Categories
Experiments

Valve ear speaker teardown

Ever wondered what was inside those neat off-ear speakers on the Valve Index?

“Balanced Mode Radiator” (BMR) ear speakers use custom drivers made by Tectonic for Valve. Valve list their audio solution as having these characteristics:-

Built-in: 37.5mm off-ear Balanced Mode Radiators (BMR), Frequency Response: 40Hz – 24KHz, Impedance: 6 Ohm, SPL: 98.96 dBSPL at 1cm.

In use onboard the Valve Index headset, the BMR ear speakers are unrivalled in terms of sound quality and sense of spatial soundstage for VR headsets, Emily Ridgway and her team at Valve certainly worked some magic here!

After experimenting with the BMR ear speakers and different audio headphones, I kept using the ear speakers as their excellent audio combined with quality of life (off-ear, on-board) was a great combination.

During the past year I did need to RMA a number of ear speakers, Steam support were very supportive and shipped them all as advanced replacement, and didn’t ask for the defective ones back. So I obtained some spares…

2 problem developed:-

1. Unwanted speaker movement. This seemed to worsen after lots of active gaming in Pistol Whip and Best Saber. Over time the ear speakers stopped holding the set position, drooping during a session or sudden movement. It appeared that the mechanism spring force degraded over time/use.

2. Vibrating. A slow developer but eventually the speaker pods started to vibrate at higher volumes or on bass hits. Not a malfunction of the driver but the physical connection between the speaker pod and speaker arm. This is felt as a looseness (slop) with light finger pressure, it’s easy to wobble the speakers pods.

Despite these problems I continued using the BMR ear speakers and gave Valve some feedback to help with further iterations.

How do they attach?

These attach to the Valve Index headset using a circular ‘pogo pin’ mounting system retained by a single torx T6 bolt through the headstrap.

Pogo pin mounting socket on Index headstrap

I decided to teardown one of my faulty BMR ear speakers to have a good look inside.

BMR Ear Speaker removed from Index, ‘Pogo Pin’ mounting system on the right
Carefully removed plastic cover (glued), I pushed it open using a flatblade screwdriver through the open slot for the height adjuster
Sliding height adjuster at minimum, power cable routing accommodates movement of height adjuster
Sliding height adjuster at maximum
Pogo pin mechanism taken apart showing pogo pin springs, bolts, sliding plate and circular pogo mounting

Looking closely at the “pogo pin” system, it’s cleverly designed using the springs to apply pressure to the pogo pins (to ensure contact with headstrap audio pads) and also allow vertical adjustment of the speaker with enough resistance to prevent unwanted movement.

Perhaps these springs are stretching over time/use, as springs do tend to stretch, to a reduced clamping force allowing the speaker pod to droop. It may be possible to tighten the small bolts to increase spring pressure, or pad the spring with steel washers to achieve the same.

The Circlip in the image above locks the speaker pod axle to the speaker arm.

This axle socket has gone sloppy (flogged out) on several of my ear speakers allowing the speaker pod to vibrate at higher volumes or during bass heavy audio. I’m unsure how this can be resolved without a different type of fitting, or perhaps a polymer bushing.

Circlip removed from axle
Speaker pod removed from arm, at this point I cut the power cables
Carefully working blade around speaker basket to break glue seal
Finally! Speaker assembly coming apart…
Wire basket removed, showing foam damper covering rear of driver unit
Wire basket removed from ear speaker pod
Thick foam damper from rear of ear speaker
Inside face of foam damper with moulding detail
BMR Ear Speaker pod stripped as far as possible, the plastic moulding was heavily glued to the front ring and resisted considerable force
Power wires to driver hanging inside speaker assembly
Detail showing driver and some type of baffling
Carefully cutting away the plastic moulding to access the driver
Driver core and baffling
Outside wire basket and diaphragm of driver
Cut apart driver showing magnet, copper coil, power leads
Close-up of coil

More information on the Valve Index audio is found in this blog article

https://immersivecomputing.org/2020/02/25/valve-index-ear-ergonomics/

Thanks for reading! Rob Cole.

Categories
Experiments

Experiments with ette…

“ette is the first of its kind finger-tracking controller. Powered by TG0’s patented technology, etee allows user to control VR without gloves, camera, or other encumbering equipment. Etee is lightweight and intuitive to use. The controller has a battery life of 8 hours that enables the user hours of time to build, explore and immerse themselves into the world of VR”

Product Description, TG0

After reading Tony’s (a.k.a Skarredghost’s) recent review of the ette controllers I was intrigued and wanted to know more – especially in light of some ergonomic concerns he raised in his article:

This was a comment I posted on his blog post after reading the review:

“Hi Tony, Great review, very detailed. The ergonomics aspect is an interesting problem, due to wide variation in hand sizing. Typically we have 3 blunt measurements:

1. length: measured from tip of the longest finger to crease under the palm.
2. breadth: measured across widest area where fingers join the palm.
3. circumference: measured around the palm of the dominant hand below the knuckles excluding the thumb.

Creating a gripped or held object to cater for the wide variation is an interesting challenge.

This challenge effect many objects whether pistol grip on weapon, household item like cutlery or cooking tool, and of course VR controller; the Valve Index controllers (Knuckles) have the ability to use clip-on Palm boosters.

I’d be interested to try the TG0 Etee controller, without getting my hands into them it’s impossible to make any valid judgement; however some physical adjustment within their structure (a clip on spacer or sliding component with locking) may be beneficial to optimise the fitting for different hands.” 

Happy XR! Cheers. Rob Cole.


Tony introduced me to TG0 who are also based in London, UK. They quickly arranged a courier delivery and within a couple of days a small box containing an Etee dev kit had arrived; many thanks to TG0.

Packaging was neat and minimalist, although I must admit I damaged the cardboard box trying to remove the controllers as their handles were very firmly wedged into the foam. I resorted to pulling the entire foam slab upwards to release them from the box, which freed the controllers but also separated the adhesive tape holding the box together.

Physical examination

The etee controllers are wonderfully simple, yet very sophisticated in terms of material technology compared to the mechanical switches and sensors we commonly see in motion controllers.

TG0’s “secret sauce” is their patented material technology used to build the sensing strips, which I understand is already used in different commercial applications – this being their first foray into motion controllers for XR.

TG0 list these as the key features of their “Etee” motion controllers:-

-Multi function TG0 thumb-pad technology

-Wireless connection

-5 finger 100 level of sensing

-6hr continue using battery life

-Gesture sensing

-Magic Trackpad with swiping, scroll, rotation and pressure sensing

-Soft silicone shore A 60 touching surface

-Weight: 75g

-Size: L136 x W60 x H30 mm

Despite coming in a little heavier than advertised at 88g vs.75g, a single Etee controller is still considerably lighter than one of my Valve Index controllers which are 197g each.

I’m also using 3D printed Valve Index ‘Palm Booster’ clip-ons, which I have to use to put controls at correct “reach” and to increase controller body “volume” to suit my medium sized hand fit, but this takes a single Index controller up to a weighty 224g.

Releasing the Palm Booster was a smart move by Valve’s designers to widen the range of hand fits possible on a single device, and many Index owners placed orders with 3D printing shops soon after Valve released all the Index mod files for free under the Creative Commons License.

From nearly a year’s experience of using “naked” Index controllers and “Palm Booster” Index controllers, the clip-on added an important tactile enhancement as the additive printing process creates a warmer, more textured surface.

This helped balance the “material mismatch” I’d experienced as a proprioceptive challenge between the soft, warm fabric strap and cold, hard plastic controller body in the weeks following the Index launch.

