raisch@Control.COM (Robert Raisch) (01/16/91)
I have recently seen a piece of marketing hype that uses the "famous" pool table image that was supposed to have passed the "Turing Test" for computer generated images. (It's the one where a cueball is seen striking a set of pool balls, (1,5,11,12), and what makes it really spectacular is the use of motion blurring.) My question is, who owns the copyright to this image, and how do I get in touch with them? It seems to me that an artist should be remunerated for his work. -Rob Raisch, Control Technology Corp.
mcastle@mcs213f.cs.umr.edu (Mike Castle (Nexus)) (01/16/91)
In article <1084@cthulhuControl.COM> raisch@Control.COM (Robert Raisch) writes: >I have recently seen a piece of marketing hype that uses the "famous" >pool table image that was supposed to have passed the "Turing Test" >for computer generated images. (It's the one where a cueball is seen >striking a set of pool balls, (1,5,11,12), and what makes it really >spectacular is the use of motion blurring.) The only image I could find was one using the balls 1,9,8, and 4 (of course the name of the image is 1984 :-). It's in "3D Computer Graphics" by Andrew Glassner (image C-14) and "An Introduction to Ray Tracing" edited by Glassner (p193). It is credited to Tom Porter, Pixar. Is this the one you were thinking of, or is there another one also?? -- Mike Castle (Nexus) S087891@UMRVMA.UMR.EDU (preferred) | ERROR: Invalid mcastle@mcs213k.cs.umr.edu (unix mail-YEACH!)| command 'HELP' Life is like a clock: You can work constantly, and be right | try 'HELP' all the time, or not work at all, and be right twice a day. |
billd@fps.com (Bill Davidson) (01/18/91)
In article <1971@umriscc.isc.umr.edu> mcastle@mcs213f.cs.umr.edu (Mike Castle (Nexus)) writes: >The only image I could find was one using the balls 1,9,8, and 4 (of course >the name of the image is 1984 :-). It's in "3D Computer Graphics" by Andrew >Glassner (image C-14) and "An Introduction to Ray Tracing" edited by Glassner >(p193). It is credited to Tom Porter, Pixar. The first place I saw this image was at the art show at SIGGRAPH '85. It also appears in the January 1986 issue of ACM Transactions on Graphics. It is one of the most photorealistic computer generated images I've ever seen. The motion blur is not the only thing that is impressive. The reflections are well done also. There is a neon "BUD" sign and it is even motion blurred on the eight ball. There is a window in the pool hall which is reflected in the balls and shows a person and palm trees outside. There are pool table lights. I have a photography background so I did notice one "flaw". There is no reflection of a camera lens in the balls. A real photograph would have one (actually, a lot of photographers would love to find a way to produce images of shiny spheres without camera lens reflections so this is probably a good thing ;-). Another thing is that I've never seen any pool balls so pristine as these other than right after they come out of the box new. The same goes for the felt. He should have shmegged them up a bit ;-). --Bill Davidson
jef@well.sf.ca.us (Jef Poskanzer) (01/18/91)
If this is the image I'm thinking of, there are visible jaggies in some
of the reflections on the balls.
---
Jef
Jef Poskanzer jef@well.sf.ca.us {apple, ucbvax, hplabs}!well!jef
"I don't feel right... tonight!" -- Dungknoll@well.sf.ca.us (John Knoll) (01/18/91)
I think the image you are thinking of is "1984" by PIXAR. The balls are numbered 1,9,8, and 4, left to right. A very realistic image.
