First Impressions From Rift Owners

Post Reply
Doc Ok
One Eyed Hopeful
Posts: 23
Joined: Sun Mar 17, 2013 7:41 pm

Re: First Impressions From Rift Owners

Post by Doc Ok »

lmimmfn wrote:
Mel wrote:
I just turned off anisotropic filtering and anti-aliasing and the Tuscany demo is waaay better.

Haven't tried TF2 yet. Maybe later tonight.
Im a gfx whore, so no AA on such a low rez screen as the rift would be a head wreck for me, did you try downsampling after AA to see if its a better solution to using AA regarding blur/screen refresh?
Oooh, it just occured to me that AA is applied during world rendering, i.e., before lens correction. AA depends deeply on the precise grid structure of the display, so if the grid is warped afterwards, all kinds of -- very subtle, admittedly -- weird stuff might happen.
lmimmfn
Cross Eyed!
Posts: 138
Joined: Mon Apr 01, 2013 6:44 pm

Re: First Impressions From Rift Owners

Post by lmimmfn »

Doc Ok wrote:
lmimmfn wrote:
Mel wrote:
I just turned off anisotropic filtering and anti-aliasing and the Tuscany demo is waaay better.

Haven't tried TF2 yet. Maybe later tonight.
Im a gfx whore, so no AA on such a low rez screen as the rift would be a head wreck for me, did you try downsampling after AA to see if its a better solution to using AA regarding blur/screen refresh?
Oooh, it just occured to me that AA is applied during world rendering, i.e., before lens correction. AA depends deeply on the precise grid structure of the display, so if the grid is warped afterwards, all kinds of -- very subtle, admittedly -- weird stuff might happen.
Ahh man you have such a head start on us regarding how things will work, lol, very well pointed out, i never even thought of the rift warping the image i.e lens correction, that creates a mess regarding AA, however i do think downsampling overall would produce a better image than in game AA followed by lens correction post processing, unless the lens correction post processing is a bit crap and doesnt base the correction off an original higher rez raw image.

*EDIT* having a bit of a think about it, i dont think any type of AA will work with the rift other than downsampling and even that will be prone to aliasing( and i would imagine in some cases severe aliasing ) appearing after the post processing for lens correction.
Last edited by lmimmfn on Sat Apr 06, 2013 7:03 pm, edited 1 time in total.
User avatar
KBK
Terrif-eying the Ladies!
Posts: 910
Joined: Tue Jan 29, 2013 2:05 am

Re: First Impressions From Rift Owners

Post by KBK »

Persistence of vision varies between individuals. Part of the wiring package that we all get, it is individualized to some degree.

One way to ascertain if this is the issue, is: does the person who sees 'blurriness', what do they think of DLP projectors? Are DLP projectors stuttery and very problematic in the 'dlp rainbow' kind of way? Did CRT monitors, back in the day, need to have their frame rates cranked, so you could stand them? Do you notice florescent light flicker?

If the answer moves toward being a yes for all of the above, then it may be a factor of low persistence of vision, in the individual sense. And the pixel flipping has risen as an issue.

This is part of the kind of things to ask, as part of the problem set of potential sources. To eliminate it as a source of the issue.

I had to ask this all the time, every day on the phone, when advising people on projector choices and Home Theater design criteria that is aimed at covering all potential problem areas. It plays out here, with addressing potential problem areas with the Rift and it's LCD screen base design parameters.
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

KBK wrote:Persistence of vision varies between individuals. Part of the wiring package that we all get, it is individualized to some degree.

One way to ascertain if this is the issue, is: does the person who sees 'blurriness', what do they think of DLP projectors? Are DLP projectors stuttery and very problematic in the 'dlp rainbow' kind of way? Did CRT monitors, back in the day, need to have their frame rates cranked, so you could stand them? Do you notice florescent light flicker?

If the answer moves toward being a yes for all of the above, then it may be a factor of low persistence of vision, in the individual sense. And the pixel flipping has risen as an issue.

This is part of the kind of things to ask, as part of the problem set of potential sources. To eliminate it as a source of the issue.