As well as the weight difference between the Etee and Index controllers, there is a big difference in physical size as can be seen in the side by side image below:

The Valve Index is used for comparison here as its the most advanced VR motion controller currently available to the consumer, whilst Etee offers a new approach to hand input championing a revolutionary material technology and different way of thinking.

Unlike many controllers past and present, Etee lacks physical buttons or triggers, instead providing full finger tracking across a rubberised controller body – an oval shape 26mm deep facing the palm (palmer side) and 32mm wide across the hand, combined with a large tracked thumb pad up top.

The low durometer (60a shore) controller skin has moulded vertical ribs which curve around a glowing status window, giving a soft, tactile feel with a small moulded ridge to separate the index and ring finger.

This moulded ridge protrudes 6mm from the surface, what I took to calling the “index” ridge (not to be confused with Valve Index!) to assist in locating and separating the index and ring fingers.

Controllers are marked “L” (left) and “R” (right) at the top of each handle as they are orientated specifically for each hand.

Perhaps the standout piece of the controllers is a rose gold coloured “finger bar” which provides a stiff brace for a sculpted foam cushion which is glued with adhesive tape to a channel along the inside face of the bar.

This foam bar “sandwiches” the finger against the rubber controller body, effectively capturing Etee to the hands without requiring any adjustable straps or safety lanyards which is a good “quality of life” improvement if regularly putting on and taking off.

This foam cushion is punched through with 8 triangular holes which increases its flexibility, perhaps helping to accommodate larger fingers. The foam is a relatively high density, possibly a requirement to maintain structural integrity as its only 12mm wide and 10mm deep at maximum.

The foam also has a forward extension protruding 9mm at the same vertical height as the “index” ridge on the controller skin, although rotated approximately 30 degrees around a vertical axis to the side of each ridge.

This locates the index finger of each hand and secures firmly in place by encircling that finger, whilst the remaining fingers are left “open” especially the little finger (pinky) which floats about unencumbered.

FITTING NOTES

My first quick setup attempt at fitting Etee was strange, being in a rush I pushed them on horizontally through the finger bar foam, finding a comfortable position across my “middle phalanxes”. I had tried pushing the controllers further up my fingers but found the opening between the body and foam a bit too small, I was concerned to snap or bend the finger bar if I forced them.

They felt odd but seemed to fit in some way, until both Tony and TG0 pointed out from a photo I’d emailed them, that I had them incorrectly fitted! So much for reading the instructions….

The hand anatomy image below shows the names of the 27 different bones that make up each human hand. From this image we can see the Middle Phalanx where I’d incorrectly first fitted them, and the Proximal Phalanx which was the correct location.

Image: Paul Jarrett, Murdoch Orthopaedic Clinic

TG0 advised me to fit by “sliding” the controllers vertically down over stretched hands, aiming for the foam to make contact with the proximal phalanxes of the index, middle and ring fingers. I watched the videos on their website again and it started to make more sense.

I tried fitting them again, and despite being a tight squeeze that felt sure to rip the foam off the finger bar, I finally got them on and gave my fingers a quick outstretched wiggle which felt very liberating as I now had “hands free” controllers firmly clamped in place.

I also found it a lot more comfortable to remove my wedding ring as this was being uncomfortably squashed into my ring finger during the fitting attempts.

SETUP

Setup was very easy with a USB dongle, an email I had earlier received from Tg0 contained a download link for their software.

These dev kit controllers have 3 degrees of freedom (3DoF) tracking like the older Samsung GearVR and Google Daydream headsets and remote controllers.

This means only rotational motion is tracked around a fixed location:- pitch, yaw, and roll, but not able to move forwards, backwards, side to side or up and down.

All contemporary PCVR systems and the Oculus Quest stand-alone use 6 DoF tracking for headsets and controllers; 6DoF allows movement forwards and backwards, up and down, left and right (translation in three perpendicular axes) combined with the rotation (pitch, yaw, and roll) of 3DoF systems.

I was interested to see how 3 DoF motion controllers would mesh with a 6DoF headset, in my case the Valve Index. My previous experience of using two different tracking systems together was limited to brief sessions with Lenovo’s Mirage Solo and HTC’s Vive Focus.

Both of those devices 6DoF headsets felt comfortable and immersive but compromised by a “fixed point” 3DoF controller which acted like a strange laser pointer fixed to my hip.

Is 6DoF possible?

It is possible to attach an HTC Vive Steam VR tracking puck to a special bracket, but after speaking to TG0 they advised me to wait for the new 6DoF SteamVR version of ette which is due later this year.

For the purposes of these experiments I limited myself to using the Etee visualiser whilst viewing the desktop through my Index headset, as I was more interested in the direct “hand feel” of using the controllers rather than interacting with virtual worlds.

After charging the controllers using 2 USB leads provided with the dev kit, I installed the software and took them for a test drive after starting steamVR alongside the Etee visualiser.

Calibration and rotation are easily handled through the Etee application, with a simple keyboard input starting each process, and on-screen instructions which indicate the position the controllers should be orientated against.

After getting them calibrated and setting rotation, I started my journey…

SO WHAT ARE THEY LIKE?

They feel surprisingly light in the hands, though unstable as they roll inwards towards the palm as the fingers are closed.

For quick comparison I looked at what happened when squeezing an Index controller and what happened when squeezing an Etee controller

Because Etee is clamped to the upper bones (proximal phalanxes) of the fingers rather than across the palm like index, closing my hands causes the controller to rotate as my phalanxes change angle, a large gap of 30mm before making contact with my palm.

This also causes the thumb to pivot backwards, if using the thumbpad whilst trying to close your hand! The image below shows the immediate difference in how the Etee and Index controllers are clamped to the hands.

Moving my fingers produced an immediate response in the Visualiser application and I started noticed haptic feedback loudly buzzing inside the controllers as “gestures” were made using combinations of fingers.

The most impressive aspect was their finger tracking, which I found to be very reliable once the controllers were fitted correctly and calibrated.

Individual finger tracking was almost flawless, even when lifting fingers off the body surface and placing back down in a slightly different pose; unlike my Valve Index controllers which sometimes struggle to maintain accurate per-finger tracking despite the “drum roll” recalibration and hardware reset tricks.

Etee’s pinky (little finger) tracking was especially reliable, and allowed me to finally include my little finger as a reliable participant in my experiments

Another impressive feat was their sensitivity to finger pressure, as mentioned in their literature “100 levels” per finger and its easy to understand this is no idle boast but a reflection of the precision of their material sensing technology.

This sensitivity allows controlled ramping up of pressure, and it was great fun “driving” my fingers and watching the finger levels rising from green to orange on the Visualiser.

I found the calibration process which uses outstretched hands created a strange finger input pressure if I tried to relax my hand into a neutral pose; try relaxing your arm and hand, and watch your fingers curl inwards as if holding a drink.

I tried to recalibrate from this neutral position but it caused the calibration to go a little wonky so I reset it to the outstretched fingers position again. This meant unless I kept my fingers outstretched (an unnatural pose) the input pressure rose typically on the pinky and ring fingers.

In terms of hand fit, I immediately found the foam finger bar too close to the controller body, causing my fingers to feel compressed to the point of discomfort, limiting my session time.

This sensation felt much more pronounced as Etee are securing the controller to the hand by clamping a foam block across the dorsal side (backside) of the Proximal Phalanxes.

The dorsal side of the hand is unused to pressure as its pretty much unused in everyday life, unlike the palmer side which is familiar with contact as we hold objects by using fingers to pull objects against our palms (try picking up your smartphone).