uselton@nas.nasa.gov (Samuel P. Uselton) (01/19/91)
In article <14869@celit.fps.com> billd@fps.com (Bill Davidson) writes: >In article <1971@umriscc.isc.umr.edu> mcastle@mcs213f.cs.umr.edu (Mike Castle (Nexus)) writes: >>The only image I could find was one using the balls 1,9,8, and 4 (of course >>the name of the image is 1984 :-). It's in "3D Computer Graphics" by Andrew >>Glassner (image C-14) and "An Introduction to Ray Tracing" edited by Glassner >>(p193). It is credited to Tom Porter, Pixar. > >The first place I saw this image was at the art show at SIGGRAPH '85. >It also appears in the January 1986 issue of ACM Transactions on >Graphics. It is one of the most photorealistic computer generated It was also the cover photo for a "popular" computing or technology magazine in 1984 (I think). I think the magazine is no longer with us, but I have the copy in a box at home. There was a real hand with cue matted in on the cover. I've moved twice since then so the box is probably deeply buried. >images I've ever seen. The motion blur is not the only thing that is >impressive. The reflections are well done also. There is a neon "BUD" >sign and it is even motion blurred on the eight ball. There is a >window in the pool hall which is reflected in the balls and shows a >person and palm trees outside. There are pool table lights. > >I have a photography background so I did notice one "flaw". There is >no reflection of a camera lens in the balls. A real photograph would >have one (actually, a lot of photographers would love to find a way to >produce images of shiny spheres without camera lens reflections so this >is probably a good thing ;-). Another thing is that I've never seen >any pool balls so pristine as these other than right after they come >out of the box new. The same goes for the felt. He should have >shmegged them up a bit ;-). ^^^^^^^^^^^^^^^^^^^^^^^ You are not the first to suggest this. In fact, Tom Porter did another image that is certainly better in this respect, and I think better in overall realism. It is the bowling scene "Textbook Strike". It is used on the cover of _The_Renderman_Companion_, but you can see a lot more detail in a blow up. I have a poster of it on my wall. > >--Bill Davidson Sam Uselton uselton@nas.nasa.gov employed by CSC working for NASA speaking for myself
robert@texas.asd.sgi.com (Robert Skinner) (01/19/91)
In article <22705@well.sf.ca.us>, jef@well.sf.ca.us (Jef Poskanzer) writes: |> If this is the image I'm thinking of, there are visible jaggies in some |> of the reflections on the balls. |> --- really? where? The only imperfection I could find was some strobing of one of the motion-blurred reflections. But you have to look REAL close to see it. Of course, as others have pointed out, the balls and table are too perfect, which is another imperfection. Robert Skinner robert@sgi.com During an operatic recital at the White House, while the nervous soprano was doing her best to please the First Family, one of the guests turned to President Coolidge and asked, "What do you think of the singer's execution?" "I'm all for it," Coolidge replied.
aipdc@castle.ed.ac.uk (Paul Crowley) (01/19/91)
robert@sgi.com sez in <1991Jan19.031630.3683@odin.corp.sgi.com> >Of course, as others have pointed out, the balls and table are too >perfect, which is another imperfection. I saw another Turing-test which was a photo of a classroom and a raytraced copy. The only way you could tell the difference was because the board was dirty in one. Has anyone done any work on making things look dirty and worn? Everything in the computer-graphic world is shiny and new. \/ o\ Paul Crowley aipdc@uk.ac.ed.castle /\__/ Trust me, I know what I'm doing.
slamont@network.ucsd.edu (Steve Lamont) (01/21/91)
In article <7926@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul Crowley) writes: >Has anyone done any work on making things look dirty and worn? >Everything in the computer-graphic world is shiny and new. James Kajiya of UCLA has made similar comments. From what I have heard, he is working in that area. Anyone associated with Dr Kajiya care to comment? spl (the p stands for polishing up my pixels) -- Steve Lamont, SciViGuy -- (408) 646-2572 -- a guest at network.ucsd.edu -- NPS Confuser Center / Code 51 / Naval Postgraduate School / Monterey, CA 93943 "It's not what you know, it's who you know to go ask..." - Richard W. Hamming
musgrave-forest@cs.yale.edu (F. Ken Musgrave) (01/21/91)
In article <7926@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul Crowley) writes: > >Has anyone done any work on making things look dirty and worn? >Everything in the computer-graphic world is shiny and new. In my images, I work my tail off to make things not appear too "computer- perfect". This is relatively easy to do with fractal models of nature. Unfortunately, not all complexity in nature is fractal. Fractals generate mondo complexity form a simple description; the kind of complexity I observe in old chalkboards, worn poolballs, asphault, etc., is just not (very) fractal. Thus it may require a lot more work to imitate convincingly. One can resort to the good old manual paintbox solution, but a procedural method would be really neat to have, eh? "We're working on it." Ken -- The Fundamental Dilemma of Existentialism: Eschew obfuscation. Ignore alien orders.
sbeagle@kennels.actrix.gen.nz (Sleeping Beagle) (01/25/91)
> In article <7926@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul Crowley) writes > >Has anyone done any work on making things look dirty and worn? > >Everything in the computer-graphic world is shiny and new. Of course, people with non high colour computers have an advantage. If I am tracing at a low res, I just turn dithering on and it seems to add texture to objects... 1/2 :-) Still hoping to swap QRT files, thanks to those who have answered already. -- ** Official Signature for Sleeping Beagle (aka Thomas Farmer)! ** sbeagle@kennels.actrix.gen.nz || Disclaimers are for sick societies ** Thomas.Farmer@bbs.actrix.gen.nz || with too many lawyers.
