Is 2560x720 the ideal resolution?

[krimms]

2560x720 resolution would mean you'll see the game as it was originally made, at 1280x720, with no modifications. You won't see it at 640x800 like you do now. All games support 1280x720, so Oculus Rift would be compatible with them by default from this point of view.

However, although this would make games compatible, I'm not sure of the impact on the eye. As you're basically seeing the game as if it were with one eye, no? Because in that case I think the viewing space is a lot more square than 16:9. 16:9 may only make sense when you see an image with both eyes.

How does the image appear when you use the Oculus Rift? Do you see it in 640x800 with both eyes, or the whole 1280x800? I haven't used one and I don't know. If you're seeing the whole resolution, then something like 720p or 1080p makes sense, as you'll see the game as you normally would. If you only see "half" of it, then you have to find out what's the right resolution ratio for each eye. Is it 4:5? Is it 1:1? Is it 7:5?

Someone needs to get the eye science for all of this, as in the end it's probably more important for the games to be "eye-native" than maintain compatibility with the old games. If Oculus Rift is that great, which I think it is, then the "new" games will arrive for it anyway. If you see the whole screen's resolution with it, then that's perfect as old games will be compatible, too. If you can't, and you're only seeing half, then the resolution will have to be adapted to "perfectly" match the eye's viewing space.

[PasticheDonkey]

i don't think it would help in the way you are thinking. the rift would work better with a 2:1 ratio screen so each eye gets a square that could then be warped into circles, you know like your eyes are.

normal tvs are 16:9 because that is close to a phi ratio. and people like that kind of shape. it's in painting, architecture, the way things a framed, music etc.

keep in mind that for 3D games it's fairly trivial changing the resolution on pc.

[krimms]

i don't think it would help in the way you are thinking. the rift would work better with a 2:1 ratio screen so each eye gets a square that could then be warped into circles, you know like your eyes are.

normal tvs are 16:9 because that is close to a phi ratio. and people like that kind of shape it's in painting architecture the way things a framed etc.

Yes, I realized that may be true after I posted, and then I edit it for more questions. But then wouldn't a 800x800 resolution make more sense? And I'm not sure if with one eye it's perfectly square (or round).

After a quick Google search I found this, which says one eye can see at a ratio of 1.17:1

http://wiki.answers.com/Q/What_is_the_a ... _eye_sight

If it's accurate then each eye's resolution needs to be more like 800x683. So almost the other way around than it currently is. So if you can only see with one eye at a time, therefore making the 16:9 ratio inaccurate for one eye, then at least the display should be as close to 1.17:1 for one eye as possible.

In the end the current ratio is close enough for 1st gen, but maybe for the second generation...

[MSat]

There is no readily available "ideal" display in terms of both size and resolution that Oculus could use. The 7" form factor is the best choice for the developer kits.

[pilzbefall]

the actual ratio is a reasonable compromise. In simple words: 2560x720 isn´t available.

[Inscothen]

There was a 2:1 aspect ratio lcd Sony 8" 1600x768 which would have been great for a DIY Rift type HMD, if only the specs were better. it's not in production anymore. It was basically a little taller than the 5.6" lcd. but the width would have filled the horizontal fov of the lenses

[cybereality]

Thats kind of a strange resolution/aspect ratio so I doubt you will find many (any?) displays with that format.

Though I do wonder if a Rift-like HMD could be made with two smaller cell-phone screens (ie around 4" each). You would lose some vertical FOV and maybe introduce sync issues, but the aspect ratio would be better suited for a lot of existing content (ie movies, most games, etc.). Something to think about.

[MrGreen]

Oculus/VR needs to grow enough to have custom panels built specifically for it.

I'd say we're off to a good start.

[Inscothen]

Thats kind of a strange resolution/aspect ratio so I doubt you will find many (any?) displays with that format.

I bought two of those 8" lcd for cheap and was going to make a couple rift type hmd. but I didn't realize how much of a mess the panel/controller compatibility was between manufacturers. I was told by some lcd controller manufacturers that the panels wouldn't work with their stuff.

I am learning this stuff slowly. I think if I had a custom lvds cable for the panel and a chalk-elec hdmi to lvds converter , maybe I could make them work. The specs of the displays are really poor though.

[Moriarty]

Oculus/VR needs to grow enough to have custom panels built specifically for it.

I'd say we're off to a good start.

Yes, I was wondering about this too. It's clear that the smartphone/phablet market is producing a lot of potential display candidates but latency etc is not an issue for these companies. How big do you think Oculus needs to get before it can have the luxury of dictating its own design specifications and order a panel that is tailor made for a great VR experience at a reasonable price, 100 000 units ?...millions of units ?

