Why we should care about ergonomics in virtual reality
(Originally published as guest article for Skarredghost, August 2018)
“I’m at the technology store today in my local shopping mall because I want to get a new VR headset.
The latest model by “Virtual Dream” was released a few weeks ago, but I’ve had to wait until the end of the month for payday. I approach the sales assistant manning Virtual Dream’s in-store concession, he sees me coming and greets me with a friendly “Good morning!”
He’s already noticed I’m looking keenly at the new headset inside the display case to the right of his booth, and knows what I’m going to ask next.
“Have you got time to do a fitting?” I excitedly blurt out, and he nods enthusiastically before replying, “That would be a pleasure, however, it will take about twenty minutes to print your face mask if that works for you?” I tell him that’s ideal as I need to get some breakfast, so I’ll come back and pick up my headset once I’ve eaten.
He asks how I’d like to pay, and I swipe my phone across his terminal, taking a third of my pay cheque. “Okay, if you can come around to the left side of my booth we can get you measured up”, he asks, “You might have seen this online but it’s pretty cool to try out in person” he adds.
On the counter of his booth is a small camera attached to a circular support frame, with a padded chin rest. He gestures for me to lean forward, and after I place my chin on the rest, he lowers the frame over the top of my head. “Please keep as still as possible, and make sure to keep your eyes closed” he suggests, before pressing a button on the rear of the camera. I hear a gentle whirring as the camera moves around my head, and less than ten seconds later, my 3D scan is captured, “That’s all done, we got a great capture here. Okay, your circumference is 58 cm so we recommend a medium, and your head shape is ideal for the Y-style harness”. ”
He picks up a tablet computer from the counter, and makes a few keypresses before turning back to me, ‘That’s all done, if you want to come back in about twenty minutes I’ll have it boxed and ready to collect.’ As I walk away to find my breakfast, I hear the 3D printer starting to do its work.“
I returned my ******* because of the discomfort and bought a ****. The ******* didn’t fit properly and hurt my forehead after 30 min. Really wanted to keep the ******* but if I can’t wear it, there’s no point
The *** was super uncomfortable for me and gave me a headache, whilst the original **** was very comfortable for me, especially once I got the ***.
I literally couldn’t go more than half an hour on the ****** without getting a nasty headache, so I can’t say if you will like either, you should try and demo one first before buying?
The **** just made my head hurt, and I found the controllers too small (I have large hands), but the **** fitted great and the controllers were way more comfortable.
These are all real responses on a virtual reality forum after posters have asked: “Which headset is best?”…only the names have been changed to protect the innocent headsets. One person recommended headset X because they found it fitted them well, whilst headset Y was a poor fit that they found painful to use during longer gaming sessions. A second person has replied with an answer that is the opposite; they thought headset Y was fantastic and spent hours using it, whilst headset X had them quickly reaching for the pain killers. So who is right?
Actually, both of them are right, based on their “headset fit”, or more specifically their “individual craniofacial anthropometry”, the technical term for the measurements that are used to describe the human head and face. When you start looking closely at craniofacial anthropometry, you will see a wide range of variance and facial asymmetry, as there are few if any people with perfectly symmetrical faces. And this is where things become more complicated when designing consumer equipment like virtual reality headsets, as there are few other products on the market that are worn so intimately against the face whilst secured to the head.
Other examples are scuba diving masks or full-face motorcycle helmets, although neither of these have the added complication of housing a head-mounted display that relies on a good fit for the correct optical presentation of the display to the eyes.
Historically, virtual reality systems were prohibitively expensive and were limited to being used in controlled conditions by a small number of researchers, scientists, and groups like astronauts. It was typical to customize headsets to suit these users, just as space suits were made to fit specific astronauts. Additionally, the headsets were used for task-specific work purposes, where comfort would take a back seat to the task in hand, and the duration of sessions was often short.
