NOTE: Due to the shear amount of data involved I will be updating this post multiple times over the next few days adding additional images, documents (research, bill of materials, etc.), etc. (there really is a lot to share).
Quick Specs:
-Project Name: Infinity Display Engine
-130+ Deg FOV @ 100% overlap)
-1280x800 full resolution per eye
-3D stereoscopic ready
-Large eye relief (eyeglasses compatible)
-Fully prototyped and working
Hello all.
The release of Palmer Luckey's HMD represents something incredible for the gaming, VR and simulation communities. If you're listening Palmer I commend you for your persistence and accomplishing all that you have - how far you've gotten is really nothing short of amazing!
As happy as I am for Palmer and his success I must admit that the news initially had me (quite selfishly) bummed. Like Palmer, I'm also a longtime VR enthusiast and have also been working on my own large FOV HMD design that I hoped to release as a product one day. It has literally been a passion of mine for many many years. Over the course of that time I've made so many great strides and have gotten so close… But living life (working, getting married, having a child (and just recently another!
) has made it a bit too easy for me to procrastinate...and unfortunately the years have slipped by and of course I never got around to releasing it. Let that be a lesson to anyone out there with an idea that you're passionate about... don't procrastinate! Make it a priority and hammer away at it with great intensity until it's realized!Now with Palmer releasing his offering I believe there's no reason to continue my sporadic development under a shroud. In fact I really should have shared and collaborated with the community from the beginning - so much more progress could have been made that way. So, while a bit late... I'd like to share my design with the community in the hopes that it may still be useful or even spur some new ideas. If you're reading Palmer I must say it's so crazy to read through your threads - although we were on different timelines there is so much overlap between your R&D and mine. Even much of the prototyping I've spied in your videos is very similar to things I've built as well.
Now with all that said I'll add this (and this was hard for me to swallow) - Palmer's design is superior to mine. IMO his design is inherently superior on 4 primary fronts - simplicity, weight, weight distribution and cost. The only item my design really wins on is resolution. However this is only because I use a separate imager (lcd panel) for each eye - and even with that it's only a temporary win as eventually a panel will be available that will surpass visual acuity even when split in half as Palmer's design dictates.
The basis for our designs share a common theme: Large single-element non-pupil-forming optics combined with relatively large imagers. We've also no doubt faced many of the same challenges along the way - such as determining lens specifications that would allow for the largest FOV possible for a particular panel, while at the same time still offering a comfortable eye-relief and as close to on-axis centering to the pupil as possible for the projected user demographic. Did I mention also needing to leave someplace for the nose to go? Damn nose always screwing up a good, simple optical system.
Where our designs go after those initial similarities is vastly different. Palmer's design trades resolution for simplicity and performance...where I traded simplicity and performance for resolution.
Palmer's Rift uses a very simple (and very efficient) optical design consisting of a lens assembly and a single imager (split, with each half dedicated to an eye). (As a quick aside... It's funny (in an "I'm kicking myself now" kind of way) as I actually considered Palmer's approach early on but completely discredited it because of the lack of resolution - props to Palmer for having the foresight to keep things simple and design around a panel that would obviously be there eventually...Good move Palmer.)
So, my design makes use of two imagers - dedicating the full field and resolution of each to a corresponding eye. This is accomplished via a slight twist on the 'Planar 3D' configuration many will be familiar with. For those unfamiliar, the planar configuration allows for 3D stereoscopic viewing by making use of 2 panels in conjunction with a beamsplitter and polarization to ensure that each eye only "see's" it's corresponding panel. The 'twist' mentioned is as follows... while stereoscopy is an inherent byproduct, this particular configuration is employed in my design for another reason entirely. With the 5"-7" panels that are common to both of our designs it would be physically impossible to center a panel to each eye without one of the panels being visually obstructed by the other due to the comparatively inadequate size of a human's IPD (inter-pupillary distance). This limitation can be overcome, however, with a slight modification to the planar configuration. In the standard planar setup two LCD panels are laterally aligned to each other. In my design, however, the panels are laterally aligned the pupils of the eyes. With this approach each eye is able be at 0 convergence (looking straight ahead into infinity) while having an entire panel (of just about any size) visually unobstructed and centered to that gaze. This approach paired with a simple set of plano convex lenses (or the custom lens assembly I've designed and cnc'd) allows for an HMD with a very large Field of View (130+ deg) at 100% overlap and a high resolution (high as the panel offers) while still maintaining a very breathable (eyeglasses compatible) eye relief.
While there are many additional items of note, in my opinion the primary tradeoff with my design when compared with the Rift's is weight and weight distribution. The optical system itself is inherently heavier than the Rift's (due to the additional panel and beamsplitter) and will make an HMD quite a bit more front-heavy than Palmer's Rift without proper counter-balancing (but thereby also increasing overall weight, Inertial mass , etc.)
In conclusion, with my current understanding of the Rift's design, if the highest resolution possible is the driving factor for a particular application (and the resolution offered by the rift at that time is not acceptable for that particular application) than I would suggest my design. Otherwise at this point I believe the rift to be the better choice and would suggest it whole heartedly. It is a simple, elegant solution and with a proper consumer grade enclosure (and eventually a very high resolution panel) there will be nothing better IMHO regardless of price point.
Lastly, to Palmer: I wish you luck in your journey - you've done a fine job and I'm envious.
If you ever wish to collaborate with someone else who truly shares your passion for HMD design please don't hesitate to get in contact.-Mike Krzenski
Below you will find many items related to my project - Everything from Images of prototypes I've built.. to sketchup and solidworks files...to bill of materials... and anything in between. As mentioned at the head of this post there is a lot of data and I'll be coming back and adding additional files, etc. as I have time.
Some things to note:
-The wording 'Infisense, Inc.' (as seen in some of the photos) is the corporation (still active) that I registered for handling the business side of things.
-As mentioned I'm releasing this information for the good of the community so that other's interested might benefit (building one for themselves, spawning new ideas, etc.) However, if you're simply taking my hard earned research and then turning around and mass producing it (or including it in your own product that you're profiting from) I would like to be fairly compensated.
I don't even care about head tracking. Just ghost free stereo3d with a large FOV is enough to blow my mind! 
