ask me about the time I stayed in a hotel room that was awash with infra-sound which I presume to have been around 17Hz. I don’t believe in ghosts, but I very nearly did that night!
Our microphones will pick up sounds outside our hearing range, although they’re manufactured to work best within our hearing range. Digital audio at 44.1KHz and 48KHz doesn’t record too high above our hearing range (22.05KHz and 24KHz respectively is the highest frequency these sample rates can cope with) but the lower frequencies can really mess with our processing – particularly compression – if it’s present. Even if we can’t hear the sound we could well hear the undesired effect the processing of very low frequencies has. Another reason to use a high pass filter!
Fletcher-Munson curves
I don’t know why, but I really like this next bit of audio geekery. I like it so much I bought the t-shirt. Fletcher-Munson curves are also known as curves of equal loudness and they demonstrate that our ear not only has a far from flat frequency response, but that it alters massively depending on how loud the sound is. Along the X-axis is the frequency of sound(s), and the y-axis is a measurement of loudness.
I think Mr Fletcher and Mr Munson (or it could have been Mr and Mrs Fletcher-Munson – I can’t be bothered to look it up) used dB SPL, but other people have replicated the experiments with concurring results but used phons or sones rather than dB. I’m not going to bore you with what phons and sones are other than to tell you that they’re different ways of measuring sound levels. The way they worked the experiment is that they played a tone at a certain volume to a volunteer, and then they played a second tone of a different frequency and asked the volunteer to make the second tone the same volume as the first by ear alone. The graph plots how much the volumes needed adjusting to make the different frequencies appear to sound the same. What we conclude is that at the edges of our hearing range we don’t hear as well as near the middle of our hearing range.
Not only that, but at lower volumes the lower and higher frequencies need turning up a lot more than at louder volumes. We also see a dip at 2.5 – 5KHz. This is known as the presence range and is the range of frequencies where the bulk of the human voice sits. We have twice as many cilia in our inner ear at this range, so it’s not a surprise that this range needs turning down to make it seem to be the same volume as other frequencies.
What this means for us is that when we’re doing a mix we need to listen at a variety of volumes to make sure we’re getting the best result possible. Not only does this help prevent our ears from getting fatigued, but it means we’re checking across the different volume levels as well and ensuring we’re not falsely compensating for a lack in a certain frequency range when actually it’s just because of the level we’re listening at.
Critical bandwidths
One other thing to tell you about, but before I do that there’s a little bit of a sub-point. It’s to do with how our brain handles sound. It’s difficult to understand, so I’ll explain it like this.
Imagine it’s one of those days when you have a very complex set of jobs to get through. You don’t just jump straight in and tackle them all at once, you break them down into smaller jobs and tackle them in small chunks. This is a terrible illustration, but that’s sort of how our ear works. It breaks sound down into what are called Critical Bandwidths. Each bandwidth is a third of an octave and they absolutely govern how we hear.
If you take 2 sounds and reproduce them so they measure at the same volume (dB SPL) one will sound louder than the other. It will be the one that covers more critical bandwidths that will sound louder. The often quoted example is that a jet engine will sound louder than a single tone of the same dB SPL. But the loudness factor isn’t what I want to mention here.
Masking
The point I want to raise is a practical point for if we’re doing a mix. If you have 2 sounds that occupy the same critical bandwidths we’re in trouble! If one is louder than the other, you won’t hear the quieter one. This is called Masking. Both sounds are still present, but because they occupy the same space in the frequency spectrum the quieter one will simply disappear.
So what do we do? We shift one or both of them. It may be that the effect is desirable in the project we’re working on and maybe a bit of ducking to let the quieter one punch through at key moments is what we need. But more than likely what we’ll have to do is EQ the sounds so the critical bandwidths are shifted and the quieter one cuts through the mix and takes its place.
Shepard tones
One final thing as a reward for getting through this far. This isn’t in the realm of psychoacoustics at all. This is an auditory illusion called Shepard Tones and is an audio file with an ever-rising (or falling) tone.
Give yourselves a pat on the back for getting through this. Give yourselves extra pats for every point you understood!
Cover image photo by kyle smith on Unsplash