[EdZ]

I'm not sure they need custom panels, more a custom controller. MIPI and eDP are emerging as the two connection standards to replace LVDS, MIPI moreso than eDP. Unfortunately, while eDP should be relatively easy to interface with being very close to DisplayPort, there is currently no way to connect a MIPI display to anything other than a SoC. Development of a custom controller to convery DVI or DisplayPort to MIPI would be complex and expensive, and require either an expensive (on the order of £30-40 per chip, due to the high bandwidth required) FPGA/CPLD or enough of a guaranteed run to justify a custom dedicated chip. This would have the benefit of making dual-panel setups more feasible, due to the fine control over timing.

If a commercial DVI/DP -> MIPI interface became available (with an acceptable low latency) then that opens up a lot of 5"-10" high-DPI panels that are currently unusable. If more panel manufacturers build eDP panels, those would also be feasible to work with, as most PC GPUs (and many laptops) now have DP outputs.

[PasticheDonkey]

we know they'll be able to get a 1080p panel and then probably a 4k panel since those exist and one of thems gotta have suitable specs other than res at some point. now the occulus would never need more than 8k but no one's going to make one for a tablet. but since it's the last they'll need and it'd be the only display going forward, if they can make enough money off the 1080p and 4k version then they perhaps could afford the R&D costs to develop an 8k panel of the right size. or some new display tech could drive the production of smaller pixels that could be more easily adapted by a vendor to fit the requirements. holographic tvs using lenticular screens would need higher and higher res for example.

[Randomoneh]

Now the Oculus would never need more than 8k but no one's going to make one for a tablet. but since it's the last they'll need and it'd be the only display going forward...

Eventually, there will be 8K at 7 inches and even more that but let's not get ahead of ourselves. There is really no need to discuss those things this early on.
Time will come, and it'll happen.

[PasticheDonkey]

why would you need a dpi beyond high quality print?

[Randomoneh]

why would you need a dpi beyond high quality print?

Eventually we'll have enough computational power that today's resolutions, including 8K will be a piece of cake to process.

At that point, it won't be a problem to build a display that'll match, let's say - Vernier acuity at minimum distance an eye can focus on (few inches).

Basically, to imagine what the future might look like, first you have to imagine you have infinite computational resources, infinite storage and infinite bandwidth. Now figure how you'd be able to benefit from it.

[cybereality]

why would you need a dpi beyond high quality print?

Because you don't look at magazines with huge magnifying glasses that make the image wrap around your entire vision.

[PasticheDonkey]

you haven't followed the conversation cyber. i am positing that resolution for a tablet needn't go beyond high quality prints dpi. there for they will not be a path to the res required by the rift after than point. which needs higher for the reason you just gave.

[PasticheDonkey]

why would you need a dpi beyond high quality print?

Eventually we'll have enough computational power that today's resolutions, including 8K will be a piece of cake to process.

At that point, it won't be a problem to build a display that'll match, let's say - Vernier acuity at minimum distance an eye can focus on (few inches).

Basically, to imagine what the future might look like, first you have to imagine you have infinite computational resources, infinite storage and infinite bandwidth. Now figure how you'd be able to benefit from it.

I don't believe that is an accurate way of assessing the future. things work on a needs must basis. hence why sounds resolution is generally 48khz right now even though higher could be available easily, and detectable by some. no matter the resources you've got it's best to put them to some actual use rather than waste them on things too many don't care about.

[Randomoneh]

why would you need a dpi beyond high quality print?

Eventually we'll have enough computational power that today's resolutions, including 8K will be a piece of cake to process.

At that point, it won't be a problem to build a display that'll match, let's say - Vernier acuity at minimum distance an eye can focus on (few inches).

Basically, to imagine what the future might look like, first you have to imagine you have infinite computational resources, infinite storage and infinite bandwidth. Now figure how you'd be able to benefit from it.

I don't believe that is an accurate way of assessing the future. things work on a needs must basis. hence why sounds resolution is generally 48khz right now even though higher could be available easily, and detectable by some. no matter the resources you've got it's best to put them to some actual use rather than waste them on things too many don't care about.

I believe there is no debate on whether you can or cannot benefit from extra resolution when you're two inches away from an object. At this day and age, displays are made to be "good enough" for average viewing distances. As things become cheaper, they won't be made for "average viewing distances" but for "as close as you can get to it".

[PasticheDonkey]

well i think high quality print is at that res that you can get as close to it as you like and see no need for improvement. without a magnifying glass of course.

[Endothermic]

Unlike sound though Visual resolution wise it will be noticeable. 8k will only be 3840x4320 per eye which while definately great, at a 90 FOV you will notice a difference and want that difference with a higher resolution. FOV also won't stay at 90 forever either so as FOV increases the more the max res threshold will go up.

