Slashdot Mirror

Even more impressive is the Campanile movie, where an entire 3D model of the UC Berkeley campus and a fly-by shot was generated from just 15 still pictures. This was done a whole decade ago, for SIGGRAPH 97.

What she was mentioning could also be interpreted as a layman's impression of some kind of a high dynamic range setup. Like this camera for "photo-realistic lighting of CG characters" (a still camera). And look at this example of HDR motion blur, which would make a lot of sense in speed racer, as it would let them put more detail and brighter, more vivid color into the blur, which they could use to blur landscapes they are racing through, or CGI cars that are racing by the camera.

They could be combining HDR and multiple zplanes/lenses to capture the entire light field, perhaps capturing the same scene in layers for automated effects similar to bluescreen (think layers in photoshop or the gimp) and bringing parts into focus as they like in post (by calculating the image from synthetic aperture data perhaps). The description does not mention multiple light beams or multiple camera lenses though so it seems HDR and setting up shots laboriously, from z=infinity to closeup, which would explain why the actress had to wait so long just to shoot a short part. The description of brilliant backgrounds and "all the colors that were not in matrix" also points to a hyperrealistic style which makes the backgrounds much brighter and sharper, like in HDR photography.

Try these links-

http://en.wikipedia.org/wiki/High_dynamic_range_im aging

A very good one here - the original authority on the matter-
http://www.debevec.org/Research/HDR/

Some technical research (with good examples and clips)
http://www.anyhere.com/gward/hdrenc/hdr_encodings. html

Does that help? Probably should have included it in my earlier post.

> Since "true" HDR consumera camera's don't exist [...]

Depends on what you mean by HDR.

For some people, HDR simply means "a very high dynamic range" (compared to competing products, or the "normal" standards). That's the case with these monitors.

For othr people, HDR means a dynamic range that is greater than your output medium. By definition, a "HDR" monitor can't comply with this definition, although a monitor can certainly be compatible with "HDR input", if both it and the graphics card support it.

And for other people HDR means virtually unconstrained, floating-point representation of light properties (including negative light, for example). Again, a monitor can't really represent this, but it can accept an input format that supports an approximation of this model. Think of it as "extreme self-calibration".

Most modern digital still cameras capture at least 20% more information than they store in JPEG format. In other words, their raw format is "HDR" (second definition above, not the third one) when compared to their JPEG output. When you load a raw file into an application, you'll typically have the option to respect the original exposure (and basically discard those extra 20%) or lower the exposure so that all the information is compressed into the "visible" range.

Photoshop's support for floating-point HDR is in its infancy (they don't even fully support LDR 16-bit images). If you want to play around with HDR, a better choice is Paul Debevec's HDRShop. In fact, Paul's site has long been one of the most important resources for digital HDR development and research (and here I include the third definition above, which is what "HDR" normally means when applied to CGI).

Unfortunately "HDR" has become a marketing buzzword, which means that in most cases it's used incorrectly, simply to mean "higher precision" or "with light blooming effects"...