I had to ask this all the time, every day on the phone, when advising people on projector choices and Home Theater design criteria that is aimed at covering all potential problem areas. It plays out here, with addressing potential problem areas with the Rift and it's LCD screen base design parameters.
Definitely something to consider,but if the blurriness you're referring to is the motion blur, i think the consensus (from virtually everyone) is that it's a real problem, and not user-specific. As for the focus issues I was experiencing, the explanation given above about AA and warping sounds valid to me, so again, would be non user-specific.
3dvison
Diamond Eyed Freakazoid!
Posts: 718
Joined: Sun Oct 24, 2010 7:25 pm

Re: First Impressions From Rift Owners

Post by 3dvison »

Mel, over all, are you happy with the dev. kit Rift and would you take it over your Diy Rift ?
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

3dvison wrote:Mel, over all, are you happy with the dev. kit Rift and would you take it over your Diy Rift ?
I am happy with it, and I would take it over my DIY, but that's only because it has the tracker and a skookum housing...both things missing from my DIY. Also, I think the FOV is somewhat better on the Rift, as I can't see the borders, but I can on my DIY. But that may just be a difference in lenses. One thing for certain, though, is that my DIY's higher pixel density is a win.
Doc Ok
One Eyed Hopeful
Posts: 23
Joined: Sun Mar 17, 2013 7:41 pm

Re: First Impressions From Rift Owners

Post by Doc Ok »

lmimmfn wrote:*EDIT* having a bit of a think about it, i dont think any type of AA will work with the rift other than downsampling and even that will be prone to aliasing( and i would imagine in some cases severe aliasing ) appearing after the post processing for lens correction.
The crux is that right now lens correction is done by resampling a 1280x800 pixel image (after AA, mipmapping, and anisotripic and trilinear filtering have been applied) to another 1280x800 pixel image. That's pretty much the hardest resampling problem there is.

There is code in the SDK to render the world into a higher-resolution off-screen buffer in the first step, and then resample that higher-res buffer into the final 1280x800 frame during lens correction. However, all that code is currently dormant. They probably just didn't get it done for the initial shipping deadline. Proper antialiased downsampling during lens correction (it's based on a quadratic radial distortion formula) is surprisingly difficult to implement.

I'm sure the option will appear in a later SDK version.
User avatar
KBK
Terrif-eying the Ladies!
Posts: 910
Joined: Tue Jan 29, 2013 2:05 am

Re: First Impressions From Rift Owners

Post by KBK »

best would be pre-post frame analysis with nyquist added in, but that's quite a bit of latency introduced by crunching numbers, among the larger given problem set. That is, relatively speaking, the Teranex way. Some set parameters that 'appear' to be effective to the eye will probably end up being the order of the day. Which will require extensive testing.
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?
LukePoga
Two Eyed Hopeful
Posts: 89
Joined: Mon Aug 20, 2012 3:49 am

Re: First Impressions From Rift Owners

Post by LukePoga »

it would be good to have the option of internally rendering 2560x1600 and only scaling to 1280x800 at the last step. it would produce high quality results. and a good option for those with fast gpus
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

I just want to reiterate that I disabled all filtering/AA settings on my video card, and the Tuscan demo looks pretty darn good now. The focus 'issue' I was experiencing was directly attributable to those settings. I run the demo at 1920x1080 and let the Rift controller do the scaling. No focus complaints now, so that's good.

As for motion blur: I've pretty much established now that, contrary to my previous (poor) 'observations', the motion blur is the same for both head turning and mouse turning. With head turning, the blur of any particular object is easy to see because you can maintain a steady visual lock on the object as you rotate just by rotating your eyeballs in the opposite direction. But with mouse rotation (essentially, full-body rotation), for whatever reason it is much harder to maintain a steady lock on an object, so the blur is harder to discern (I think Mr. Abrash may have described this phenomenon, just in a way my brain couldn't grok).