There is a point where a higher resolution may be technically perceivable by a person but they simply won't notice the difference however 8k is not that resolution even at a 90 FOV.

[Randomoneh]

a point where a higher resolution may be technically perceivable by a person but they simply won't notice the difference however 8k is not that resolution even at a 90 FOV.

Sh*it, this is my specialty and I want to stay but I just have to get some sleep. Don't have too much fun without me. Last piece of advice: 1 arcminute is not enough! Good night...

[PasticheDonkey]

i think the rifts properties of having higher res towards the centre than edges may make it enough. if it's not one doubling later and it will be.

[PasticheDonkey]

so was just thinking that at some point it could be better to use a small rear projection dlp. because of the enclosure the light doesn't have to be bright so no heat or power usage problems. that is unless those chips run hot.

[Dilip]

@PasticheDonkey

Here you are may like to check EPSON MoVeRIO BT-100 it was a product that elimineted need of CAM for AR as it was projecting on prism jaded on plain glass just like normal specs but were bulky for normal specs style.

besides epson does BS JOB to not to allow any video source other then their own closed ecosystem & they were way costly too,

http://www.expertreviews.co.uk/gadgets/ ... rio-bt-100

this also explains why it got failed

something with twin pico projectors with 1280X720 twin display can have rest alll things ready like games & middlewears like DDD & NVIDIA

some one need to DIY in that direction too.

[cybereality]

@PasticheDonkey: OK, I see what you are saying now. Even so, technology seems to advance even without clear reason to. Companies will always want to up the specs of their hardware so they can sell more units. I mean, you think when the first 8K 120Hz tablet comes out all the companies are just going to pack up and stop making new devices?

[PasticheDonkey]

nope they'll find new differentiators to compete on. the sort of stuff SD TVs did for decades. and add new gimmicks folding or roll up screens etc.

[PasticheDonkey]

actually they may try for autostereoscopic 3D on a tablet at 4k and that would give the right screen for the oculus with whatever overlay is on it removed.

[EdZ]

a point where a higher resolution may be technically perceivable by a person but they simply won't notice the difference however 8k is not that resolution even at a 90 FOV.

Sh*it, this is my specialty and I want to stay but I just have to get some sleep. Don't have too much fun without me. Last piece of advice: 1 arcminute is not enough! Good night...

I'll drop another hint: look up 'Vernier Acuity/'Hyperacuity'. The human eye may only be able to distinguish separate lines or dots down to 1 arcminute, but distinguishing separate lines is hardly all your vision system does!

[Fredz]

Actually parallax barriers are now inside the display, between the LCD panel and the backlight. It's certainly possible to dismantle the whole thing like these guys did, but I'm not sure it's that easy for a production run or that the LCD display would still work correctly after that.

[PasticheDonkey]

a point where a higher resolution may be technically perceivable by a person but they simply won't notice the difference however 8k is not that resolution even at a 90 FOV.

Sh*it, this is my specialty and I want to stay but I just have to get some sleep. Don't have too much fun without me. Last piece of advice: 1 arcminute is not enough! Good night...

I'll drop another hint: look up 'Vernier Acuity/'Hyperacuity'. The human eye may only be able to distinguish separate lines or dots down to 1 arcminute, but distinguishing separate lines is hardly all your vision system does!

well the test on this page http://michaelbach.de/ot/lum_hyperacuity/index.html uses antialiasing to test you below sub pixel limits. so the smallest pixel your eye can see may be the only limit required to be met. something in the space between those pixel would be represented on 2 of them.

[PasticheDonkey]

Actually parallax barriers are now inside the display, between the LCD panel and the backlight. It's certainly possible to dismantle the whole thing like these guys did, but I'm not sure it's that easy for a production run or that the LCD display would still work correctly after that.

yeah but at least if the tech exists to make the right sized pixels then an arrangement can be made to have custom batches made. as long as the order is big enough.

[Randomoneh]

well the test on this page http://michaelbach.de/ot/lum_hyperacuity/index.html uses antialiasing to test you below sub pixel limits. so the smallest pixel your eye can see may be the only limit required to be met. something in the space between those pixel would be represented on 2 of them.

I'd disregard that webpage. It really seems resolution-dependant. I get low and from then images are absolutely the same.

Vision is really not a 1+1=2 thing. There's minimum discernible acuity, minimum separable acuity, hyperacuity, Vernier acuity...

EdZ, you're going to like this paper from 2003(!) by Darrel Hopper (Air Force Research Laboratory) titled "Capability of the Human Visual System".