I refer back to my ergonomics article on the Valve Index controllers:-

The skin of the palmer side including the fingers is tough, thick and hairless with your fingerprints (double rows of papillae) help you grip objects as well as protect the skin from ripping.

Straight away with the Index controller I have the odd sensation of having something clamped across the back of my hand, it’s unusual as it goes against a lifetime of not having anything clamped across the back of my hand!”

The Etee controllers go one further than the Index controllers in clamping to a more sensitive part of the hand with thinner skin, and clamping using a high density foam of narrow width which is readily felt as it localises clamping pressure causing a proprioceptive challenge like Index.

Where Etee is less problematic is that the hand has soft rubber on side and foam on the other with Index having fabric on one side and hard plastic on the other which causes a material mismatch.

There wasn’t any way to adjust the position of the foam, as the bar was a fixed length to the controller body, so I turned my attention to the foam itself; thankfully this was only secured with adhesive tape so it quickly came loose.

Now with a blank finger bar, I searched through my fitting toolbox and found a number of different pads from bicycle helmets as well as foam pipe lagging of different thickness. Getting to work I trimmed different pads and stuck them in place, quickly realising the difference between ‘just too tight’ and ‘just too loose’ was critical to hold the controllers firmly in place without discomfort.

The alternative cushioning was interesting to try out, but made the controllers less stable as the material was a little too soft allowing free movement of the fingers, it needed to be a higher density like the original Etee foam cushion.

Moving further I removed the finger bars completely and tried using the controllers “naked” which was interesting, though a bit frustrating as without clamped fingers the sensitivity was all over the place making reliable inputs less frequent.

After spending a while trying the Etee controllers without their finger bars I decided to strip them down further and had a good look inside after peeling off the rubber controller body. The sensor strips were clearly visible, and some internal hardware.

Before putting the Etee controllers back together, I wanted to examine how changing the shape of the controller body would alter the hand fit so cut and taped together a larger diameter foam shape with a forward offset to create a relaxed hand pose.

Without the finger bar the controller felt like an oversized GearVR or Daydream remote and surprisingly comfortable to hold though the sensors did not register buried far underneath the foam. The big difference was that the controller did not “roll” in the palm when pulling the fingers as there was no gap with my hand “filled” by the new shape, creating a more stable grip pose.

Looking at other grips used for different sports equipment, handlebar grips on bicycles were an immediate comparison.

Round, constant diameter rubber grips are used on flat-bar sports and mountain bikes to allow the hand to move dynamically as the rider moves their body about on the bicycle, especially when standing up or sitting down. This instability of grip in the hand is what helps give a sports bicycle their agility, but at expense of stability and comfort.

In contrast, shaped ergonomic grips are commonly used for leisure bikes which are typically ridden in a seated position; another use for ergonomic grips is for riders with ulnar nerve damage whom benefit from a reduction in pressure on the wrists by having the palm fully supported.

After my quick test with the foam the solution to adjust the shape of the controller body and rubber skin would be the decision of TG0 as it would require a new design of controller (though using the same technology). Satisfied that adjusting the body shape could prove very beneficial. I refitted the rubber skins and thought more about the finger bar.

Adjustment time

I needed to create an adjustable distance finger bar; using the original high density foam piece and accommodating for different hands by moving the finger bar in or out, rather than trying to adjust a fixed distance using different thickness foam pieces…which hadn’t proven successful so far.

After sketching out the existing design and looking at the measurement I realised that creating a method of adjustment was going to be tricky due to the lack of physical space around the finger bar and body junction.

Following a number of iterations in my workshop using different cog parts (commercial off the shelf goods) like metal brackets from bicycle mudguards, plastic light fixings and different sized metric fitting bolts I finally managed to create a working prototype which actually surpassed my expectations in terms of its adjustment range.

By removing the finger bar’s attachment block (which has 2 angled keys that lock into slots on the body) I created an empty volume of space in which to fit an adjustment mechanism; 2 steel plates with oval slots to allow “infinite” sliding adjustment within its range.

A fixing bolt was captured with a knurled plastic knob with its own internal screw thread, and various bolts, nuts, grub screw used to attach the controller body to one of the steel plates, and to attach the finger bar to the other steel plate.

Finally I cleaned up the finger bar with isopropyl alcohol and firmly stuck some male velcro in place, and female velcro on the rear of the original finger bar foam so that the foam cushion could be adjusted from left to ride along the chanel of the bar – this would allow fine tuning of the foam ridge relative to the index finger on each hand.

After assembling my prototype I sent an email to TG0:

Previously I found the controller squashed my fingers which was quite uncomfortable and limited my time using them, unfortunately no way to adjust the clamping force though I tried a number of different cushions inside the finger bar (always slightly too loose or too tight). 

I quickly realised a method of adjustment was required. 

I went through many iterations to try and get this right as space was very limited whilst it needed to be easy to adjust whilst wearing, and offer “infinite” adjustment within the range of movement (clicks/notches are often just too tight or too loose).


This modification allows 3 adjustments to accommodate a wider range of hands/fingers. 

1. Distance (depth) from controller body to finger bar 

2. Angle of finger bar relative to controller

3. Lateral position of foam relative to index finger

Adjustments 1 and 2 are made using the black knurled knob which loosens and tightens a bolt, easy to do whilst wearing. 

Loosening the knob allows forward and backwards sliding movement to set the distance and adjust angle if required; tightening the knob locks the position. 

Adjustment 3 is made by repositioning the foam on the finger bar, because the foam is now velcro backed it’s easy to adjust left to right(can also swap to different shapes or materials)”.

The proof is always in using something practically, so I slipped both controllers on and started the Visualiser. For the purpose of this experiment I had modified only the left controller, as I didn’t want to risk damaging both if it didn’t work properly.

Thankfully, this was not the case, with the new adjustable left controller allowing me to find the ideal depth and best angle for the finger bar foam to maximise comfort whilst maintaining enough clamping pressure to retain the controllers in place.

This adjustment was easily done whilst wearing both controllers, with the knurled knob very easy to turn due to its vertical orientation.

The adjustment angle of the finger bar foam was limited by its square cross section, perhaps cutting a slight chamfer on each outside edge would soften its presence and allow more angle adjustment whilst providing enough flat contact against the load bearing bone to stabilise the controller.

Having the adjustable left controller and fixed right controller to compare against each other allowed easy assessment of the differences in comfort; the adjustable controller now accomodating my hand fit without squashing my fingers.

In terms of function, still very similar with the controller body rotating towards my palm every time I closed my hand.

Clamping the controllers to the proximal phalanxes will always cause this rotation but without a 6DoF version to assess using steamVR applications, I was unable to properly assess the etee controllers as virtual reality input devices.

The question of whether the mechanical switches, for example the trigger common to existing motion controllers, can be replaced by TG0 sensing technology; wasn’t something I managed to properly assess during my experiment.

The day following the completion of the adjustable left controller, the rigid plastic finger bar of the right controller snapped across its middle whilst I was removing the controller, putting an end to my experiments with Etee.

I suspect it fractured from being overloaded by my fingers being too large to accommodate, stressing the plastic over time and failing at the weakest part of the finger bar.

After speaking to TG0 they advised me the plastic was not production strength but a weaker 3D printed piece; this failure should not occur on production controllers and so my failure with the Dev kit should not be considered representative.

Immediate improvements to my adjustable concept would be a more rigid adjustment structure as the steel plates were rigid enough for my testing purposes but would benefit from being stiffer for day to day use especially for larger hand users with more strength and finger leverage.