[edit to remove nonsense]
User avatar
Dilip
Certif-Eyed!
Posts: 692
Joined: Sat Dec 13, 2008 9:23 am
Location: Ahmedabad//INDIA

Re: First Impressions From Rift Owners

Post by Dilip »

DoC OK

Why you think games need special effort or its hard to make Primary VR game with intention to port it on all patform

You can see AVATAR which was NAtive 3D Mulitplatformer with many 3D options yet Equally NICE in 2D, Another Example Batman AA though not native 3D its 3DVISION Highlighter worked Very well on my LG D2342+Tridef Ignition

I think as RIFT SDK seems to have ROOT LEVEL intigration and some asthetics keeping in mind while developing game like

A. Avoiding Object Placemnet at Wrong Depth in 3D Space or Z Pan
B. Avoiding Mixing UP 2D elements with 3D elements
C. Avoiding comman 3D incompatible Shadders
D. Avoding certain shadow rendering methods making them look awful in 3D
E. Testing game in S3D using all populor middlewears in BITA TESTING only in (generic 3d profile) before release (Does Any studio bother to do this other then those Fu***ng 3D VISION CERTIFIEd or HD3D Certified titles, Come on you are game developer developing for mass and this is not too much to ask)

Keeping this simple things while developing any "Tom Dick Herry" game can be S3D supported then why....????

It can always kept as option which can be turned off. it can also be completly desabled on certain incompatible platforms

Where its hard work???
Last edited by Dilip on Sun Apr 07, 2013 12:54 am, edited 1 time in total.
Pyry
Two Eyed Hopeful
Posts: 85
Joined: Mon Aug 13, 2012 5:55 pm

Re: First Impressions From Rift Owners

Post by Pyry »

Doc Ok wrote:Proper antialiased downsampling during lens correction (it's based on a quadratic radial distortion formula) is surprisingly difficult to implement.
There's the proper way, and then there's the brute force sampling way (in pseudo glsl fragment shader; offsets is a list of vec2s of sub-pixel offsets from which to sample and average):

Code: Select all

varying vec2 destPos; // the coordinates of the pixel in the final warped image
vec4 val = vec4(0.0,0.0,0.0,0.0);
for(offset in offsets) {
  vec2 srcPos = warpFunction(destPos + offset);
  val += texture2d(srcImage, srcPos);
}
gl_fragcolor = (val / offsets.length); 
Doc Ok
One Eyed Hopeful
Posts: 23
Joined: Sun Mar 17, 2013 7:41 pm

Re: First Impressions From Rift Owners

Post by Doc Ok »

Pyry wrote:
Doc Ok wrote:Proper antialiased downsampling during lens correction (it's based on a quadratic radial distortion formula) is surprisingly difficult to implement.
There's the proper way, and then there's the brute force sampling way (in pseudo glsl fragment shader; offsets is a list of vec2s of sub-pixel offsets from which to sample and average):

Code: Select all

varying vec2 destPos; // the coordinates of the pixel in the final warped image
vec4 val = vec4(0.0,0.0,0.0,0.0);
for(offset in offsets) {
  vec2 srcPos = warpFunction(destPos + offset);
  val += texture2d(srcImage, srcPos);
}
gl_fragcolor = (val / offsets.length); 
I'm glad you posted that, because it gives me a chance to address it. The code above implements a box filter, which has pretty poor spectral properties. It removes a lot of frequencies that could otherwise be represented, meaning it's blurrier than it should be, and at the same time it lets a lot of high frequencies slip through, causing aliasing. Given the relatively low resolution of the Rift's screen, you'd want to use a better subsampling filter, at least a Hermite filter or maybe a Lanczos. The problem is computing the sampling weights to weigh the individual fragment contributions in your shader. I.e., instead of val += texture2d(srcImage, srcPos) you'd need to do val += texture2d(srcImage, srcPos) * weight(srcPos, offset). The weights are not only offset-dependent, but also position-dependent, because the distortion is non-uniform across the screen.

It can definitely be done, but it's a bit hairy, and will take lots of texture lookups in the shader. I'm guessing that's why the code is currently disabled in the SDK.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

Mel wrote:I just want to reiterate that I disabled all filtering/AA settings on my video card, and the Tuscan demo looks pretty darn good now. The focus 'issue' I was experiencing was directly attributable to those settings. I run the demo at 1920x1080 and let the Rift controller do the scaling. No focus complaints now, so that's good.

As for motion blur: I've pretty much established now that, contrary to my previous (poor) 'observations', the motion blur is the same for both head turning and mouse turning. With head turning, the blur of any particular object is easy to see because you can maintain a steady visual lock on the object as you rotate just by rotating your eyeballs in the opposite direction. But with mouse rotation (essentially, full-body rotation), for whatever reason it is much harder to maintain a steady lock on an object, so the blur is harder to discern (I think Mr. Abrash may have described this phenomenon, just in a way my brain couldn't grok).

