Fully operational cantilever HMD

CantileverThis is the HMD I showed at the VanVR April event.  It is a fully operational display that incorporates the following design improvements:

– Clear plastic construction: This seemingly trivial improvement actually has some strong benefits for wearing headsets in complex environments, say a home setting with pets, children and other moving obstacles.  By being able to retain your peripheral vision you are able to anchor yourself in the real world while enjoying a virtual world in your central field of view.  I received many positive comments from “long time” Oculus wearers on this approach and only one tester commented that they prefered the blackout experience out of the twenty or so who tried it out.

– Adjustable Interpupilary Distance: By mounting the lenses in 3d printed nylon (Taulman Bridge) clasps and constructing a lateral box frame with a slot that  a pull tab can fit through I was able to provide the ability to physically adjust the interpupilary distance between lenses by simply pulling or pushing the tabs on either side of the helmet.  Although the physical system worked reasonably well, the feedback for users (including myself) to know when the adjustment was correct was missing.  I was however able to allow one gentleman with a particularly narrow IPD the ability to see a convincing VR experience without optical distortion for the first time, as this is not currently a feature of the standard Oculus Rift DK1.  This single event made the effort in adding this adjustment completely worthwhile for me.

– No face touching:  The design incorporates a front back head clamp design that uses an elastic strap and static friction to hold the head mount firmly to the head while the whole weight of the HMD is supported by the top strap (not shown in picture).  This provided sufficient rigidity to completely remove the need for a nose bridge or cheek pads which are typically used to prevent axial sway as the head is turned rapidly.

The headset design however does have several drawbacks:

– The head clamp system was only comfortable for a subset of head shape/sizes and was incapable of working on a childs head.

– The headset design did not allow the user to wear a pair of over the ear headphones.

Overall a very succesful iteration and the feedback from the people and VanVR has provided lots of good fuel for future improvements.

Look who-who is here

Just wanted to introduce you to my version of Mythbuster’s famous “Buster”.  No explosions or dead-falls I am afraid but he is ready, willing and able to act as the head model for some somewhat ungainly and ugly head gear.

Ladies and gentlemen, please welcome Hedwig!

Coming up for air

MasksIt seems that a raft of dive mask style HMDs are due to come out later this year including Oculus, Sony, Microsoft, Gameface to name but a few.   I am truly puzzled by the total lack of diversity of design of HMDs.  There are so many designs to explore yet the industry already seems to be “fighting the last (cellphone) war” by focusing bigger screens, faster electronics etc. thinking that these incremental improvements will magically drive the VR over the “presence” threshold.  I agree that there is plenty of room for improvement here but the optics and ergonomics also need dramatic improvement to make presence a sustainable activity, not to mention the critical “Content design patterns for successful presense” which has yet to be written let alone adopted.

My plan is to systematically study, explore, prototype and refine alternative HMD designs.  Since there are so many dimensions to explore I have chosen to start by studying the physical properties of trying to hang several hundred grams of stuff in front of your nose comfortably and securely without having to resort to the dreaded “face corset”.


FaceSo, lets start with the face itself.  Check out Zygote’s fabulous interactive 3D model of human anatomy.  Pull down the slider on the left a little and you will be able to easily study the structures under the skin that dictate where Head mounted displays should and shouldn’t touch.  The importance of mapping HMD touch points was brought home to me in both my initial Rift helmet design as well as my second one shown here:

As it turned out the reason that the cheek pads failed was simply due to that part of the face having complex musculature and thus nerves and circulation that did not respond well to continuous pressure.


Rift Dev Kit 1 3D Printed Headset V1

The Goal

My goal was simply to see if I could make something that matched and hopefully exceed the physical comfort of the current “dive mask” Head Mounted Display (HMD) concept used by Oculus.  Ultimately I would like to build my own HMD from scratch but will iterate towards this piece by piece.  The Oculus Dev Kit 1 is a great starting point as I can explore physical HMD characteristics without getting bogged down in all the electronics, optics and software (yet).

The Result (Yay!!)

The build

After many (many) iterations I was finally able to complete my first fully functional Head Mounted Display using the electronics, screen and lenses from a Oculus Rift Dev Kit 1 headset.

Lens and screen mount

Oculus-Eye-Cups-V1I created five separate lens mount variations and ended up using the one shown below.  It is nice and stiff while still being quite light and held the lenses firmly in the correct position.  The base of each cup exactly matches the shape of the rendered images in the standard farmhouse demo (which just shows how much of the screen real estate is wasted in this version).  The nose rest slides into a square tube and is replaceable.  The lenses are those of from the Dev Kit and are held firmly in place by two, friction fit, rings.  The screen and accelerometer are held snugly in place by a back plate which is attached to the back of the eye cup component.  The eye cup piece shown is printed as one piece and took around four hours to print.  The back plate was also printed as a single piece.