It is interesting that Darrel Hopper also wrote some papers on the subject of military HMD displays and how many pixels is "good enough". He's a chairman of the DoD (USA Department of Defense) Roadmap for Displays . Who wants this job?

Also an amazing paper from 2005.

[Endothermic]

One thing i've always wondered but never seen a paper on is motion perception in relation to resolution. Well i've seen some but they are all on like the difference between lower resolution and higher resolution nothing on the maximum percieved human resolution and going beyond that.

So you have a screen with a resolution and at a distance that it's the maximum you can interpret, so if you double or 10x the resolution you still do not see anymore additional detail it looks excatly the same.

Even though you can not see any additional detail with the higher resolution screen the pixels are there, smaller and changing, so is your eye/brain able to detect when those smaller individual pixels change even though you can not actually 'see' them, hence having you respond quicker to movement or maybe not quicker but notice more movement then on the lower resolution screen which looks just as detailed or is your ability to detect motion tied to your ability to percieve detail?

[PasticheDonkey]

well the test on this page http://michaelbach.de/ot/lum_hyperacuity/index.html uses antialiasing to test you below sub pixel limits. so the smallest pixel your eye can see may be the only limit required to be met. something in the space between those pixel would be represented on 2 of them.

I'd disregard that webpage. It really seems resolution-dependant. I get low and from then images are absolutely the same.

Vision is really not a 1+1=2 thing. There's minimum discernible acuity, minimum separable acuity, hyperacuity, Vernier acuity...

EdZ, you're going to like this paper from 2003(!) by Darrel Hopper (Air Force Research Laboratory) titled "Capability of the Human Visual System".

It is interesting that Darrel Hopper also wrote some papers on the subject of military HMD displays and how many pixels is "good enough". He's a chairman of the DoD (USA Department of Defense) Roadmap for Displays . Who wants this job?

Also an amazing paper from 2005.

even at .05 they are different you could check the colour information in a paint program.

vernier acuity isn't a measure of the resolution the human eye can see at. it's the measure of how well it can discern past it's resolution using colour information. and it'll do the same on a display that is actually good enough to fit it's resolution.

[Mystify]

One thing I haven't seen discussed:
if you have mroe resolution, but the graphics being displayed aren't of a higher quality, how much improvement does it actually yeild?
Modern games still aren't photrealistic, so to some degree, more resoultion just lets you see the imperfection better.

[Inscothen]

One thing I haven't seen discussed:
if you have mroe resolution, but the graphics being displayed aren't of a higher quality, how much improvement does it actually yeild?
Modern games still aren't photrealistic, so to some degree, more resoultion just lets you see the imperfection better.

less aliasing, and seeing things more clearly that are further away are two benefits of higher resolution

[PasticheDonkey]

One thing I haven't seen discussed:
if you have mroe resolution, but the graphics being displayed aren't of a higher quality, how much improvement does it actually yeild?
Modern games still aren't photrealistic, so to some degree, more resoultion just lets you see the imperfection better.

there is a balance between resolution and texture and geometry detail that matters. but currently console games are under serviced in the resolution department because they don't have the power to push those textures and geometry as they are meant to be seen. on pc people with expensive powerful rigs render at higher resolutions than the screen that are then down sampled to be displayed. anyway processing power for driving those graphics will scale with resolution and high res displays can always show lower resolution material upscaled.

[EdZ]

if you have mroe resolution, but the graphics being displayed aren't of a higher quality, how much improvement does it actually yeild?

It depends on how much effort you put into scaling. Poor scaling may introduce a drop in quality over 1:1 pixel mapping for some images, but in the majority of cases, with proper scaling (preferably performed client-side rather than using inferior on-controller scaling) a higher resolution display will look better than a lower resolution display fed with the same resolution source.

[snorelab]

I would just like to derail the technical conversation for a moment and make a comment about predicting the future of display technology based on tablets and smartphones. The iPad was released less than 3 years ago. Many, many people didn't believe there was a niche for tablets. Now tablets are everywhere and the displays are manufactured at high volumes. The iPhone was released at a time when everyone was trying to make their phones smaller. Do you remember those days? That was less than 8 years ago. No one knew for sure if it would be successful. But in the end it became the Jesus Phone and people went crazy for them.

My point is that you may not want to prognosticate years ahead based on tablet and smartphone displays and how they will drive the market for head mounted displays. It is entirely possible that our phones and tablets will have a new form factor by the time 8k screens are being mass produced.

It is also possible that the Rift becomes Jesus Goggles and creates a brand new consumer device niche. Wouldn't that be sweet? After the hype we saw at CES it's not out of the question.

Perhaps I should create a new thread. Jesus Goggles: Could it Happen?