[edit to remove nonsense]
Then you DISAGREE with John Carmack (regarding the Rift prototype, and Rift Dev Kit to some degree):
At http://www.mtbs3d.com/phpBB/viewtopic.php?f=140&t=17066&p=116924#p116924, geekmaster wrote:
At http://www.mtbs3d.com/phpbb/viewtopic.php?f=120&t=14967, John Carmack wrote:... Adding explicit motion blur to get rid of the strobing might be a win, but the real answer is a 120hz super AMOLED panel when they are available in the right size and resolution. ...
IMHO Oculus did an admirable job with the Rift Dev Kits at addressing John's concerns about various details of the prototype in the rest of his post quoted above.

The above quote may be WHY the Rift SDK intentionally adds motion blur. According to Carmack, that would be to reduce the 60Hz stroboscopic effect during rapid head motion. I did observe such ghost images, and a 120Hz display would only make them twice as close together. I think an OLED display with no physical motion blur would need even more software motion blur.

I know that watching movies on my HeadPlay HMD (with high speed LCoS screen) has horrible ghost images while moving my eyes to track moving objects on the screen, but MUCH WORSE because of the sequential RGB, causing a trail of RGB ghosts instead of full color ghosts. I suspect that OLED will be similar to LCoS, but at least it will be parallel RGB instead of sequential RGB. For now, I think the Rift LCD is better than OLED when used on older or slower computers, because some of the motion blur is built into the slower pixel switching time.

So, in my opinion, OLED displays will need faster computers to support strong motion blur, to minimize ghost image trails during head motion.
SUMMARY: Removing software motion blur may make "ghost trails" (60Hz motion strobing) of high contrast vertical edges (which I personally observed in my Rift) more obvious during rapid sideways head movements. However, removing (expensive) motion blur may allow the Rift to be used on older/slower computers.
Doc Ok wrote:... Oooh, it just occured to me that AA is applied during world rendering, i.e., before lens correction. AA depends deeply on the precise grid structure of the display, so if the grid is warped afterwards, all kinds of -- very subtle, admittedly -- weird stuff might happen.
Another good reason to turn off software motion blur, until we get the algorithms and hardware processing power to do this (for each Rift lens eyecup, and for each of the eleven screen distance adjustments). Perhaps NOT doing all this expensive filtering until final pre-warp processing is the best solution until we get the correct algorithms and adequate processing power to implement them.

I am just doing a choice of simple mean-sampling (pixel averaging), or less processor hungry monte carlo (random) sampling, when downscaling and pre-warping images for my Raspberry Pi. Those are very low overhead compared to what goes on in the Unity Tuscany demo. Here is a sample screenshot from my code, demonstrating my "fisheye pre-warp":

Image
*** Credit goes to mediavr for the original Belmont Park 360-degree panaramic stereoscopic image pair from which the above Rift-compatible screenshot was extracted.

That 180-degree fisheye image is about 170-degree FoV (fullscreen), but the Rift needs a bit less (perhaps about 150-FoV fullscreen, with only about 110-degrees visible depending on viewing angle through the lenses). To reduce the FoV to what the Rift needs, I just need to clip a little off all sides and expand it to fullscreen again. Then the small visible curvature (in the Rift) of vertical lines at the extreme outer edges will become linear, so the lamp posts near the edges do not look even a little "bent".

Note that the Emerson pre-warp algorithm used in cybereality's Vireo drivers was derived from images that John Carmack said did not have enough pre-warp (for his Rift prototype), as can be seen in John's quoted post above. It appears taht the warp parameters have already been adjusted for best linearity for the Rift Dev Kits, at least in the Oculus Rift SDK. The Unity Tuscany demo looks pretty darned good.