stam@dgp.toronto.edu (Jos Stam) (01/26/91)
F. Ken Musgrave writes: >Paul Crowley writes: >> >>Has anyone done any work on making things look dirty and worn? >>Everything in the computer-graphic world is shiny and new. > > In my images, I work my tail off to make things not appear too "computer- >perfect". This is relatively easy to do with fractal models of nature. > > Unfortunately, not all complexity in nature is fractal. Fractals generate >mondo complexity form a simple description; the kind of complexity I observe >in old chalkboards, worn poolballs, asphault, etc., is just not (very) fractal. >Thus it may require a lot more work to imitate convincingly. > Why are fractals so important? Any stochastic model will do. One way to add "dirt" is by texture mapping or solid texturing. (see for example the cover of the Renderman manual). I wouldn't mind seeing more dirt in computer generated images, just for the sake of variety. By analogy, figurative art is not only restricted to hyperrealism. Jos
eugene@nas.nasa.gov (Eugene N. Miya) (01/26/91)
Re: Discriminating realistic appearing synthetic imagery. A test image generator must consider a lot of factors when trying to fool a viewer. You can spend a lot of time, but the benefits of realism do not come simply for the cost. For this reason we need what I call "computational test pilots" [critics of a sort] who are more knowledgeable than the general public and can look for image artifacts. In article <1991Jan25.142659.3914@jarvis.csri.toronto.edu> stam@dgp.toronto.edu (Jos Stam) writes: >I wouldn't mind seeing more dirt in computer generated images, just for >the sake of variety. By analogy, figurative art is not only restricted to >hyperrealism. 1] first attmept 1) take computer generated image, go outside. 2) grab a hand full of soil, sprinkle lightly. or 2] second attempt 1) view local 'red' light district, 2) generate NC-17(tm) images. ;^) --e. nobuo miya, NASA Ames Research Center, eugene@orville.nas.nasa.gov {uunet,mailrus,other gateways}!ames!eugene AMERICA: CHANGE IT OR LOSE IT.
billd@fps.com (Bill Davidson) (01/26/91)
In article <1991Jan25.142659.3914@jarvis.csri.toronto.edu> stam@dgp.toronto.edu (Jos Stam) writes: >I wouldn't mind seeing more dirt in computer generated images, just for >the sake of variety. By analogy, figurative art is not only restricted to >hyperrealism. The pool table picture was an attempt at photo-realism; something we are getting closer to but can't quite do. It was an extremely good example (especially for 1984) though it attacked something that was relatively easy (spheres and green felt). I think the real emphasis was coming up with a good model for motion blur (it made my jaw drop when I first saw it). It almost looks like a real photograph with strobes. There are two major clues: the lack of imperfection and the lack of camera lens reflection in the balls. In many professionaly produced pictures of shiny spherical objects, the camera lens is retouched out or obscured by other reflections around it (you can't shoot it without the reflection; it's impossible). If you were really trying to fool the viewer (given that that is a goal) you might even put in a camera lens reflection. More generally, lighting models are getting really good; to the point that objects look as if they could be real, even though in all cases you know they aren't (I have yet to see a computer generated image that I couldn't tell was not a photograph of a real scene though some took a few seconds of looking). Of course we could debate how important photorealism really is. While in general, I don't need it or really care, it would be nice to know that you could do it if you wanted to. As far as being for variety, it seems like an awful lot of work for mere variety. You'll get more pleasing variety from creative artists using computer graphics than from a small technical improvment (gained at great expense). This is not to say that I am against increased photorealism but "variety" just seems like a dumb reason. Increasing the state of the art and available tools for artists and scientists is a better reason. --Bill Davidson
musgrave-forest@cs.yale.edu (F. Ken Musgrave) (01/26/91)
In article <1991Jan25.142659.3914@jarvis.csri.toronto.edu> stam@dgp.toronto.edu (Jos Stam) writes: >Why are fractals so important? Any stochastic model will do. One way to add >"dirt" is by texture mapping or solid texturing. (see for example the cover >of the Renderman manual). Yes, the reference to fractals in my posting was a bit indicative of my peculiar world view. Fractals are the most compact way of describing complexity in nature. "Dirt", in order to be realistic, should be complex. It can also be quite heterogeneous and thereby non-fractal, as it is the quality of self-similarity that characterizes fractals. By the same token, no simple texture can represent dirt well. A hand-painted texture map is trivial; you might as well scan photographs of dirt and scuffing. If it's not described with a formula or an algorithm it's not computer graphics, it's art or craft. (Again, I'm showing my prejudice in favor of procedural models.) Have a good close look at a "trompe l'oeil" still life painting (e.g., Ken Davies) to get an example of the kind of complexity involved in artful depiction of dirt and wear. We have a long way to come, in computer graphics. Yet the inclusion of such imperfections is essential to the computer graphics "Turing test", and to the aesthetic flexibility of the computer as a medium for visual art. Soap Boxken -- The Fundamental Dilemma of Existentialism: Eschew obfuscation. Ignore alien orders.