However, following the release of consumer equipment like the HTC Vive, Oculus Rift, Sony PSVR and the various mobile VR headsets, we now have a different situation with a much larger group of people (the general population) using virtual reality on a regular basis, for longer periods of time. The larger this group of users, the more craniofacial variation is found, which has a direct impact on the quality of the virtual reality experience and whether users continue to use their headsets on a regular basis.
A “good fit” is comfortable to wear and ensures that the optical presentation is correct to truly immerse the user in their virtual world. A “bad fit” is uncomfortable (can be painful) and often causes a poor optical presentation which breaks immersion. A bad fit is simply a bad fit and not something that can just be easily adjusted away with headset straps or fiddling with the IPD (interpupillary distance) adjuster. If it doesn’t fit properly, it won’t ever fit properly, like a pair of ill-fitting shoes that don’t get better with time but continue to cause problems.
A bad fit can cause physical fatigue to the soft tissue of the face and hard bones of the head, creating stress which makes the experience unpleasant, and can contribute to a tendency towards motion sickness especially when combined with heat inside the headset. A bad fit with poor optical alignment has a negative effect on the feeling of “immersion” and “presence” as it interferes with the suspension of disbelief, by constantly reminding the user they are wearing a headset – its hard to relax into your experience when one of your eyes is slightly out of focus, the stereoscopic effect is reduced or you have noticeable lens artefacts in one eye and not the other.
A number of headsets currently on the market have no IPD adjustment, or software-based IPD adjustment (with a fixed lens), neither of which give suitable optical alignment unless you are lucky enough to have the same IPD as the headset. As an example of IPD variation, in men the 5th percentile for IPD is 58.5mm, the 50th percentile is 64.0mm and 95th percentile is 70.0mm, therefore the mean (average) is 64.0mm. Using a headset with the wrong IPD can be “interesting” to say the least; fixed IPD headsets should probably not be on sale as it’s detrimental to the user and their experience of VR.
The end goal when designing Virtual Reality equipment is “complete transparency” where it fits so well, you don’t notice you are wearing it; you simply relax into the virtual world with true immersion where you feel “present”. As we start to see higher resolution displays coming to market, the notorious “screen door effect” will diminish, and more sophisticated lens designs will reduce or eliminate optical artifacts like “god rays”, meaning that good fit will become ever more important so as to not become a barrier to true immersion.
An interesting aspect of our ability to accommodate new experiences is a psychological trait called “bohemian adaption”, where something novel soon becomes the new normal, causing us to seek further novelty as we become dissatisfied with the normal. A side effect of this is our ability to overlook flaws becomes diminished once novelty wears off, something many VR enthusiasts have experienced once they become accustomed to their equipment and start noticing the flaws, which start breaking immersion.
The headset that had the “wow!” factor soon become irritating because it puts pressure on the forehead, their left eye always seems a little out of focus, their left hand gets a cramp when using the controller, or they keep getting entangled in their tether. Good ergonomic design seems to reduce or eliminate these concerns, and good ergonomic design relies on good data.
When designing equipment, an “Ergonomist” (human factors design specialist) will look at anthropometric data. To suit the general population we’d look at a range from the 5th percentiles to the 95th percentiles of the dataset, which will then suit 90 percent of the population, and accept that people outside of these percentiles will not be catered for by that design. This anthropometric data has been gained by measuring huge numbers of people over a number of decades through large studies by many organizations including those in the medical profession and the military. As an example, a functional arm reach for “5 percentile female” is 73.5cm whilst for “95 percentile male” is 94.2cm. This difference in reach of over 20 centimeters is one example of the wide range seen in the human population, and variations are also present in craniofacial measurements.
As the old saying goes “You can please some of the people some of the time, but not all of the people all of the time”.
The aim is to please as many people as possible unless you are aiming your product at a specific group of people using a specific dataset, as there are different sets of anthropometric data to target different groups of consumers. Some examples would be general datasets for Men, Women, Adults, Children, or more specific datasets, for example, Caucasian or Asian. Appropriate anthropometric data is a critical ingredient for good ergonomic design, but data can be limited for smaller population groups which have not been widely studied.