RMN
~~~

http://www.debevec.org/Research/HDR/

I'm a little curious to know why no one on this thread has mentioned the guy who first started implementing it the way it is used now in the first place. We might as well just forward a link to his site and Siggraph papers.

Paul Debevec's homepage

HDRI has been around for a long time --since the late '90's. I don't understand why this is considered new, especially since Paul Debevec introduced this at Siggraph in '99 (?) in Fiat Lux. It's been in almost all the latest big VFX movies to date. HDRI is not a "a lighting process that's been designed to emulate in-game or artificially generated lighting". It is a method of lighting scenes using real-world lighting scenarios. I suppose this is new to the video game industry, but this has been around for quite awhile, so the article is a bit misleading. For more info about HDRI, go to Paul Debevec's site: http://debevec.org/

I think, if you want to be precise, what Valve did in Lost Coast should be called Paul Debevec's High Dynamic Range.

http://debevec.org/ lots of info here

This is not targeted at the home computer user. This is a technology for high-end video gear, and a few years down the road, for high-end home theaters (assuming anything gets released to the public in a yet-to-be-determined HDR video format).

Above the contrast ratio of film or DLP projection, I'll agree it's close to useless, however. Unless some exotic sci/med visualization stuff needs it.

"For the average enthusiast, there are only three things to remember about HDR:

1. Bright things can be really bright
2. Dark things can be really dark
3. And the details can be seen in both"

Yes... and it's good if you like to tweak the image. * I don't get how this will help with 3d-rendering.

Current display devices can display a limited range of contrast and colors, which is one of the main reasons that most image acquisition, processing, and display techniques use no more than eight bits per color channel. This course outlines recent advances in high-dynamic-range imaging, from capture to display, that remove this restriction, thereby enabling images to represent the color gamut and dynamic range of the original scene rather than the limited subspace imposed by current monitor technology. This hands-on course teaches how high-dynamic-range images can be captured, the file formats available to store them, and the algorithms required to prepare them for display on low-dynamic-range display devices. The trade-offs at each stage, from capture to display, are assessed, allowing attendees to make informed choices about data-capture techniques, file formats, and tone-reproduction operators. The course also covers recent advances in image-based lighting, in which HDR images can be used to illuminate CG objects and realistically integrate them into real-world scenes. Through practical examples taken from photography and the film industry, it shows the vast improvements in image fidelity afforded by high-dynamic-range imaging. [more]

Paul Debevec has probably already done something like that. He's done a lot of work with "image based rendering", including reconstructing scenes from photos and extracting "light probes" from photos of shiny spheres. He's got a lot of papers and demos on his home page there.
Sorry bud. I know how you feel! I've had similar experiences myself.

the gamuts of available monitors do not encompass the gamuts of all possible printing techniques. (Please make us better monitor scientist nerds) Your wish is their command: http://www.debevec.org/HDRI2004/

Actually, yes. The poor man's way (for stills at least) is to take a number of shots with the camera a few stops apart to capture an extremely wide exposure latitude, and recombine them in HDRshop.

There are also a number of companies making HDRI backs for digital still cameras that automate some portion of this.

For motion, a beam splitter and a few calibrated and synced cameras could be rigged up in a similar way, although I've not heard of anyone actually doing this.

The idea isn't to actually produce end results beyond the exposure latitude of film, but rather to be able to color-correct more agressively without banding, clipping, or losing detail.

Besides, an HDR display is targeted mainly at compositors who actually want to see a film-like contrast range on screen so they know better what the end result will look like; much the way that film sound editors do the final mix in a theater-like room to get the best idea of how movie house acoustics will sound with their mix.

Whilst it might not take cool pictures, there's certainly some cool pictures of it - or, rather, the important elements of it.

Very, very spherical
That is about as close to a typical POVray render reality has ever gotten. Those spheres are the most spherical objects created by mankind.

They would make excellent HDR Probe reflectors if they weren't so expensive.
( In fact, you can run the manu photos through HDRshop and derive panoramic images of whatever environment they were in. Cute. )

More on fabrication and measurement.
From reading that, it appears they could've made them even more precise if they really, really wanted to (i.e. would be able to put up with tossing out even more of the spheres than they already did).

I think you're confusing two things here...

OLEDs are a replacement for, say, LCD, CRT, Plasma and LED displays.