Additionally, a “captive” bolt for the sliding adjuster plate to make it even more responsive to adjustment (the existing bolt can start turning if the knurled knob is unscrewed too far).

Finishing my time with the Etee controllers, I thought about the improvements that could be made to their ergonomic shape, and adjustment to suit a wider percentile of users.

The sensing technology itself was very impressive, and with a successful Kickstarter now backed and a SteamVR version due towards the end of the year there is a lot of potential waiting to be unlocked from this interesting product.

Thanks to TG0 for supplying this sample, I wish them luck in their developments. And thanks to you for reading! Rob Cole, immersivecomputing.org

Categories
Experiments

Re-index

Originally published by SkarredGhost. Edited to remove my original spelling mistakes and some images updated. All images by Immersive Computing unless otherwise stated.


Introduction by Tony at Skarredghost:

June, 28th, one year ago, Valve launched the Valve Index, one of the best virtual reality headsets on the market. To celebrate its first birthday, my friend and renewed XR ergonomics expert Rob Cole of Immersive Computing has decided to write another one of his very interesting guest posts about his experiments with this device, that you can read here below. The post is long and detailed, and will teach you all the modifications that you can apply to improve your experience with the Valve Index.

Happy birthday, Valve Index!


Re-Index

It’s hard to believe it’s already been a year since I received my Valve Index PCVR kit on launch day, June 28th 2019. I wish the Valve Index a very happy 1st birthday, with reports of new Indices being shipped to customers as fast as they can make them!

This article looks at some of the different modifications I’ve experimented with during the past year to improve the fit of my Index headset, Index controllers and Index ear speakers, as part of what I like to call my Re-Index project. 

I’m not going to get bogged down by talking about RMA claim as it’s not the focus of this article. But yes, there have been many problems since launch including replacement of the headset, tether, controllers, and ear speakers. I lost about 1/3rd of the past year waiting on replacement parts.

I’m currently on my 7th left Index controller and 6th right Index controller which informs me that controller issues have not yet been resolved. But I’ve been impressed by Valve’s active warranty support with a number of my RMA claims resolved as “advanced replacements” where the new product was shipped before the faulty product was returned.

Despite the loss of VR time and frustration over sometimes the long delay in receiving new parts, overall the sheer sense of presence and immersion provided by Index is unlike anything else currently available on the consumer market, and for that reason alone I have stuck with my Index.

Wide Face Gasket Base

For some unknown reason Valve released the Index headset in June 2019 with only a single, relatively narrow facial gasket, which immediately became a problem for my medium-sized head (59cm) as the Index simply did not fit onto my face.

I was puzzled by this omission, as their previous collaboration with HTC saw the Vive released in 2016 with both narrow and wide facial gaskets to offer 2 fitting options out of the box. 

Later I learned that Index development kits were originally supplied with narrow and wide facial interfaces, as listed by the Valve Index webpage from early April 2019 which clearly listed:

2 Face Gaskets (narrow and wide)

But by the end of April this was changed to just:

Face Gasket

Much later (early 2020) physical evidence of the “L” development face gasket emerged after Steam Support mistakenly sent one to a Redditor, who then posted a picture on Reddit asking how to attach the gasket, as it did not have magnets (production gasket) but clip-in tabs (pre-production).

This has led me to consider why the wide facial gasket was officially dropped, and despite asking Steam Support on three different occasions over the past year, I always get the same reply:

Unfortunately we do not have any information on potential future products. For information on current sales and promotions please refer to the Steam Storefront, the Steam News section, or on the Valve Index store page.

For reasons unknown it seems there is no interest in a wide face gasket within Valve, which seems bizarre as it cannot be true that all of Valve’s staff and 3rd party developers happened to have narrow faces.

One theory is that the Index is display-limited, and the wider face gaskets simply showed an unflattering view with black borders. This idea is supported by an IPD upper limit for Index of only 70mm, compared to the Vive Pro at 74mm. From Steam user “Edhem”:

My IPD is 74mm, and as soon as I tried the Valve Index, I started getting a headache. I currently have the HTC Vive and I have tried the Vive Pro, both with 75mm IPD vs the Index’ 70mm, and neither of the HTC headsets gave me any issues. I am sad that I will never be able to buy the Index as it appears to be a great device, but only made for smaller people.

Either way, there was no wide face gasket product offered by Valve, although Valve did the modding community a big solid by releasing the Index modding CAD files which included the plastic base model for the face gasket.

Modder “Anonymous Hermit” quickly stretched out the Valve base design to create a wider base, and released it on Thingiverse under the Creative Commons Licensing scheme. You can find it at this link: https://www.thingiverse.com/thing:3738708

After getting quotes around £200 from several UK 3D print shops, I managed to get a much more reasonable price from Ninja Prototype who offered to print my base in their Chinese print shop and ship direct to the UK.

This was the first time I had used a 3D printing service and I was really impressed by their smartphone application which made the entire process incredibly easy.  About 2 weeks later, my wide base arrived and I got to work building a wide face gasket for my Index.

Wide face gasket build

Top: Valve Index face gasket; Bottom: Wide face gasket build

With Anonymous Hermit’s wide base design now printed, I set about building my own wide face gasket. The dramatic difference in the width and radius between the Valve face gasket and my wide face gasket can be clearly seen in the image above and demonstrates how crucial the wide gasket actually is for fitting wider faced Index owners.

I purchased a pack of 1 x 5mm Neodymium magnets, industrial-grade self-adhesive Velcro, and aftermarket Vive face cushions, and carefully bonded the magnets and Velcro to the 3D printed base after degreasing with Isopropyl Alcohol to remove any skin oils or other contaminants.

The Vive cushions were then cut to fit the larger Index eye box and positioned to give the most comfortable position whilst presenting the optics in a neutral position.

After letting the adhesive cure overnight I started testing my wide face gasket which immediately solved the “narrow fitting” issue and proved that a wide face gasket could work for Index.

Other “comfort” issues would arise in the longer term from the use of PU face cushions, but at least for now I could finally wear my Index without having my face squashed and my eyes in the wrong position.

Magnet Stacking

After building my wide face gasket, I discovered it was possible to make fine additional adjustments to the face gasket fitting by “stacking magnets”, as seen in the image above.

To get a 3D printed facial gasket to attach to the Index headset, 4 magnets are bonded onto the plastic base, 1 in each corner; I used readily available 2-pack epoxy adhesive and left to cure overnight. 

These tiny magnets (5mm dia x 1mm thick N42 F351-50) are very strong (0.3kg pull), and quite capable of supporting several stacked magnets without the stack becoming unstable (which could cause the face gasket to shift around).  

This led to a realisation that some asymmetry could be partially accommodated beyond the standard adjustments offered by the Index headset, using different arrangements and numbers of the magnets on each of the 4 mounting posts of the face gasket.  

One noticeable difference for me was placing stacked magnets on the 2 top mounting posts which seemed to sharpen the picture quality as well as creating eyebrow clearance, I also experimented with different left/right bias which made further subtle changes. 

I wrote an article about building my wide face gasket for this blog and you can find it here.

VR Cover Valve Index face gasket

After months of using cut-up Vive face cushions in my 3D printed wide facial gasket base, I was pleased to learn that VR Cover were bringing their own facial gasket to market – the first after market facial gasket replacement for the Index to hit the market.

Tony at SkarredGhost was kind enough to put me in touch with VR Cover who quickly sent me some production samples to test. 

Their well finished plastic base was the same width and radius as the Index original, with thick and thin PU face cushions offering 2 different fits on the 1 base.

