MTN Weekend: Your ears may be confusing your spatial audio
In 1998, a group of Dutch researchers gave people new ears. Not surgically, fortunately. They placed custom-made plastic moulds inside the folds of the participants’ outer ears. The ear canals remained open, but sounds now encountered a subtly different landscape on their way in.
The effect was immediate. The volunteers could still judge whether a sound came from the left or right, but struggled to place it above or below them.
Over several weeks, their performance improved substantially, even though the moulds had not changed. Their brains were gradually learning the unfamiliar acoustic patterns. When the inserts were removed, the volunteers could use their original ears normally again.
The experiment, described in Nature Neuroscience as “relearning sound localisation with new ears”, revealed something important about human hearing. The visible folds on either side of your head form part of a positioning system that your brain has been calibrating since childhood.
This helps explain why spatial audio can sound uncannily convincing to one person and rather muddled to another. Personalised spatial audio tries to solve that problem by adapting headphone sound to the shape of the individual listener’s head and ears.
How your ears locate sounds
When a dog barks to your right, the sound reaches your right ear fractionally earlier and, particularly at higher frequencies, more loudly than your left. Your brain compares these differences to determine where the bark came from.
Locating a sound above, below, in front or behind requires additional clues. As sound passes across the curves and cavities of the pinna—the visible outer ear—some frequencies are strengthened while others are reduced. The resulting pattern changes with the angle of arrival.
Your head and upper body affect the sound too. Together, these anatomical effects are described by a head-related transfer function, or HRTF.
What is a head-related transfer function?
An HRTF is a mathematical description of how sound travelling from a particular direction is altered before reaching each ear. Because every person’s anatomy is slightly different, everyone has a somewhat different HRTF.
Spatial-audio systems use these acoustic filters to reproduce directional clues through headphones. By applying carefully calculated changes to the left and right signals, they can persuade the brain that a helicopter is flying overhead, footsteps are approaching from behind or a voice is coming from across the room.
At least, that is what should happen.
Why spatial audio sounds different to different people
Many headphone systems traditionally used a generic HRTF, derived from measurements of a person or an acoustic mannequin.
That works reasonably well for plenty of listeners. Human heads are broadly similar, after all. But the finer details vary enough to influence the illusion.
It is a little like borrowing someone else’s prescription glasses. You may still see the room, but it will not look quite right.
With a poor HRTF match, sounds intended to come from above may remain stubbornly at ear level. Front and back can become confused. Instead of expanding into a convincing three-dimensional space, music may seem hollow or artificially processed.
Dolby Atmos, HRTFs and head tracking are not the same
Part of the confusion arises because several technologies are sold under the broad label of spatial audio.
Dolby Atmos is an immersive audio format that allows a mix to contain sounds positioned in three-dimensional space. A playback system renders that mix for the available speakers or headphones.
For headphone listening, an HRTF helps convert those positions into suitable left and right signals. Head tracking, meanwhile, detects movement so that the virtual screen or soundstage can remain in place when you turn.
Atmos supplies the spatial information. The HRTF adapts it for headphone listening. Head tracking stabilises the illusion.
How personalised spatial audio scans your ears
Personalised spatial audio replaces the generic model with a profile intended to resemble the individual listener more closely.
Apple’s system uses the TrueDepth camera on a compatible iPhone to capture the front of the head and a view of each ear. It then creates a personal profile for supported devices including the AirPods Pro, AirPods Max and Beats Studio Pro. According to Apple, the camera data used to develop the profile is processed on the device and the images are not stored.
Sony offers ear-shape analysis for 360 Reality Audio through supported apps and services. Users photograph both ears so that playback can be optimised for compatible products, including models such as the Sony WH-1000XM6. Availability depends on the music service, app and device, so compatible headphones do not automatically personalise every form of spatial audio.
Neither process produces a complete laboratory measurement of your individual HRTF. Consumer systems use visible anatomical information to estimate a suitable profile or select a close match from existing acoustic data. The exact calculations are largely proprietary.
Researchers are also developing machine-learning methods that extract measurements from ear images and use them to estimate individual HRTFs. The goal is to approach the benefits of specialist acoustic testing using little more than a smartphone camera.
Does personalised spatial audio work?
A well-matched profile can make virtual sounds easier to locate. It may also improve “externalisation”: the feeling that a sound exists somewhere around you rather than inside your head.
The effect is not equally dramatic for everyone. The recording, playback app, headphones, earbud fit and individual hearing all influence the result. Personalisation also cannot rescue a poor spatial mix.
Before buying new headphones, check whether your current combination of phone, headphones and streaming service already supports the feature. Perform the scan in good light and move hair away from your ears.
For a useful comparison, choose a film or game scene containing dialogue in front of you, movement behind you and a sound from above. Listen with personalisation on and off without changing the volume. Louder can easily be mistaken for better.
If you use earbuds, check the fit as well. A poor seal alters the sound before any sophisticated processing begins. Trying another size of replacement ear tips may accomplish more than repeating the scan.
Some listeners will hear the virtual scene click into focus. Others will notice only a subtle improvement or prefer the standard setting.
There is no universally perfect spatial-audio profile because there is no universal pair of ears. The cleverest component in the entire system may not be the headphones, phone or algorithm, but the two peculiar structures your brain has spent a lifetime learning to understand.
This article contains affiliate links. If you purchase something through one of these links, Move The Needle may receive a small commission at no additional cost to you.