Companies like Oakley manufacture their sunglasses in “Regular” fit and “Asian” fit, as do a number of bicycle helmet manufacturers. You will also see different “foot forms” used when designing shoes for a Germanic foot, Celtic foot or Greek foot, just to name a few of the different foot shapes found across the human population. The companies that manufacture VR headsets choose a particular dataset (head model) when designing their headset, and will make design choices or compromises to try and suit the largest possible user base of their target market.
If you fall outside of this dataset, or have unusual features (as many of us actually do), you may find your headset is uncomfortable, slightly blurry or even painful to wear. This explains why a VR headset designed for the Asian consumer is often a poor fit for a Western consumer unless you actually better suit the fit of an Asian headset.
You can now start to understand why there are different opinions on VR headsets, and why it’s difficult to recommend a particular VR headset to a friend, colleague or stranger on the internet based on your own experience of wearing that headset. A problem with the different “fits” that the current headsets have, is that you may want to choose a particular model based on its technical features, access to its content, or wish to support that company, but you try/buy that headset only to find it doesn’t fit you properly.
Within each headset design, there is only a little that can be done to adjust fit, apart from some of the radical modifications people make. These tend to be “band-aid” fixes that only partially resolve fit issues to make that headset into something that can be tolerated, other fixes involve severe modifications that void the warranty which is less than ideal on an expensive headset that is only 2 months old. And sometimes the official upgrades can make the fit worse, for example, some Vive users bought the DAS only to find it much less comfortable than the original soft harness.
You might have heard the word “ergonomics” when people discuss virtual reality headsets, such as “that headset had poor ergonomics!”. It’s a slight misuse of the word, as the definition of ergonomics is:-
Ergonomics (or human factors) is the scientific discipline concerned with the understanding of interactions among humans and other elements of a system […] in order to optimize human well-being and overall system performance” (International Ergonomics Association).
Ergonomics is also known as “biotechnology”, “human engineering” or “human factors engineering” and has played an important role since the introduction of desktop computing in the office, where workers often spend hours doing repetitive tasks with a limited range of movement.
An Ergonomist is a specialist in the applied science of equipment design, intended to maximize productivity by reducing operator fatigue and discomfort. As we move into the realm of immersive computing, the focus of the Ergonomist will shift from the desktop computing environment to the immersive computing environment which we access through virtual and augmented reality headsets. This in itself presents a new set of challenges, ranging from equipment design to user interface design, and physical movement limitations inside virtual environments.
So what can be done to ensure headsets (and hand controllers, which we have not even mentioned) are comfortable? It’s a tricky problem to solve, as during the early adoption of consumer VR the development budgets are limited by the small size of the market. This doesn’t leave much room to spend money on tackling difficult ergonomic challenges by offering a headset model with different “fits”, or providing much in the way of customizable or modular headset designs.
Magic Leap’s first consumer AR headset has just come to market and shows some interesting thinking as its’ offered in two sizes, as well as each size coming with a “fit kit” offering a number of shaped pieces for the nose and the head padding. This headset is different to VR headsets in that the waveguide technology does not allow for moveable displays to adjust IPD, so they’ve had to offer 2 headset sizes to accommodate users with a smaller IPD range and users with a larger IPD range. However, if we run with this thinking, it could be possible for a manufacturer of VR headsets to offer different sizes, just as bicycle helmets are often available in small, medium and large sizes.
The current “one size fits all” model of the headset is a bad compromise, just as “one size fits all” cycle helmets rarely satisfy anyone, resulting in a loose fit or pressure points as the adjustment system cannot accommodate such a wide range of heads. Each individual size of bicycle helmet has further adjustment using a ratcheting, radial harness system to really fine-tune the sizing; a medium cycle helmet will usually adjust from 55cm-59cm, as well as offering 3 volume settings using a rear vertical adjuster.
In addition to sizing options for VR headsets, different style of headset harnesses could attach to the same headset base to accommodate users with different shaped heads, especially to suit the occipital bone (the pointy ridge at the back of your head).