VR Cover’s thick cushion gave the same narrow fit as the Index original so was unusable for me, whilst their thin cushion was a closer match to my facial width, but at the expense of headset comfort. Index is a relatively heavy headset with a long tether requiring clamping force from its head strap to keep the headset in position on the face especially during active gaming. 

I quickly found the thin cushion was unable to isolate my face from the weight and clamping force of the Index headset, whilst the PU material got hot and sweaty very quickly allowing moisture to build up inside the headset, making longer sessions uncomfortable with red marks left on my forehead, sweaty skin and sometimes a mild headache.

My recommendation to VR Cover was to produce the plastic base in 2 widths (original and wide), offer a pre-curved (no ripples) face cushion for each width of base, but using a medium thickness breathable foam and skin-compatible surface layer to provide better comfort for longer sessions, kinda like what Valve did with the original Index facial interface.

Another concern was the lack of absorbency to protect the headset from sweat corrosion damage: as seen in the image above the PU cannot absorb moisture causing the lenses and inside of the headset to quickly become wet. This is not ideal in the long term as it may cause similar damage to that seen on the HTC Vive when used with PU cushions.

Unfortunately the Coronavirus pandemic soon followed, and production facilities around the World shut down which saw stock shortages of all VR products including headsets, controllers and after market accessories.

Hopefully VR Cover and many other VR companies with their manufacturing partners can all get back to prototyping, building and distributing their products very soon as many VR enthusiasts especially Index owners await further improvements!

You can find my full review of the VR Cover cover for Valve Index in this blog as well.

“Pogo biscuit” working prototype

After getting a workable face gasket installed in my Index headset, I turned my attention to my ears which were unhappy with higher sound pressure blasting my left ear.

After looking at every audio related software solution, including reducing left volume which seemed to just let in more background noise rather than resolve the imbalance, I realized it was a physical imbalance caused by my perfectly normal, human asymmetry. It’s common to have asymmetrical ears, faces, hands, feet, etc.

With a normal pair of audio headphones, most people can get a good fit with the adjustments available on the sprung headband, and unless an outlier in terms of fitting, will find the experience comfortable.

However, with the off-ear “Ear Speaker” design on the Index, there was no depth adjustment to accommodate different size and shape ears, and wonky heads! This meant that the left and right speakers might be a different distance between the left and right ears due to variations in head and ear shape.  

A physical adjustment was required, and I measured my gap difference at about 10mm on the left, which was a good starting point. Quickly going through ideas for redesigning the arm mounting mechanism, before settling on a simple physical spacer with matching “pogo-pin” pass-through electrical connections for the headset and ear speakers.

Soon, a collaboration on Reddit with Anonymous Hermit produced the Thingiverse model they published; using an East London 3D print shop I had a dozen biscuits printed and built some working prototypes after using cut down bicycle spokes to create electrical pass-through pins. I took the best-finished prototype, installed it into my head strap with a longer M2 screw; it worked well and proved the concept was firm.

However, it was tricky to build at scale as the 3D printed solution wasn’t ideal in terms of small detail strength, requiring an injection mould tool and pogo pin fittings would need to be custom made with large minimum orders.  

Whilst I was figuring this all out, the Hermit launched a much smarter and simpler solution…

5.4 degree ear speaker spacer

Anonymous Hermit had figured out with some maths that a simple shim could be added to the existing BMR ear speaker setup whilst maintaining electrical and structural connection through the existing pogopin connector. You can find its model on Thingiverse as well.

In an act of kindness (many thanks!), the Hermit mailed a number of samples they’d 3D printed from their base in New York,USA; as seen in the image above.

I have 1 of these installed on the left side of my Index, and it sat there quietly for many months since just doing its job, which is the sign of a great design. These spacers are essential to have in your modding box if you own an Index and want to make ear speaker adjustments.

Depending on its fitting orientation, the spacer can be used to move the ear speaker in or out relative to your ear. Due to the pogo pins requiring a good connection but having limited engagement depth, its not advisable to double up the spacers.

The improvement to the audio was immediate, giving a balanced and comfortable audio presentation to both ears. Feedback was provided to Valve so hopefully, we will see a future iteration of the ear speaker with a depth adjustment mechanism. 

Audio Lead Extension

During the first few weeks of ownership, my Index was really underpowered in the audio department, which was somewhat disappointing. 

This caused a big loss of immersion, as audio is incredibly important in VR, and living in a busy city I immediately noticed background noise bleeding in where my previous PCVR headset with closed headphones had proven very immersive in the same room.

My immediate solution was to remove the ear speakers and switch to audio headphones to reduce this unwelcome background noise. Removing the ear speakers and using headphones proved less easy as I’d already switched to using a 3D printed wide face gasket.

This gave less clearance for headphone cables, as the audio port behind the face gasket had an awkward orientation and needed a deep jack plug with a sharp right angle, causing many cable plugs and leads I tried to foul the face gasket, uncoupling the magnets and lifting the gasket off the headset mounts.

Eventually, I modded a Logitech headphone splitter cable by removing one lead and shaving down the rubber plug grommet so the cable could clear the tight gap without fouling the face gasket. 

Using the Index with headphones was beneficial in terms of reducing background noise, which is important for City dwellers or at location-based entertainment venues where you don’t want outside music overpowering virtual experiences. 

For quieter applications with little body movement like “The Blu” my Sennheiser and Logitech headphones worked well by removing the noise of my PC and cars driving past my street in London.

Though it became obvious the amp inside the Index was specifically designed and tuned for the BMR ear speakers, as the tonal quality was quite different to when these headphones were plugged directly into my PC or even my smartphone. The headphones also seemed a little underpowered which led to the next finding.

All audio settings were tried during these weeks of quiet audio, until eventually it was tracked down to the Nvidia audio driver feeding the HDMI and Display port cables. The driver was applying minus-6 in the pre-amplification stage dramatically reducing the amount of power available to the ear speakers.

Later, Equalizer APO was installed on my PC allowing me to resolve the lower sounds power issue by adjusting the pre-amplification stage to finally provide full power to the ear speakers, which then truly “came alive” with a huge difference in audio presence.

Once the sound power issue was resolved, I spent most of my time using the BMR ear speakers as their sound quality was very good, though to be critical slightly lacking bass weight when compared to studio headphones. For their tiny size and weight (53 gramme) though, they pack a big punch and have excellent clarity throughout their range including going very loud without distortion. 

The other important reason I stuck with the BMR ear speakers was for the sheer quality of life and comfort improvements of using headset mounted, off-ear audio, combined with their really immersive and open soundscape interacting naturally with the unique shape of my wonky ears.

The only negative has been the reliability of the connection between the ear speaker “pod” and the mounting arm, the plastic joint tends to wear quite quickly, causing a small amount of free slop to develop which can allow the speaker to vibrate at higher volumes. In an ideal world, Valve could redesign the mounting arm to provide depth adjustment and eliminate this free movement.

I wrote on this blog a long deep dive about my experiments with the audio of the Valve Index, and if you are interested, you can find it here.

Valve Index palm boosters

The big disappointment for me of the Index launch was actually the Index controllers, which left me a bit shocked and confused.  They looked very cool and it was definitely a special feeling to handle them after following their development for several years following my purchase of an HTC Vive in 2016.

However, they didn’t actually fit my hands very well, with the controls in the wrong place, the controller body feeling too small in size, and a strange material mismatch between having a fabric strap clamped across the back of my hand whilst the hard plastic body controller pressed into my palm. I started wondering how they could have got it so wrong

Soon enough Valve came to the rescue with their “Palm Boosters” clip-on design released for 3D printing. Straight away I placed an order through Etsy.com and had them printed and shipped to the UK by Printlix in Romania for a very reasonable price.