mcdonald@aries.scs.uiuc.edu (Doug McDonald) (01/26/91)
In article <15115@celit.fps.com> billd@fps.com (Bill Davidson) writes: >In article <1991Jan25.142659.3914@jarvis.csri.toronto.edu> stam@dgp.toronto.edu (Jos Stam) writes: >>I wouldn't mind seeing more dirt in computer generated images, just for >>the sake of variety. By analogy, figurative art is not only restricted to >>hyperrealism. > >The pool table picture was an attempt at photo-realism; something we >are getting closer to but can't quite do. It was an extremely good >example (especially for 1984) though it attacked something that was >relatively easy (spheres and green felt). > In many professionaly >produced pictures of shiny spherical objects, the camera lens is >retouched out or obscured by other reflections around it (you can't >shoot it without the reflection; it's impossible). But you can come arbitrarily close: take the picture through a neutral density filter whose front surface reflects onto a large surround of black felt. And put the camera in the dark. A 10 percent transmission filter works well. Doug McDonald
billd@fps.com (Bill Davidson) (01/27/91)
In article <1991Jan26.144101.27840@ux1.cso.uiuc.edu> mcdonald@aries.scs.uiuc.edu (Doug McDonald) writes: >In article <15115@celit.fps.com> billd@fps.com (Bill Davidson) writes: >> In many professionaly >>produced pictures of shiny spherical objects, the camera lens is >>retouched out or obscured by other reflections around it (you can't >>shoot it without the reflection; it's impossible). > >But you can come arbitrarily close: take the picture through a >neutral density filter whose front surface reflects onto a large surround >of black felt. And put the camera in the dark. A 10 percent transmission >filter works well. While it's true that this will make the camera reflection pretty much disappear it can cause other problems. Can you say flare? I knew you could. Also, you have to make your apparatus cover the field of view of the spherical object (180 degrees in all directions) without having the fact that it is an apparatus visible. In most cases that's very problematic. Large gel's tend to exhibit all sorts of imperfections which reflect in those shiny objects. These imperfections also translate onto your film directly (since you are shooting through them. Plus there's more places for dust to collect and it all adds up to make that picture you thought was going to crisp and vibrant into a fuzzy mess. Just retouching the camera out in the first place makes for a better image. I've spent a lot of time in the studio trying to learn how to do advertising work. Crisp, vibrant photos are a main-stay in this type of work. I used to shoot 4x5 most of the time and spend a hell of a lot of time trying to make the crispest photos possible. I know what I'm talking about because I've tried it. This is getting away from CG; follow-ups to rec.photo (which I don't read anymore). --Bill Davidson
jaapv@accucx.cc.ruu.nl (Jaap Verhage) (01/27/91)
In article <15115@celit.fps.com> billd@fps.com (Bill Davidson) writes: >In article <1991Jan25.142659.3914@jarvis.csri.toronto.edu> stam@dgp.toronto.edu (Jos Stam) writes: >>I wouldn't mind seeing more dirt in computer generated images, just for >>the sake of variety. By analogy, figurative art is not only restricted to >>hyperrealism. > >The pool table picture was an attempt at photo-realism; something we >are getting closer to but can't quite do. It was an extremely good >example (especially for 1984) though it attacked something that was [...] And more discussion of same. By this time, I'm getting highly interested in this photograph/image. I've read in earlier posts that it's in a book; is there an electronic version of it (the original?) available somewhere? And ftp-address would be highly appreciated. -- Regards, Jaap. Jaap Verhage, Academic Computer Centre, State University at Utrecht, Holland. jaapv@cc.ruu.nl +<-*|*->+ I claim *every*thing and speak for myself
pfl@eniac.tmc.edu (Pedro Faria Lopes) (01/29/91)
In article <7926@castle.ed.ac.uk>, aipdc@castle.ed.ac.uk (Paul Crowley) writes: > robert@sgi.com sez in <1991Jan19.031630.3683@odin.corp.sgi.com> > >Of course, as others have pointed out, the balls and table are too > >perfect, which is another imperfection. > ...some text deleted here... > Has anyone done any work on making things look dirty and worn? > Everything in the computer-graphic world is shiny and new. not everything! have you seen "Paris 1789"? it's a computer generated film done in 1989 to celebrate the bicentenaire of the french revolution. Most of the scenes that include 3D buildings and gardens have that warn and non-perfect look. They avoided the use of solid textures and, instead, they used projected maps. the images to map were created with the help of a paint system and done manually. that's how they avoided the clean shiny new look... > > \/ o\ Paul Crowley aipdc@uk.ac.ed.castle > /\__/ Trust me, I know what I'm doing. pedro faria lopes -- Carl, Computer Animation Research Lab. @ INESC, Phone: 351 1 545150 pfl@eniac.inesc.pt pfl@eniac.uucp Fax: 351 1 525843 eniac's statement: "I'm young at INESC and still learning how to send email"