An additional design solution is to offer a user-specific, custom facial interface (face cushion). A custom facial interface helps accommodate the asymmetry that is found in our faces, and the direct influence this has on the optical presentation when wearing a headset, and user comfort. Current headsets are designed so that the face is centred within the headset, and the IPD adjustment works off this premise so that an IPD set at 64mm has equal spacing of 32mm left and 32mm right. But what happens when your face is not centred within the headset because of your facial asymmetry, and your left eye is actually further from your nose (34mm) than your right eye (30mm)?
Take an image of your face, and draw a vertical line straight down the centre, then measure the position (width) of each eye relative to this centre. Repeat the exercise by drawing a horizontal line across your face, and measure the position (height) of each eye relative to this line, the results are often surprising.
Further analysis shows variations between the depth of each eye within its socket (orbit), the volume of the cheekbones (zygomatic or malar bone), the shape of the forehead (frontal bone) around the brows and reports of 19 different nose shapes. We also see optical variations with eye dominance or eye-specific refractive errors requiring specific adjustments (it’s not ideal to wear glasses inside a headset). These physical variations can be accommodated by a custom face cushion, whilst optical variations can be catered for by automatic IPD adjustment and focus adjustment using eye tracking systems to cater for each eye individually. We are starting to see new headsets coming to market with these automatic optical adjustments, which is very encouraging.
Currently the aftermarket face cushions you can purchase for headsets like the Rift and Vive are offered in different thicknesses and types of material, but follow the same concept of the OE (original equipment) face cushions of a symmetrical shape, equally spaced for left and right. These don’t allow any asymmetrical accommodation, which means one side may be more comfortable than the other, this is seen in “witness marks” as the bony facial structures compress the padding used in the cushion. What is required is a method of supplying a face cushion which properly fits that individual’s face, and ensures they align correctly with the optical display.
One solution we have been studying is the emerging technology of 3D face scanning, for example, the Bellus3D Face Camera Pro, which is plugged into an Android smartphone. This combines 2 proprietary state-of-the-art technologies that measure 500,000 3D points on the subject’s face creating a very accurate high-resolution face model in seconds. We could take this face model and integrate it with a design model of the headset to generate a data file for a custom face cushion. This could be 3D printed using hypoallergenic, self-wicking, antibacterial materials to provide a skin-sensitive layer to prevent any irritation, reducing facial stress and heat build up. A layered approach with a memory foam base could ensure a supremely comfortable fit!
The same technology can also be used to scan the head to assist in selecting the correct size of the headset and correct style of headset harness to suit the head shape. Once installed in the headset, the custom face cushion and correct headset harness will give the user a truly customized fit addressing their craniofacial asymmetries, whilst eye tracking systems adjust IPD for asymmetry and depth focus on a per eye basis. As described in my fictional story at the start of this article, in the near future we could see virtual reality and augmented reality headsets offering an individually customized fit that we believe is required to truly get the best from the experience.
VR enthusiasts often focus on technical specifications, obsessing over resolution, pixels-per-degree, refresh rates, tracking systems…perhaps forgetting the most important element of the equation is the human interface. As one colleague commented last year, the clue is in the name “headset”, a “set” you wear on your head.
As consumers, we can purchase custom footbeds for running and cycling shoes and heat molded ski boots, which have been successfully sold through retail outlets for a number of years. With some forward thinking this business model can be incorporated into headset sales, with consumers only needing a “one-time” capture which could be applied to whichever headset brand they purchase in the future, similar to the data your optician holds on your prescription.
We look forward to the immense benefits to be gained from customizing our immersive computing headsets, to provide the best comfort and optimum optical presentation so we can relax in our virtual worlds, at least until its time for a bathroom break!
© 𝑖𝑚𝑚𝑒𝑟𝑠𝑖𝑜𝑛 𝑚𝑒𝑐ℎ𝑎𝑛𝑖𝑐𝑠 𝑙𝑡𝑑 2023