The Palm Boosters fundamentally changed my fit, control, comfort and material mismatch.

I should mention I have medium-sized hands and wear medium-sized bicycle gloves; it seems the Index controller design offers the “naked” controller as small size and the “palm booster” fitted controller as a medium size.

This leads me to believe that larger handed owners could further benefit from a “large palm booster” with a volume increase compared to the current 3D printed part.  

I liked them so much I bought a spare pair and then wrote an article about it for SkarredGhost, that you can find here, if you are interested.

Studioform Creative VR Apache DAS head strap

In the ongoing search to further improve headset comfort for long VR sessions, I found this Apache DAS head strap from New Zealand company Studioform Creative which was advertised as being compatible with the Index.

This had been originally made for the Vive DAS and so didn’t fit my Index particularly well with the front strap fouling the BMR speaker mount, making fitting a bit awkward.

One benefit was I could reduce the overall strap pressure for the top strap and rear harness, although this could cause a loss of stability if too loose, so it was about getting the tension just right. 

Admittedly I continued to struggle getting the headset sat right each time and didn’t enjoy the extra layer of “friction” from having more straps to fuss with, sometimes I found the Apache DAS straps causing interference and moved the headset from its optimum position. 

It also caused my head to get a little warmer during active experiences like Beat Saber and Pistol Whip, so eventually I went back to using the Index with its standard top strap setup.

VR Cover Valve Index headset Cover

Soon after VR Cover had sent me their Index Face Gasket kit to test, they sent me their new Valve Index headset cover, which arrived with 2 covers in the pack.

This was designed to prevent wear and tear to the rear cushion of the Index headset hardness, as the cushion is not user-replaceable.

I found during my testing it added a bigger benefit of more firmly gripping the rear of my head, which stabilized the headset during active movement. I didn’t notice any additional heat compared to the standard head strap. This accessory was genuinely useful and a great day one installation to prevent your Index headset from getting a dirty or worn rear head strap. 

The full review on this interesting gadget is in this other post.

Soft top strap modification

During my ongoing search to improve comfort for long term sessions in my Valve Index I tried cutting up my Apache DAS strap to make the top strap of the headset more cushioned.

I had never found the standard top strap very comfortable as its thin and unpadded which can be felt on the head especially for bald headed people like myself. I’d noticed some Index users recommending repurposed strap covers from satchels and hand bags, but my DAS strap was sitting there doing nothing…

Whilst this modification added a good degree of comfort, it also increased the heat by covering a substantial part of my head, so I soon removed it when I got back into playing Beat Saber.

Soft tether routing

The standard tether routing on the Index caused my Index launch headset to start sparkling after 2 short weeks of use, after simply opening and closing the headset a number of times had put a nasty kink in the tether which was firmly anchored by 2 plastic mounting clips.

After receiving a replacement headset I switched to modified cable routing which ditched the plastic mounting clips in favor of “soft” Velcro attachments. Initially, I used a 3D printed Index tether belt clip mounted using a Velcro strap above the head strap rotary dial but noticed a crease developing where the tether exited the cable clip.

This led to understanding that the tether is relatively soft and quite easily damaged from being coiled up, badly stored or being stressed whilst being worn due to poor cable routing; additionally any blunt edge (cable clamp) applied against the tether over time starts to crease the plastic cover. 

Talking of storage, its worth looking at videos of “how to coil audio cables” as its very easy to damage the tether from improper storage with the cable becoming internally wound.

Eventually I settled on a “soft” solution as seen in the image above, which firmly holds the tether with large pieces of soft Velcro that stick to matching Velcro strips on the harness. Since this image was taken I have extended soft support for the tether where it comes off the head strap to prevent any crease developing in the tether. 

Used with a “soft” belt clip (Velcro wrap) and shorts/trousers with a suitable belt, this stops the tether from getting easily damaged and should reduce its unwelcome presence felt on your head or back as you move around.

Frunk

Valve designed the Index headset with a “Frunk” (front trunk) which contained a USB port to offer expansion options to the modding community. At the launch event for journalists, Valve had a small display panel installed displaying different messages.

The expectation was that the community could come up with all kinds of cool stuff to bolt into the Frunk, and the community has certainly delivered: at this page you can find the list of all mods created by the community.

Personally I found the Frunk ideal for housing my Xbox controller wireless dongle, or rarely for charging an Index controller with a flat battery whilst continuing to play by plugging it straight into the headset.

It’s neat but non-essential as the Xbox dongle works fine when plugged into my PC, and charging is much quicker using wall-mounted USB mains chargers.

Then a question arises, would removing the Frunk improve the headset design? This could remove a big lump of plastic from the front and provide a lighter/slimmer headset.

Bear in mind the Index is a heavy headset with a bias towards the front, so any reduction in weight, complexity and manufacturing cost would be welcome. I’ve already seen modifications where owners have removed the Frunk cavity and replaced it a 3D printed blank, though warnings exist that this can upset the calibration of the tracking system so is best approached with caution.

Frunk removal would have to be weighed against the benefit that some Index owners really appreciate, the most obvious being the “Chilldex” cooling system that fits neatly into the Frunk, or an Ultra Leap module for hand tracking experiments. For now, I’ll keep using it for my Xbox dongle as it frees up another USB slot on my PC which is always useful. 

Cranial cap with occipital support

After all my experiments with face gaskets, top straps, and rear covers, I looked at a more radical proposal. Instead of adding extra padding and material to solve comfort issues (which just added weight and heat), perhaps a fundamental redesign of the headset mounting would be more beneficial.

I drew inspiration from the Kathryn Bigelow film “Strange Days” where the lead character Lenny wears a very cool head-mounted FBI developed device. The “Squib” records brain waves as the user lives an experience, but also allows playback causing another user to relive the experience, starting an illegal black market of recordings.

This device had a skeletal structure designed to closely fit the skull, and this was my launching point for a “cranial cap”. 

Image: Strange Days, 20th Century Fox.

The benefit of the cranial cap is to use the skull itself to support the headset, and remove pressure from the face, perhaps using a soft gasket like PSVR to seal the eye box against light ingress . I stripped down a bicycle helmet and removed its “Mips” safety liner to use as a working model to develop the concept. After rebuilding the liner with attachment points for the headset, I pulled back the Index harness and tried some fitting sessions to check clearance.

The idea seems to work well, with the next step being a 3D head scan so I could build a head model, but as the Coronavirus pandemic unfolded it became a non-option with my booking at a specialist scanning facility in London “put on hold”.

The other option was the old-school route of making a head cast using plaster of Paris, which would provide me with the ability to then plaster cast a physical model of my head onto which I can directly build a cranial cap unique to my head.

The cranial cap concept has a minimalist design of wireframe structure with CoolMax contact pads in key places, and an occipital support structure at the rear. Weight, pressure, adjustment, cooling, skin comfort are all considerations. More on this soon…

Closed headphones for BMR ear speaker mounts

Another project I’ve been looking at a while is a “closed headphone” for the BMR ear speaker mounts on the Index headsets.

After my experiments with audio leads, headphones, and ear speaker spacers I wanted the option of a closed headphone for users like myself with noisy backgrounds or using quieter applications/experiences where a closed headphone could be ideal.

This would simply replace the existing “open” BMR ear speaker using the same pogo pin fitting to provide a hassle-free swap-out.

The trickiest part is not physical packaging and adjustment, but working out the best headphone drivers to use which can play nicely with the amp inside the Index headset, as mentioned previously the amp has been specifically tuned for the BMR ear speakers.

