Hi Palmer,
"Cb"wan told me about your "Oculus/Rift" project, in the beginning I thought it was just a chip diy project but after reading your post, I think that's really impressive... I hope you'll achieve it!
If you're still in need for logos, I'm no pro but I would be glad to help you for that part ( btw, I think Oculus is a very nice name)...let me know if you're interested: balthazar.auxietre@gmail.com

Good luck!

The res is 1280x800 split in half. It should be able to use VGA, HDMI, and DVI connections. Still figuring out the best way to do drivers.



Great! That is exactly what I am hoping people will do, write support for it into their own games. Integrated head tracking is probably a good idea as long as it can be done without a huge cost, which looks very possible now. If people want more advanced tracking (Like the Hydra) they can do it themselves.



Your tracker probably is a good choice, it would fit this project well. 1.5x1.5 is fine, size wise. USB is probably fine for most people, but bluetooth would be good for people using wireless video links for standup rigs. I would love to get in on whatever first batch you make for testing, and if it works well, we can probably work out something from there.

Everyone else, thank you for the interest! Looks like there are about 20 people who have said they would want one for sure, that is probably enough to hit at least a third of the Kickstarter goal I have in mind.

If your looking to gauge interest I'd definitely get one, and I might be able to convince a few friends as well...
Looking forward to it Palmer

I haven't started to research the cost of adding head tracking to my games, but with so many products in the market (without a standard API) it may not be so easy. However, I don't understand why only a few games have native stereoscopic support and none of them offers a configurable fov (without hacking config files). The cost of adding those features (in any game engine) is so small in comparison with the benefits you would get... I really don't understand.

I suppose most innovations comes from indie developers nowadays: while many indie developers try to make money finding a niche and making good products for that small market, most big studios are working in multi-million projects and are too afraid of trying new ideas because of the financial risk involved.

I most certainly am intrested, but have very little cash. Will this still be available in many months when I have saved up enough?

If you support the Freetrack API most trackers can be supported through FreePIE or GlovePIE

Thanks for the information, I will study that approach. I don't want to hijack this thread, so I will start a new thread if i need more information about Freetrack and other motion tracking systems.

No worries, just one more thing, Freetrack only takes care of headtracking. For moving the entire view (that you normally do with the mouse) you could support absolute joystick values

Don't forget us lurkers! I'll want one for sure.

The trouble with Bluetooth is the inherent latency. End to end, it'll be at least 50ms and probably over 100ms. You can below that if you license access to CSR's bluetooth stack and tune the HCI, L2CAP, and RFCOMM layers for latency instead of using whatever the premade modules are set up for, and also develop a USB dongle to use on the desktop side to avoid getting ruined by whatever the Bluetooth stack there is set up to do. That would require making a lot of the things for it not to add a lot of expense per unit.

The latency is acceptable for controllers, but for something controlling your field of view, it might be nauseating.

While I agree latency is something undesirable for "something controlling your field of view", I wonder if it would be so bad. Most multi-player games don't send a network package every frame (but every 100 ms) and use interpolation and movement prediction algorithms to offer the player the feeling of smooth movements. Maybe having that latency isn't so bad, we talked about redirected walking in this thread (http://www.mtbs3d.com/phpBB/viewtopic.php?f=120&t=14069&#p67559) and it seems our brain can accept a small disparity between our real head movements and the virtual viewport movements, but I suppose we won't know until we try it.

Bluetooth in my (admittedly little) experience is laggy, supported by all sorts of wierd implementations that may or may not work, and wastes a lot of power on mobile devices.
Most of the decent wireless solutions that i've ever seen, seem to have their own dongles and systems.



I think you are underestimating the latency issue. I've experienced it on the Virtuality SU1000 setup (not that I needed to, the detrimental effects of lag in the head movement/display loop are pretty well documented) and its not going to be acceptable to have 50ms of lag shoved on top of your frame times. You can not interpolate this - while a tiny amount of interpolation may be useful just for anti noise purposes, any substantial amount will just cause similar immersion issues to the lag in the first place. The issue is that you turn your head, and the world takes a moment to 'catch up' - this breaks immersion.