Head mount

Head-MountI found the job of actually making a comfortable head mount quite daunting.  I built a number of prototypes using combinations of elastic straps, headbands, a tight-fitting beanie as well as printed components but none of these provided both comfort and a good physical mounting point for a HMD.  I finally opted for a quick workaround in the form of an insert from a safety helmet from Home Depot.  The insert was not useable as is so I built some clasps that fitted exactly into the insert to form an adjustable X-shaped crown.  Once complete the insert worked well; it was adjustable, comfortable and quickly disappeared into the background like a well-fitting hat.  It also had two “studs” that protruded from either side of the headband at the widest lateral position as shown which I was able to use as the mounting points for the screen assembly.


Adjustable Screen Mounts

Side-MountThe two side assemblies proved quite challenging to make light, strong and adjustable.  This assembly, that joins the head mount to the screen, has to allow for horizontal rotation, axial rotation and length adjustment as well as being rigid and light.  My solution was to break the side mount into two sliding rails, a rotating end connector and a head mount clasp as shown.

Putting it together

All the various components fitter together easily using standard fasteners available from any hardware store thanks to the precision of my wonderful Printrbot Plus.

Fully assembled the whole thing weighed around 200 grams without the Oculus electronics.

Testing it out

I really liked the experience wearing the completed head mounted display.  The adjustable headband was super comfortable, I did not have anything touching my face apart from the nose rest and did not get the “hot face” that I get very quickly with the standard Dev Kit 1 mount.  The one weakness of this design is that the full pivoted weight of the display had to be carried by the nose rest making it uncomfortable to wear after a short period of time.

I had the chance to show the display at our awesome local Vancouver VR meetup and received a very positive response from the group, many of whom had a chance to try it on.  They all agreed that the weight on nose rest was the one failing, but, apart from that, found the experience very pleasant.  The lack of blinds to block out peripheral vision actually received a positive response from several testers.  They felt that it was not at all distracting and, in fact, allowed them to sense when others were approaching them so they could remove the headset and be social.  Interesting.

Next steps

The key focus will be to reduce the weight of the pivoting screen mount as much as possible as well as look at ways of moving the weight away from the nose, perhaps onto the cheeks.  The rest of the device is working well.  Once I have worked out how to have the nose rest have just the right amount of force to keep the eye pieces registered correctly with the face and the remaining weight taken up somewhere else I will move on to making the optics adjustable.  Any comments and suggestions are most welcome.



Measuring Interpupillary distance

Interpupilary-DistanceOne of the key physical requirements of VR headsets of any type is to be able to adapt to the viewer’s individual facial and optical characteristics.  A key component of this is interpupillary distance.   The distance between pupils varies considerably from individual to individual with the extremes being 52mm and 72mm.  As you can see in the image, if you are unlucky enough to be at the extremes of the range shown, you are not going to get a satisfactory experience from a fixed width head mounted display.  Oculus attempts to address this by moving the virtual eyeballs to the correct setting when it renders the two images but there is no physical adjustment on the Oculus Devkit 1 and, to the best of my knowledge, the Devkit 2.  From what I can see of Sony’s Morpheus, I don’t think you can adjust the lateral distance between the lenses either.  I would like to incorporate this feature into my headset designs at some point, and, minimally, set the interpupillary distance for my headsets to be the average by default.

In order to adjust a head mounted display to the correct interpupillary distance, you need to know what your personal distance is.  To that end I developed a series of physical measuring devices that culminated in the following prototype.  Interpupilary-Distance-RuleThe idea is that you can hold the device in front of your eyes and then slide the two holes closer or further apart until each pupil is centered in the hole as seen reflected in the image of the mirror.  Once set the device can be placed on a table and the distance between the two eye holes measured.

It worked well when standing in front of a mirror in good light and was portable, but when I tried to use it on some volunteers it did not so work well.  The usage model was unintuitive to them and without a stable mirror it was virtually impossible to align it correctly.   I will build a new version that uses long parallel tubes with a light source at the end and perhaps incorporate a measurement scale or Vernier system in my next prototype.


A new beginning (Part 2)

From there I went on a very wide and circuitous path ever deeper into complex software systems, big data etc etc which ultimately led me into the intelligence space where I have been lurking until recently.  Now, with the rise of consumer 3D printers, 3D television and the Oculus Rift I have the opportunity to return at last to my passion, which leads me to this blog.

My goal is to share my thoughts and dreams about how the lines are being ever more blurred between the virtual and physical worlds as well as my personal adventures in this arena.  Welcome and enjoy 🙂