HDR, on the other hand, is not a replacement for any of those - it is a concept (that of a very high range, with implied (though not specified) fidelity equal to or greater than currently used) that can be used -with- all of the above.

If you can get a CRT to display at stupendously bright levels, it could qualify for a HDR display - assuming it would still keep the same or better fidelity than a 'regular' CRT next to it in the same black-to-white range.

The actual technology used by this HDR display is that of two LCDs.. One for regular display, and another low-resolution one backlit by bright white LEDs.
And OLEDs could be a replacement for that setup, yes - but not for the concept of HDR itself ;)

For more on HDR :
http://www.debevec.org/
http://http://www.trinisica.com/sub_learn_typediss ue.asp?lv=3&mode=1&issue=002

And HDR and fidelity :
http://www.pointzero.nl/sf/reffect_scale/ ( see the chart in section 2 )
http://p067.ezboard.com/fhdrshopfrm1.showMessage?t opicID=232.topic ( beware of pop-up )

This technique (tie camera to kite, tie kite to hand) was used to do some interesting image based modeling work by Paul Debevec, he flew his kite around the UC Berkley campus then later used the images to reconstruct 3D models of various buildings. Check it out: Paul Debevec's Campanile Movie.

This is the first graphics programming article I've seen in a long time with no visual aids. I think the writer simply wanted to write a huge "smart" article so that he'd seem impressive. Missed some good algorithms for terrain rendering (tilemap, octrees, frustrum culling). If you want a really good site about graphics algorithms, check out Paul Debevec's homepage (famous for his contributions to The Matrix)

It's about time high dynamic-resolution imaging was supported in hardware. It's not only simple to use, but makes many many things easier--no more worrying about hitting the 8-bit limit, picking colors to avoid saturation/overlighting, or struggling with realistic compositing. For compositing alone it's priceless. Debevec even has a light stage for compositing real-life objects into computer-generated scenes (instead of vice versa).

I've been working for a while now on HDRI solutions for things like POV-Ray--even if it is a pseudo-solution. (Or you can get the ml-POV patch for more native HDRI support.)

Well it's more generically called image based techniques. In these case image based modeling and in particular image based lighting or IBL. Matrix was one of the first practical examples though the technique was already known and demonstrated at SIGGRAPH in particular by Paul Bebevec's team (some of the research was seen as far back as SIGGRAPH 96). The Campanille movie was a good example. Borushkov was part of that team.

The Campanille movie

A few films that pioneered some of that stuff too included Batman and Robin, Fight Club and What Dreams May Come.

I suppose by this you should also clarify that the first Matrix used this too, for the backgrounds for those bullet time scenes. The Matrix Reloaded is more like those same techniques refined (clearly, the backgrounds are way more complex now).

There is REALVIZ ImageModeler.

The technology of deriving 3d geometry information from 2d images is typically known as image-based rendering. Gabe Newell (head of Valve Software) said that this is one of the technologies they didn't have enough time to get into Half-Life 2. It was first demonstrated in The Campanile Movie and has since been used for special effects purposes in movies (including the Matrix). It is very impressive, not only for its ability to generate the geometry from the images, but also its way of generating perfect distortionless texture mapping for said geometry.

seems to be /.'ed or something.
Meanwhile, here's another site that actually has the Campanile movie:
http://www.debevec.org/Campanile/

j

This is a little cross-pollinated. HDRI isn't used in photography, traditional or digital. You're referring to image formats with higher bit-depth.

High Dynamic Range is used in 3d rendering to more accurately simulate lighting (especially skylighting). You take multiple exposures of a scene (at 8-bit) and composite those together to form a HDR image, which is used by high-end rendering software to project onto a 3d scene to create very soft, realistic lighting.

For anyone that's interested:
HDRI Primer

D
www.zerosexlife.com

The real problem there is dynamic range. Photoshop still works in 8 bits per channel, which is clearly not enough for any sort of exposure / brightness / contrast control. You need at least 16 bits per channel, preferably 32 (in floating-point format). Photoshop can load 16-bpc images but 99% of its tools are disabled until you convert the image down to 8-bpc. In other words: the 16-bpc mode is there just for marketing.

There are some interesting HDR (high dynamic range) projects, such as HDRShop, and these formats are also used in several high-end 3D renderers, but I don't think they will become mainstream until Photoshop adopts them.

Unfortunately, Adobe insists on minor updates instead of doing what Photoshop (and Premiere, and several other of their products) needs, which is a complete rewrite.

High-end 3D renderers also have very good "film grain" simulation (film grain is not just random noise, it has very specific characteristics), and other tricks that can make CGI "feel" almost exactly like traditional analog media. But again, this is not something you'll find in Photoshop.

RMN
~~~