billd@fps.com (Bill Davidson) (01/29/91)
In article <1125@accucx.cc.ruu.nl> jaapv@accucx.UUCP (Jaap Verhage) writes: >In article <15115@celit.fps.com> billd@fps.com (Bill Davidson) writes: >>The pool table picture was an attempt at photo-realism; something we >>are getting closer to but can't quite do. It was an extremely good >>example (especially for 1984) though it attacked something that was >[...] >And more discussion of same. By this time, I'm getting highly >interested in this photograph/image. I've read in earlier posts >that it's in a book; is there an electronic version of it (the >original?) available somewhere? And ftp-address would be highly >appreciated. I personally would love to have a version of it. It's copyrighted by Pixar. Perhaps they can be persuaded to make it available (I know they have people who read this group). Some work will probably have to be done to make it ftp-able. The version I first saw was (I think) an 8x10 glossy photo-print. You could get as close to it as you wanted and you couldn't see jaggies (I didn't have a lupe handy ;-). I would guess that the original is probably 4k to 8k pixels across with at least 24-bit's of color resolution. That's a hell of a lot of data (maybe as much as 150Mb); considerably more than most sites would consider reasonable for ftp. My vote (if I had one) would be for something around 1280x1024x24 tiff or Sun Raster or some other widely known format (this is still about 3.75Mb uncompressed; compression would be necessary for obvious reasons). I would think it would compress well with all that green felt. --Bill Davidson
musgrave-forest@cs.yale.edu (F. Ken Musgrave) (01/30/91)
This is a posting on behalf of Holly Rushmeier, who is having trouble posting to the net. >All I wanted to add to the discussion about imperfections in >graphics images was that I thought the following paper >"Imperfection for Realistic Image Synthesis" by Becket and Badler >The Journal of Visualization & Computer Animation Vol. 1, No. 1, pp. 26-32. >was a pretty good discussion of the problem and a few >attempts to deal with it. Its a paper you might >me interested in if you haven't already run across it. >-- Holly This is a new European journal published by Wiley & Sons, Sussex, England. >I'd like to see the discussion move forward in time >from talking about an image that was after all published 7 yrs ago. >Also, in case no one has mentioned it, the 1984 picture and the >side by side conference rooms Greg Ward did appear in Foley, >van Dam , Feiner and Hughes. >And finally, as far as Turing tests go, there is only one paper >that I know of where anybody did direct side by side tests of >real and synthetic images (our obscure TOG article from '86). If >people on the net know of similar experiments since then, unpublished >or published I would be interested in hearing about them. (Oops, >I forgot, there is one other paper -- the automotive appearance >paper by the Toyota group in the last Siggraph). >-- Holly Hear hear! - that kind of side-by-side comparison, particularly in the kind of carefully controlled environment of the radiosity experiments, approaches the real practice of Science. Ken (for Holly) -- The Fundamental Dilemma of Existentialism: Eschew obfuscation. Ignore alien orders.
musgrave-forest@cs.yale.edu (F. Ken Musgrave) (01/31/91)
Again, I'm posting for Holly Rushmeier:
Greg Ward did the conference room comparison, and there is an article
about the software he wrote and more images (mainly in black and
white) in the June 1990 issue of "Lighting Design and Application".
There have been a few other places in which real and
synthesized images have been compared:
"An Experimental Evaluation of Computer Graphics Imagery" by
Meyer et al. , Jan '86 ACM Transactions on Graphics.
(compares images of the red-white-blue box)
"Accurate Rendering Technique Based on Colorimetric
Conception" by Takagi et al., in SIGGRAPH '90 (compares
images of cars)
The feature article on Akira Fujimoto in the May 1989 IEEE
Computer Graphics & Applications.(compares images of a
hallway in a hotel)
I would be very interested in hearing about any other
efforts to compare real and synthetic images. Also I would
like to get any other references about adding imperfections.
hr3@prism.gatech.edu (RUSHMEIER,HOLLY E)
--
The Fundamental Dilemma of Existentialism:
Eschew obfuscation. Ignore alien orders.