Some final thoughts on 1st year of ownership

Framerate

The Index was unusual in being the first headset to offer user-adjustable frame rates of 80 Hz, 90 Hz, 120 Hz, and 144 Hz. The higher frame rates and ultra-low persistence display had big implications for the user’s sense of immersion and virtual presence

Being a “power user” with an overclocked multi-core processor (8086K @ 5.2Ghz) and GPU (MSI Duke OC 2080Ti) I had lots of headroom to experiment, or so I thought…

It quickly became apparent that many VR applications were poorly optimized meaning even a power user may struggle to hit the right frame timing without wallowing in re-projection, dropping frames or having to use motion smoothing.

A handful of less graphically demanding games like Beat Saber, Space Pirate Trainer, Pistol Whip and Superhot ran well at 144 Hz, whilst most of my games worked well at 120 Hz, and some of the more demanding games worked best at 90 Hz.

I even tried 80 Hz for some of the driving simulator applications which are computationally expensive and tend to quickly grind to a halt once graphical settings are increased.

120Hz seems to be the “sweet spot” for most of my applications, giving a fresher, sharper feel than 90Hz but without going into unwanted re-projection. But for certain titles, the benefit of applying “super resolution” outweighed the desire for higher frame rates.

Super Resolution

Often called “super sampling”, Index really benefited from my powerful PC as I could increase the render resolution which is then downsampled to fit the native resolution of the Index display hardware, reducing aliasing and giving amazing clarity.

130% resolution at 90 Hz, 110% resolution at 120 Hz or 100% at 144 Hz were just some of many settings I tried, Index is fantastic in giving the user these options if they have the computing power to play with.

For some seated games like “Aircar” 90 Hz felt very comfortable, whilst increasing super-resolution in-game to 1.3 gave a big boost to immersion (visual presence) with an incredibly clear visual quality giving a good idea of how much higher resolution displays will benefit future PCVR headsets.

I was surprised how much difference super resolution made for Index, with its minimal screen door effect (partially thanks to the display diffuser) it pushed the display quality to another level, with some really breathtaking moments in the best VR titles.

However, for fast-moving action games like Pistol Whip and Beat Saber, using higher frame rates with regular resolution (for stability) provided a more substantial advantage for game accuracy and body presence.

Brightness

Something not talked about much is display brightness (luminance), but its very important in terms of contrast, visual immersion and feelings of presence.

Its certainly something I noticed when moving from a HTC Vive to an Oculus Rift CV1, which had a clearer display with less screen door effect but seemed to be quite dim with muted colours. Whilst I enjoyed the clearer picture, I never felt “present” in the CV1 partially the dimmer screen and partially the smaller field of view.  

Something noticed by many veteran VR owners using the newer LCD headsets is a big reduction in display brightness with less vibrant colours, and murky or washed out blacks in some dark scenes. The LCD displays operate very differently to the older OLED display, and simply do not provide the same levels of luminance, which can make applications colour graded and lit for the older OLED headsets, less than ideal on Index.

To give some figures, quoted by VR Dev and Redditor “Eagleshadow” who did direct testing on these headsets:

Luminance in nits:

  • Index: 95
  • Vive Pro: 143
  • Vive: 214

As you can see, Index has a relatively dim display compared to the older OLED displays in Vive and Vive Pro. Thankfully, Valve later released a headset update which allows the user to increase display brightness to 160%.

I haven’t seen any measurements yet with the display at 160%, but anecdotal comments put the increase closer to the Vive Pro but not as bright as the original Vive.

As there seems to be a correlation between display brightness and the glare inherent to the dual compound Index lenses, brightness adjustments can be made depending on applications and user requirements/preferences.

Conclusion

Well what can I say about the Valve Index after my first year of ownership? Its been a mixed year no doubt with incredible moments of sheer presence in virtual reality, tempered by the tedium and frustration of often waiting on replacement parts, typically for the Index controllers.

Ultimately I am here using virtual reality purely for the immersion, and the Index offers an immersive experience I haven’t found in other VR headsets so far!

Its been very revealing going to a number of location-based entertainment venues and events over the past year and feeling less than satisfied after using a Rift S, Vive Pro Eye or Samsung Odyssey Plus. I found with the older headsets they simply do not offer the same level of immersion as my Index. 

Valve has certainly created something very special, yes it has some flaws and quirks, but as their first attempt at an “in house” system it’s a great first step and can only get better as further developments improve the overall experience and hopefully reduce the need to regularly RMA parts of the system.

Current pain points

1. Face Gasket

Left: Valve original, right: wide custom

Wide face gasket – we need an “official” version as soon as possible, or an aftermarket version with sweat absorbency and skin compatibility

2. Wireless PCVR

The tether…nothing has really changed since Vive Pre in early 2016, it’s still a nuisance, point of failure, breaks immersion; wireless would be very very welcome. We are probably waiting on the new wireless standard to be approved for domestic use, as the standard was delayed last year. Having used wireless on the Vive Pro, I would happily pay top dollar for an Index wireless module  

3. Controller Joysticks

Whoops! Sure that’s not supposed to happen?

Controllers – upgraded joysticks are required, as a small increase (1mm longer) I measured in the shaft length of the left stick on my later replacements has not prevented joystick drift from occurring again. 


Going back to a graphic I created in early 2017 after finishing my “Room scale Plus” experiments with the HTC Vive, summing up my thoughts as an early adopter.

It’s very interesting for me over 3 years later to see how many of my wishes have been met and which are still outstanding. It seems even back then, the tether had made itself known too many times!

Thanks for reading! Rob Cole

Categories
Experiments

Daydream Cool @ immersivecomputing

Coming soon after the mind blowing Roomscale Plus experiments I assumed mobile VR would be, to be polite, a pale imitation. However, being tetherless (albeit in 3DoF) was intriguing, and any excuse to try new technology is a good one.

Google’s Daydream launched with their new Pixel smartphones which ran a stripped down version of Android with VR support baked in and low persistence display panel with a SOC capable of rendering 2 x 60Hz. I headed to Google’s Daydream public demo hosted by a retail partner in London’s West End.

The increase in resolution and reduced screen door effect (compared to the Vive) were an immediate and pleasant surprise. The vivid colours and bright screen worked really well with a low poly style application I tried.

The Daydream demo was impressive enough to give me a sense of presence with its hand controller and accurate head tracking. I placed an order for a Pixel XL and Daydream headset for next day delivery.

Very neat packaging and design. The controller was incredibly performant as my tracking test below shows and gave a great feeling of hand presence despite being a 3DoF device.

At this time Daydream’s established competitor GearVR lacked the hand controller, immediately giving Google an important advantage and something leveraged by a number of developers.

Virtual Virtual Reality, Eclipse Edge of Light, Audio Factory, Dreadhalls, Rez infinite and later on Bladerunner Revelations…some standout applications and tight integration with YouTube VR, Google photos and early releases of Firefox Reality and Chrome VR gave access to WebVR experiments.

Whilst the display, remote controller and applications were effective, the headset was uncomfortable and unstable; if looking down the headset would hang forward from the face. Tightening the strap just put pressure on the facial interface without applying much stability. So I added a top strap, seen below.

The strap was later reworked to make it easier to adjust whilst wearing the headset. Facial interface pressure was still an issue due to the small size of the face gasket with a lower contact area than many headsets. Getting to work i adapted the gasket to accept an aftermarket Vive face cushion.

Facial comfort and stability were massively improved where it was possible to use Daydream for longer sessions and freely look about without the headset moving.