Check out Paul Debevec's web site. He seems to have pioneered (correct me if I'm wrong) a lot of image-based rendering techniques. HDR images are an important part of this. He describes how to recover HDR images from photographs, how to create "light probes" (HDR environment maps), and then how to light synthetic scenes with a light probe.

Check out Paul Debevec's web site. He seems to have pioneered (correct me if I'm wrong) a lot of image-based rendering techniques. HDR images are an important part of this. He describes how to recover HDR images from photographs, how to create "light probes" (HDR environment maps), and then how to light synthetic scenes with a light probe.

Check out Paul Debevec's web site. He seems to have pioneered (correct me if I'm wrong) a lot of image-based rendering techniques. HDR images are an important part of this. He describes how to recover HDR images from photographs, how to create "light probes" (HDR environment maps), and then how to light synthetic scenes with a light probe.

Check out Paul Debevec's web site. He seems to have pioneered (correct me if I'm wrong) a lot of image-based rendering techniques. HDR images are an important part of this. He describes how to recover HDR images from photographs, how to create "light probes" (HDR environment maps), and then how to light synthetic scenes with a light probe.

A film scanner like the ones used at VFX houses can produce material with up to 14 bits per channel of color resolution. So can Panoscan's MK1 HDR camera. For reasons outlined in another thread, there are advantages to using FP numbers rather than integers to represent these values.

The CCDs used in these devices are pretty expensive and aren't available in pro-sumer or consumer devices. For now.

Apps like Idruna's Photogenics, Paul Debevec's HDRShop, and Greg Ward's Photophile can produce HDR FP images from scans of photos of the same scene using different exposures. This works with the cheap color scanner that you bought at Fry's or Best Buy.

As for synthetic images, Renderman, Mental Ray use 32-bit FP internally. They can already produce 32-bit TIFF images. We're working on making the OpenEXR display drivers for these apps available with the rest of the OpenEXR software distribution.

If you enjoyed watching all those millions of dominoes toppling, you ought to check out Fiat Lux If you enjoy anything to do with CG animation, you should also check this out. It's a rather impressive CG short of thousands of dominoes about the size of the monolith from 2001 (The small one that appeared before the apes - not the giant one in space) toppling through the interior of St. Peter's Basilica. As an added bonus, if you like shiny objects, the domino-monoliths are highly reflective. It's quite a sight to behold. My crude explanation doesn't do Fiat Lux justice. Go check it out for yourself!

You're half-right.

On the one hand, focus is the issue here. Focus is very important, and i myself HATE when the CGI looks like CGI because it's too in focus.

BUT, as someone who's done a few high-quality computer animated shorts, there's a large amount of work that goes into, and a significant number of applications to help with, post-processing effects, such as blur and focus and color temperature. It's fundamentally a very hard problem. For color and lighting, Paul Debevec has been doing a lot of new and interesting things to use real-world lighting to light computer models, and computer lighting to light real people. As for blur and focus, i'm not sure of any really good algorithms or techniques, but in practice, it's largely all hand-tweaked, which is bound to be imperfect.

It's easy to say that "damn CGI artists don't know what the word 'focus' means," but when you take a good look at the problem, it's hardly that trivial.

ben.c

Reminiscent of Paul Debevec's light probes.

-m

Cameras have been sensing depth with infra-red for years. It's called auto-focus. All this camera has added is recording that data along with the rgb. There is a nifty little method of scanning film location for later 3D reconstruction using Lidar that has already been used in films for some time. Lidar is often also implemented for geographical surveys by folks such as NASA. The resolution of the ZCAM sounds very limited and will not be useful for truly realistic keying. Possibly guy-in-front-of-weathermap applications, but it doesn't look like it can handle fine hair, water, or other transparencies that most keying software can today. If you don't want to deal with a green/blue screen, there's also the option of a "difference key" that compares the shot without anyone in it to the shot when your person walks in. Some systems can now do real time insertions this way. The ZCAM still won't give convincing film quality compositing results. With the advances in photogrammetry and it's applications in movies, I don't think that ZCAM has much use potential

Cameras have been sensing depth with infra-red for years. It's called auto-focus. All this camera has added is recording that data along with the rgb. There is a nifty little method of scanning film location for later 3D reconstruction using Lidar that has already been used in films for some time. Lidar is often also implemented for geographical surveys by folks such as NASA. The resolution of the ZCAM sounds very limited and will not be useful for truly realistic keying. Possibly guy-in-front-of-weathermap applications, but it doesn't look like it can handle fine hair, water, or other transparencies that most keying software can today. If you don't want to deal with a green/blue screen, there's also the option of a "difference key" that compares the shot without anyone in it to the shot when your person walks in. Some systems can now do real time insertions this way. The ZCAM still won't give convincing film quality compositing results. With the advances in photogrammetry and it's applications in movies, I don't think that ZCAM has much use potential

This goes by the name of "image-based modeling and rendering," and one of the the pioneers of the field developed the technique that was used for the notorious bullet-time shot. Another group demonstrated their realtime IBMR-from-video process at Siggraph 2000.