EDIT: Try "free viewing" the SBS stereoscopic image embedded above, by looking "through" the small window on your screen, at the 3D scene beyond that virtual window. Or click it to see the full 1280x800 image, then use Ctrl-mousewheel to shrink it, bringing the image centers closer together so you can free view it as above, but at higher image quality. Remember that the Rift does its own "defisheye", making straight lines linear again (when the correct FoV is uses, about 150-degrees fullscreen, with about 100-degrees visible with stock Rift lenses, as I mentioned above). I realize that I need to compress the outer edges horizontally in my "spherical panorama to fisheye" conversion, but that code is disabled at the moment...
:D

My code mentioned above will be part of my "Rift Rider" app (complete with virtual Roller Coaster rides), and my spherical "skyboxes" will have depth maps, so I can have live virtual objects interact with them (such as passing behind lamp posts and trees, with full occlusion by objects in the "skybox"). I also have working prototype code that projects scrolling text (with full "anti-screendoor" /"dynamic resolution enhancement" software subpixel "wobulation" to make various fonts and text colors readable even when only 2x2 pixels per letter). Beware that "subpixel wobulation" is "dizzy making" when applied to the full screen in the Rift (I tried it), but for text only, it works great.

Sometimes "Roll Your Own" software and algorithms pay off in the end. Where else do new discoveries arise?


P.S. I think Doc Ok should steal this image for his avatar:

Image
:lol:
Last edited by geekmaster on Sun Apr 07, 2013 1:45 pm, edited 1 time in total.
Pyry
Two Eyed Hopeful
Posts: 85
Joined: Mon Aug 13, 2012 5:55 pm

Re: First Impressions From Rift Owners

Post by Pyry »

Doc Ok wrote: I'm glad you posted that, because it gives me a chance to address it. The code above implements a box filter, which has pretty poor spectral properties. It removes a lot of frequencies that could otherwise be represented, meaning it's blurrier than it should be, and at the same time it lets a lot of high frequencies slip through, causing aliasing. Given the relatively low resolution of the Rift's screen, you'd want to use a better subsampling filter, at least a Hermite filter or maybe a Lanczos. The problem is computing the sampling weights to weigh the individual fragment contributions in your shader. I.e., instead of val += texture2d(srcImage, srcPos) you'd need to do val += texture2d(srcImage, srcPos) * weight(srcPos, offset). The weights are not only offset-dependent, but also position-dependent, because the distortion is non-uniform across the screen.

It can definitely be done, but it's a bit hairy, and will take lots of texture lookups in the shader. I'm guessing that's why the code is currently disabled in the SDK.
What specific kernel it approximates depends on what offsets you have in the table; you'll only get a box filter if you picked your offsets from a uniform distribution in a box. If you generate your offsets by drawing them from a normal distribution, you'll get an approximation to a Gaussian kernel. Weights will indeed let you get by with fewer samples, but they aren't strictly necessary. And, actually, since you're sampling in the destination space, you don't need to change the weights based on position (the position dependence comes from passing the samples/offsets through the warp function).

Edit: Also, keep in mind we're talking about sampling for a single pixel-- the warp function is so smooth that you can safely assume it's linear in the sampling neighborhood of the pixel*, in which case, if you care about tweaking the weights at all, effectively they all get multiplied by the same constant factor.

* which suggests that you might be able to slightly optimize the inner loop by passing the offsets through a locally linearized approximation of the warp function
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

geekmaster wrote:Then you DISAGREE with John Carmack (regarding the Rift prototype, and Rift Dev Kit to some degree):
Not sure why you're yelling about my disagreement, but if Mr. Carmack is right about the motion blur being deliberate, than I ABSOLUTELY DISAGREE with him, and the job Oculus did is not 'admirable'. It looks awful, and for head rotations it is completely and utterly unnecessary. Go ahead and leave it in for body rotations, though, because the blur is hardly noticeable at all.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

Mel wrote:... if Mr. Carmack is right about the motion blur being deliberate, than I ABSOLUTELY DISAGREE with him, and the job Oculus did is not 'admirable'. ...
Now who's "yelling"? :mrgreen:

Are you "yelling" at me, or are you "yelling" at John Carmack, or are you "yelling" at the Oculus team member(s) who decided to use motion blur in the Unity Tuscany demo?