I believe the redirected walking is mainly working on the effect whereby a small movement can be simulated in VR, and feel like a large one... and vice versa
That means that when turning, you can have the system 'overturn' or 'underturn' you. When moving straight, it can slowly rotate the viewpoint at a speed under your recognition. This has nothing to do with head lag IMHO.

I don't have the hardware to test it myself (yet) so I'm probably underestimating the effects of the latency, I was only trying to suggest a possible solution to reduce those effects. I thought that, as the brain can accept some disparity between the real world movement and the virtual movement (used in redirected walking), we could get away with a small lag between the real movement and the virtual movement.

I'm probably wrong, but if somebody has the required hardware and want to give it a try...

EDIT:
A quick google search has proven that I was wrong.

i sent you a pm, palmertech. i'm hyped about this project, just wish the resolution was a bit higher!

Shouldn't bluetooth 2.0 latency around 30ms to 50ms? Hopefully new gen of bluetooth 4.0 will bring that below 10ms.

I want one for sure too.

I'd be interested in one of these for sim racing on PC.

hi Palmer - how is it looking in terms of your June launch date? I am saving up those pennies!

How much do we know about the specs then...? The resolution is about 640x800 for each eye, and should be able to take VGA and HDMI. What about the FOV? 70 degrees? Are these things still being finalized?

Not going to launch in June, but I will have the Kickstarter project up in June. With luck, that will mean shipping in July!



The FOV is still being finalized, mainly because I am still trying to strike a balance between cost/exit pupil/ease of assembly/FOV. I want to have a FOV of at least 90 degrees, since that is the maximum that most games can render by default. I can go as high as 120, but it will probably be a little less than that.

Progress on the project is stalled out for about a week while I buckle down on finishing a different project that is on a deadline, but full steam ahead after that.

Perfect - thanks for the update! Can't wait!!!

90 degrees FOV is good enough. I just hope a version with 1080p panel is possible this year. With all smartphone and tablet going for retina display, we should be able to get one soon. Palmer's design is modular and allow modder to upgrade the panel when it's available.

Just a heads up that I'm still planning on working at branding for you Palmer but I'm currently trying to wrap a very big job. I might not get to it until late May (3rd week in May) as I'll be out of town as soon as this job is done. So, if that's too late for you, no worries if you need to pull in someone else.

Good luck with everything and I'm still super excited about this project!

You might have heard that IZ3D got rejected from KickStarter for their open source driver. KickStarter said it was not a "creative arts" project. Do you think this HMD project might also get rejected? What other options do we have in that case?

If they tried to set up a Kickstarter with a $800,000 goal, that probably didn't help their cause. No successful Kickstarter project has ever set a goal close to that high, let alone one to release source code for an already complete product. Kickstarter exists to bootstrap development of a project, which was not iZ3d's intent.

It clearly is PalmerTech's intent though, and Kickstarter is unlikely to reject it.

Yeah, I don't think there has ever been a project set at $800k. Even if there had been, the IZ3D thing is not paying for some new creation, it is basically to cover their past business losses in exchange for releasing something they already made.

There have been a lot of things on Kickstarter that are similar to this. The TOOB got on, it was an immersive spherical projection display, and there have been countless gadgets and machines.

If we do get rejected, IndieGo is a viable alternative.

Mmmm. I'm curious on quite a few points.

What parts of this HMD are "open-source", just the software? What about the optics and mechanical side of things? ;P (The driver board and panel being a sourced part, is pretty obviously not going to be. But lol. It's also not all that unique from the sound of it.)

How does the performance look thus far?
How much distortion is there, in particular?
Also rather curious on how much space the optics take up and how far from the eye they sit ^.^'

;P Honestly, a good compilation of information on this project in one place would do it some good.


^.^' I imagine yer pretty buisy pulling it together for kickstarter. Some sort of comparison of playing immersed on a HMD vs on a screen might make a good demonstrative item. Sort of a geek-sitting-slightly-too-close-to-a-decent-sized-monitor vs your HMD design comparison. ;P Hopefully to a flattering end.
Also, lasers.