However, longer sessions were not really possible unless just using YouTube VR as the phone tended to quickly overheat with thermal throttling leading to shutdown. The inevitable consequence of sandwiching a smartphone in the plastic front of a fabric covered headset.

Some Daydream users had reported limited success (an improvement) using aluminium foil to siphon heat away. Others tried cooling fans with battery packs adding weight and noise. Looking for a simpler solution I removed the front panel completely….

Purchasing a Pixel XL smartphone case off Amazon, I built a passive cooling system with a finned aluminium heatsink thermally coupled to a copper coldplate which would contact the Pixel’s aluminium body. Of course adding a Daydream NFC strip borrowed from the original front panel so it would trigger Daydream mode.

Attaching this to a fabric covered VR headset was less easy, but solved with strong adhesive, industrial strength velcro and faux-leather material for the lower “wrap” whilst the elasticated top strap from Daydream was repurposed.

Testing quickly revealed a big difference in thermal performance with the Pixel now barely getting warm preventing thermal throttling. Due to reduced heat stress battery life was also markedly increased allowing me to have 2 hour sessions in games like Eclipse Edge of Light.

“Daydream Cool” was born and gave many hours of service until Google released Daydream 2017… which had a new cooling solution, top strap and improved facial interface.

2017’s biggest advantage over the original was the new Fresnel lens design which increased the field of view, clarity and sense of presence.

The only downside with moving from aspherical lenses to Fresnel lenses was the introduction of God rays (lens artifacts) which spoiled many YouTube VR applications so I kept Daydream Cool for media consumption.

Daydream Cool heatsink
Daydream Cool facial interface support
Daydream Cool top strap integration
Daydream Cool front panel

Thanks for reading! Rob Cole @ immersivecomputing

Categories
Experiments

Roomscale Plus VR @ immersivecomputing

HTC Vive PCVR in 2016 was an incredible experience but the limitations of a domestic playspace (3.0m x 2.2m) soon became evident with the intrusion of the chaperone if walking more than a few steps in any direction.

The chaperone intrusions broke immersion, whilst turning the chaperone off (developer option) just led to painful impacts with walls.

A number of Roomscale VR applications also had minimum space requirements larger than my playspace, which had prevented me from being able to use them.

The desire for true roomscale virtual reality led to 2 questions

1. How large a space could steamVR track?

2. How would this change the VR experience for the participants?

An opportunity came about in early 2017 to use several large spaces over 3 weeks at a huge retail store in London’s Canary Wharf district. After moving the PCVR system to the store i arranged some early morning sessions…

Space #1: The Workshop

A very easy space to setup with base stations mounts screwed straight into the walls creating 4.5m X 3m of playspace ideal for running left to right in Space Pirate Trainer’s roomscaling design.

All of the first week morning sessions were done in the workshop with 2-3 guests per session. The rubber floor was useful when someone ran head first into the wall (the Vive was fine) and then fell backwards.

Already planning for a move to the next, larger space, how to accommodate the base stations without a handy wall to screw the mounts into?

The Lighthouse Towers

Doing some rough maths I worked out I need to raise the base stations up much higher to maintain the optimum tracking angle across a larger floorspace. Also needed:-

Small footprint to get each base station as close to the walls as possible (tripods of adequate height had too large a footprint).

Stability to prevent tracking jitter as base stations had electric motors inside which tended to vibrate if not secure.

Finding some scrap aluminium alloy beam sections was providence, softwood blocks isolated the base station mount from the beam, holes were drilled along the beam length to secure the power cables with zip ties.

Workstation bases were borrowed from the store’s workshop which had heavy (20kg) steel baseplates with small footprints; the aluminium beams were fixed to the stands with softwood v-blocks, bundles of zip ties and gaffer tape.

The freestanding Lighthouse Towers gave a stable foundation for tracking accuracy whilst allowing the base stations right against the boundary walls to maximize tracking space. Despite being heavy they were easy to quickly move and each tower had its own 25 metre cable drum for mains electricity supply.

What about the PC?

The PC was mounted on a wheeled mechanics trolley with its own 25 metre power cable drum and fitted with a wi-fi receiver on a non conductive mast making it possible to move the PC anywhere inside each space. Intel Core i7 CPU and Nvidia GTX 1070 both overclocked to squeeze out more supersampling.

After a week of exciting morning sessions in the workshop space it was decided to move to a larger space behind the workshop which was used for bulk storage. .

Space #2 : The warehouse

We moved many boxes from the warehouse to increase the size we could expand into, but left a layer against the back for impact protection.

This gave a playspace 6 metres wide and 5 metres deep which started getting very interesting in VR with very little sign of the chaperone, the illusion maintained!

During these sessions the headset tether cable started making its unwelcome presence known leading us taking turns as a “tether assistant”, doubling up as media assistant shooting video and photos of the sessions.

Management of the tether freed the user; we quickly became adept at manoeuvring the tether as the user moved around to stop them tripping, tangling or tugging the headset.

“Abbot’s Book” was a favourite in this larger space alongside Cloudhead’s The Gallery Episode 1, Space Pirate Trainer and Valve’s The Lab.

Walking around inside The Lab was truly mind-blowing as was the sheer space of Viveport home with lots of out of bounds trespassing giving interesting views.


The third week was set to be the most exciting with a move into the largest room, a huge space used for displaying retail products on plinths.

Space #3 : The retail space

Each morning started at 5am by moving everything out of the way before quickly setting up the PCVR before guests arrived

With the space empty it was cavernous, wall mounts were used to string the sync cable between the base stations whilst the PC sat on its trolley half way along the right wall.

Having full control over lighting (pitch black) and temperature with powerful air conditioning gave a thrilling effect in games like Dreadhalls where the cold of the dungeon could be easily recreated adding another layer of immersion!

This is where things got really interesting as we reached the limits of steamVR tracking. The maximum using the link cable to sync the base stations was a vast 9 metre diagonal gap between the two base stations.

We also had problems with the tether length in more free roaming applications, having to move the wheeled PC around following the user traversing the huge space. Cable extensions worked… sometimes..but caused tracking instability.

After 3 weeks of Roomscale Plus sessions?

A number of the guests had never used VR, whilst some had limited experience. Imagine the first experience of VR being Roomscale Plus?

Watching the transformation was rewarding with people shouting, shrieking, laughing uncontrollably even crying. Brilliant to witness and also session as a user, freely roaming in VR was very liberating.

Going back to the original questions?

1. How large a space could steamVR track? A huge space as we found out setting up the retail space!

9 metres diagonal gap between the base stations = 6.36m X 6.36m = 40 square metres of tracked floorspace. It tracked the tallest guest waving their hands above their head, allowing 3 metre vertical =120 cubic metres.

Within this steamVR tracked volume we had full environmental control and empty space so no obstruction, no reflections to upset tracking, clean lines of sight and lights off for maximum immersion.

2. The VR experience was fundamentally changed by being able to move in a huge space without the chaperone intruding, or the tether interfering with the headset – thanks to the tether assistant!

Roomscale Plus gave a view of a possible future with wireless PCVR or standalone headsets allowing users to freely move within their application at a large scale.

All of the guests came away raving about their experiences and several went on to buy PCVR systems. It was great fun hosting the sessions with a carefully curated selection of content.

Several weeks of amazing experiences and memories came to a close all too soon, and it was time to return to my much smaller domestic playspace.

This however brought a crashing sense of disappointment as the freedom and scale of true roomscale VR was reduced to cautious steps and teleporting or free locomotion on the spot.

One unseen consequence of”Roomscale Plus” Virtual Reality…leaving my final thoughts on PCVR in early 2017

Thanks for reading! Rob Cole, immersivecomputing