Do you consider EVERY use of a capitalized word as "yelling"? I would LOVE to see your reaction if you had to update a typical ALL CAPS COBOL program. I just use all-caps for emphasis... Some people prefer bold, or italics, or other text colors.
:shock:

But yes, I tend to agree with you, and I would prefer to NOT use motion blur. Is there a hot key to turn it off in the Unity Tuscany demo, or does it need a code change?
3dvison
Diamond Eyed Freakazoid!
Posts: 718
Joined: Sun Oct 24, 2010 7:25 pm

Re: First Impressions From Rift Owners

Post by 3dvison »

I hate it when you two fight. Please think about your 2.4 VR children and how it makes them feel...LOL..You know that .4 kid of yours is sensitive

Has turning brightness down as far as you can, helped with ghosting much ?
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

3dvison wrote:I hate it when you two fight. Please think about your 2.4 VR children and how it makes them feel...LOL..You know that .4 kid of yours is sensitive

Has turning brightness down as far as you can, helped with ghosting much ?
Turning the brightness down has the opposite affect...it makes the blur more obvious.

At least to my eyes.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

3dvison wrote:I hate it when you two fight. Please think about your 2.4 VR children and how it makes them feel...LOL..You know that .4 kid of yours is sensitive

Has turning brightness down as far as you can, helped with ghosting much ?
The ghosting is not a "CRT phosphor persistence" phenomenon. It is a perceptual phenomenon, where your brain sees different positions along the track of head motion, leaving a "Persistence-of-Vision" after-image (in your eye/brain) for each time the LCD panel was updated at its refresh rate. Turning down the LCD backlight brightness in a darkened face mask like the Rift uses just changes your eye pupil diameter to compensate to some degree, and you adapt to the lower light level AND you tend to be more sensitive to motion in a darker environment.

So, such brightness adjustments probably do not reduce perceived stroboscopic ghosting. However, lowering the CONTRAST may reduce perceived ghosting, because it is only particulaly visible for high contrast edges perpendicular to the direction of head motion (and even then, mostly visible ONLY when you are actively looking for it).

EDIT: I see that vrvision beat me pressing submit on his post, but with a less complete answer.
lmimmfn
Cross Eyed!
Posts: 138
Joined: Mon Apr 01, 2013 6:44 pm

Re: First Impressions From Rift Owners

Post by lmimmfn »

Mel wrote:
3dvison wrote:I hate it when you two fight. Please think about your 2.4 VR children and how it makes them feel...LOL..You know that .4 kid of yours is sensitive

Has turning brightness down as far as you can, helped with ghosting much ?
Turning the brightness down has the opposite affect...it makes the blur more obvious.

At least to my eyes.
That makes no logical sense whatsoever, unless its the Tuscany demo which has motion blur implemented in the SDK? try the oculus Tuscany demo and see if the brightness affects it.
User avatar
KBK
Terrif-eying the Ladies!
Posts: 910
Joined: Tue Jan 29, 2013 2:05 am

Re: First Impressions From Rift Owners

Post by KBK »

Thankfully, most of the issues will reduce in intensity and sensitivity in image manipulation, as the panel resolution goes up. Therefore, the DEV kit 1 will receive the brunt of any potentials in negativity in this area. I have yet to see a Rift, DIY or otherwise, but the issue was predictable.

edit: Reduction in (peak) brightness is a contrast constriction for the eyes; bringing up susceptibility to issues of image subtlety, at the same time persistence of vision increases. The issue is compounded. Digital systems are so dang...digital. (on or off, exact or inexact) :P
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

lmimmfn wrote:
Mel wrote:Turning the brightness down has the opposite affect...it makes the blur more obvious.

At least to my eyes.
That makes no logical sense whatsoever, unless its the Tuscany demo which has motion blur implemented in the SDK? try the oculus Tuscany demo and see if the brightness affects it.
A reduced brightness may help on a LCD panel in a brightly lit room, but in the Rift, the screen is all you see. Reduced brightness brings your "night vision" into play, with increased sensitivity to motion. Another way to describe it is that when excited more fully by bright lights, the rods and cones in your retinas fire more frequently, but taking longer to fully recover their full sensitivity. In other words, when your vision cells fire less frequently (lower frequency) they can fire more quickly (less latency/motion sensitivity), due to being less "tired".

So, you see MORE motion ghosting when the overall light levels are darker.