The optics and hardware will also be as open source as possible. If someone wanted to, they could purchase and fabricate all the parts themselves, the goal of the kit is to keep the costs down through group purchasing and fabrication. The original plan was to make everything out of laser cut sheets, but I might vac-form some of the parts to keep the HMD lighter; I am going to make the plans (And probably parts) for the highest possible FOV available, but if I compromise to around 95 degrees, then the weight can be cut right in half.

A compilation of all the info on this project is definitely needed, I am going to put that on the Oculus website. I am working with some friends who work in the film industry on coming up with ideas for the Kickstarter demo video, because if that can get the general idea across, I think a lot of people will want one of these! The performance is pretty good, there is no distortion in the central area of the image, though it gets worse in the periphery. Luckily, you can correct for this in software.

I think 95 degrees is a very good compromise for the consumer market (specially if we take into account the fov of the competing products) and I'm sure the weight reduction will improve the comfort. However, a larger fov could be very useful for research... as this project has been planned as a highly modular HMD, you should probably create several pledges in KickStarter to allow us choose if we want to purchase the 95-degrees-optics, the optics with a larger fov or both of them.

I agree; a 95 FOV with a lighter form factor will probably be better as a consumer product, but if it could be easily modded to 120 FOV, that would be perfect. I personally would prefer the larger FOV.

I can't really say which I would prefer. Having experienced nothing even close to 90 degrees, it's hard for me to imagine what the difference between 95 and 120 would be. I guess software support is important, so if most software only supports up to 90 FOV then 95 seems like a reasonable size - especially if it dramatically increases the wearability of the unit.

The higher the FOV the better, but 90+ is acceptable to me.

For complete immersion, you want all you can get. I've only ever used the 51° HMZ-T1, but people here have said that even at 120-180°, you can still see there are edges. Palmer created a 270° prototype. Any chance of a Kickstarter for that? That would spread the resolution out too far, but would be great with that really hi-res panel Palmer mentioned.

I think PalmerTech's 270 degree prototype used 3 screens in a Eyefinity type setup. It might be difficult to use a single flat screen to provide 3 views due to the optic path needing to be bent (total guess).

Yes, that makes sense. He mentioned it might require a boom mount because of weight.

I think that one important thing that can be done to improve immersion, regardless of the field of view, is to have the limits of the field of view be defined by physical boundaries near the eye, rather than by running out of pixels at the focal distance of the display. Many of the “practical” uses of HMDs want to see every last pixel straight and true, but that leaves the obviously unnatural sense of moving borders out in the distance. If the visible edges are put between the eyes and the optics, it will scan to the brain like a real helmet / glasses / goggles / glasses that occlude the world. This unavoidably will leave display pixels that can’t actually be seen, potentially a fair number of them if you have a generous eye box, but it is probably the right call for immersion.

John Carmack

Are there currently any HMDs out that feel more immersive than traditional 3D displays? Still trying to decide if it's worth investing in one, considering their price is fairly high in comparison to a 3D monitor, which would be my other choice.

@John: An interesting point. I can see the logic, but it seems so tragic to give up pixels and FOV when we have so little to spare.

Keep in mind, also, that the vertical FOV of this thing is also going to be around 90 degrees, which is much taller than most HMDs.

The 270 degree prototype is fun, but just not practical. It is big, heavy, bulky, and you need to render 4 distinct views. Even if the software supports it, the rendering load is massive.

John is absolutely right, as far as cutting off pixels go. There are benefits to having distinct edges for "normal" use, in addition to saving pixels; You get an improved senses of "Coming out of the screen!!!!" that competes well with 3DTVs. For VR use, though, having boundaries does not help with immersion. Had I not spent a lot of time with a Wide5, I might have felt differently, but it really does make a difference! Giving up pixels seems bad now, sure, but displays will only get better.

95 degrees sounds pretty good to me. I think the last prototype I tried was slightly more than that, and it was more than enough. In terms of the vertical, it covered the full FOV. Granted, you could still see the edges on the sides, but it wasn't anything like the tunnel vision of most cheap HMDs. Considering most non-custom software won't even look right at super high FOVs, this is probably the best bet for the majority of people.