Clear enough?
Pingles
Certif-Eyed!
Posts: 555
Joined: Sun Jan 27, 2013 6:15 pm

Re: First Impressions From Rift Owners

Post by Pingles »

Maybe we can bring this thread back home with first impressions from new Rifters and move the LCD screen debate to another thread.
jack612
Two Eyed Hopeful
Posts: 80
Joined: Mon Aug 27, 2012 12:20 pm

Re: First Impressions From Rift Owners

Post by jack612 »

I was hoping we could talk some more about diets.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

Pingles wrote:Maybe we can bring this thread back home with first impressions from new Rifters and move the LCD screen debate to another thread.
jack612 wrote:I was hoping we could talk some more about diets.
Back on track ALREADY? We haven't even discussed Nazis (or icons that look like Nazis) or other "adult topics" in this thread yet!

I cannot find a relevant LCD screen master debaters thread on which to continue this topic... :(
Krenzo
Binocular Vision CONFIRMED!
Posts: 265
Joined: Tue Sep 07, 2010 10:46 pm

Re: First Impressions From Rift Owners

Post by Krenzo »

Doc Ok wrote:I did notice that, when I was turning my head quickly left and right, the horizon tilted quite a bit during those motions
That's very interesting. I have an IMU from Sparkfun that I use attached to my Sony HMZ-T1, and I experience horizon tilting when I pan my head left and right. My guess was that it's an issue with the software, but I haven't looked into it.
User avatar
kenman884
One Eyed Hopeful
Posts: 38
Joined: Wed Mar 13, 2013 1:44 pm

Re: First Impressions From Rift Owners

Post by kenman884 »

jack612 wrote:I was hoping we could talk some more about diets.
:lol:

I was hoping I would see some feedback about the two Skyrim-compatible drivers. That's one of the first things I'm going to do when I get my Rift (and again when I upgrade to the consumer).
raijinspecial
One Eyed Hopeful
Posts: 38
Joined: Sat Sep 01, 2012 6:30 pm

Re: First Impressions From Rift Owners

Post by raijinspecial »

I really cant wait to try out Oblivion. I think the cartoonish environments will look really interesting and high framerates should not be a problem at all.
Machinima
Cross Eyed!
Posts: 166
Joined: Sun Feb 17, 2013 10:34 am

Re: First Impressions From Rift Owners

Post by Machinima »

raijinspecial wrote:I really cant wait to try out Oblivion. I think the cartoonish environments will look really interesting and high framerates should not be a problem at all.
Should be fine running it off an SSD but I remember getting stutter due to data loading from the harddrive when moving quickly (e.g. horseback). Personally I'd really like to see Fallout 3.
zalo
Certif-Eyed!
Posts: 661
Joined: Sun Mar 25, 2012 12:33 pm

Re: First Impressions From Rift Owners

Post by zalo »

Since this is the closest place I could find to post it (which is pretty bad), I noticed something really odd about the C++ Tuscany Demo.

When you first turn it on, the motion blur is really bad. But when you hit F2 (to turn the distortion off), the motion blur becomes barely perceptible. The difference shocked me when I first saw it.

Can someone else with a Rift try the demo and toggle between F2 and F3? I suspect that all that black space around the warped image is causing some sort of ghosting problem.
Mel
Binocular Vision CONFIRMED!
Posts: 287
Joined: Sat Nov 10, 2007 7:45 am

Re: First Impressions From Rift Owners

Post by Mel »

zalo wrote:Since this is the closest place I could find to post it (which is pretty bad), I noticed something really odd about the C++ Tuscany Demo.

When you first turn it on, the motion blur is really bad. But when you hit F2 (to turn the distortion off), the motion blur becomes barely perceptible. The difference shocked me when I first saw it.

Can someone else with a Rift try the demo and toggle between F2 and F3? I suspect that all that black space around the warped image is causing some sort of ghosting problem.
The blurring is visible on my Rift, with or without warping.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: First Impressions From Rift Owners

Post by geekmaster »

F2 does affect pre-warp blur (caused by warping AFTER downsampling) but does not affect motion blur (caused by LCD pixel switching time). There is confusion as to whether or not the Unity Tuscany demo adds software motion blur. Apparently somebody examined the code and determined that it is not turned on.

You can see the effect of pre-warp blur by looking at bricks on the inside wall of the Villa in the Unity Tuscany demo, and move a little. Different bricks will become sharp or blurred. This would not happen if the pre-warp were done BEFORE downsampling.
User avatar
mickman
Binocular Vision CONFIRMED!
Posts: 279
Joined: Mon Jun 04, 2012 5:07 pm
Location: London
Contact:

Re: First Impressions From Rift Owners

Post by mickman »

There 's a good attempt at showing actual screen pixel density of the Rift in this video. This guy simply jams a smartphone into the eye piece... I know its primitive but it does provide a vague representation of what one can expect form the Dev. Kit... & Wooo...... he's already running 2 dev kits !!

http://www.youtube.com/watch?v=Ubepu6X0LfQ


slightly off topic but worth a mention: There's a good thread on the unity3d site regarding how Unity lacks in terms of delivering decent quality 3D gaming for the Rift.. Thats nothing surprising as unity tends to be tailored more towards phone apps... UDK would be the better dev platform for V.R games.

http://forum.unity3d.com/threads/174387 ... s-Rift-but
Image Image

My Mind is screaming like a Zen Master dreaming .....
lifespan
One Eyed Hopeful
Posts: 2
Joined: Mon Apr 08, 2013 5:51 pm

Re: First Impressions From Rift Owners

Post by lifespan »

Any tips for eliminating the eye cups fogging up. For some odd reason I sweat a bit under the eyes and it's making it impossible to use the Rift for more than a minute or two.
User avatar
pizzy00
Cross Eyed!
Posts: 123
Joined: Thu Mar 07, 2013 2:04 pm
Location: Indiana

Re: First Impressions From Rift Owners

Post by pizzy00 »

You could always try cleaning them with spit like you do with under water goggle. I don't know if it will work.
Check out this forum for a beta driver to get existing games working with the Rift.
Official Vireio Perception Driver Forum http://www.mtbs3d.com/phpbb/viewforum.php?f=141
Support these games on Steam Greenlight them http://steamcommunity.com/sharedfiles/f ... d=92979040
Rig Specs - MS Windows Ultimate, i5 3470, 16 GB RAM, 2x AMD 1GB 5850 HD crossfired, SATA2 HD
OzOnE2k10
Sharp Eyed Eagle!
Posts: 436
Joined: Wed Jan 13, 2010 7:12 am
Location: UK

Re: First Impressions From Rift Owners

Post by OzOnE2k10 »

The main thing that Oculus have mentioned is to not tighten the headstraps too much.
I think the middle part of the foam is to allow the mask to breathe, so it can't be too squashed.

The Rift is supposed to be light enough to allow the straps to be quite loose without it falling off.
Not sure how valid that statement is though ofc?

I'm sure there will be fan mods done to the Rift over the coming months too. 8-)
lifespan
One Eyed Hopeful
Posts: 2
Joined: Mon Apr 08, 2013 5:51 pm

Re: First Impressions From Rift Owners

Post by lifespan »

OzOnE2k10 wrote:The main thing that Oculus have mentioned is to not tighten the headstraps too much.
I think the middle part of the foam is to allow the mask to breathe, so it can't be too squashed.
Yup, that was the trick. I didn't think the goggles felt too tight, but I found that when I loosened the up they still stayed snug in place and my fog issues went away.
User avatar
mattyeatsmatts
Binocular Vision CONFIRMED!
Posts: 315
Joined: Wed Mar 06, 2013 5:02 pm

Re: First Impressions From Rift Owners

Post by mattyeatsmatts »

[youtube]http://www.youtube.com/watch?feature=pl ... KxoEZCaKoY[/youtube]
[youtube]http://www.youtube.com/watch?feature=pl ... qqH_L8H-7s[/youtube]

not me I just found it on youtube and thought it was good.
Last edited by mattyeatsmatts on Tue Apr 09, 2013 3:15 am, edited 4 times in total.
User avatar
Callezetter
Two Eyed Hopeful
Posts: 71
Joined: Thu Jul 05, 2012 3:09 pm
Location: Stockholm, Sweden

Re: First Impressions From Rift Owners

Post by Callezetter »

Really great review, Matt. Cheers!
Post Reply

Return to “Oculus VR”