h a l f b a k e r yRight twice a day.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
This idea would make movies (mostly animated CGI movies) appear smoother.
I'm probably the only one that notices this/complains about it, but if you've watched any CGI movies, you might have noticed that when there is a panning motion, or a character or something is moving across the screen it appears
very 'jerky', this is probably because it is not very many frames per second, to save time and money.
This idea is expensive, impractical, and requires specially formatted media.
It would include multiple projectors (I don't know how many, it probably depends on how much money you can spend), and either motorized stands which can turn and aim the projectors up and down very quickly, or else a complex system of motorized mirrors. It will also require a complex computer control system with sensors to determine the position of the mirrors/projectors.
It would work by having different scenes, characters, objects, and whatever displayed with the separate projectors. Instead of having many still pictures or frames, which differentiate from each other by some amount. The projectors would move the "objects" separately, and the "objects" would simply change shape, color, size, whatever is required. The "object's" changes in size, etc. would still appear slightly 'jumpy', but overall the movie would appear much 'smoother'.
The same software used to make the CGI could still be used, but the movie would be created in a different format.
I don't know how all the systems would work exactly or work together, but I'm sure its possible.
(?) AV Forum
www.avsforum.com My companys forum. Anything you want to know about front pojection systems... and the existing matrix driver technology. [evilpenguin, Jul 22 2007]
AV Forum
http://www.avsforum.com/ evilpenuin521's companys forum. Anything you want to know about front pojection systems... and the existing matrix driver technology. [BJS, Jul 24 2007]
[link]
|
|
I'm sure this could have many other uses than just animated CGI movies. |
|
|
The software equivalent would be to not re-light individual frames, but instead just move the things that move as if they were cardboard cut-outs, their colors and shadows frozen in place. |
|
|
Hm, I wonder whether people would notice. |
|
|
How would you accomplish that? Would it basically create more frames per second? |
|
|
There are matrix processors that already exist. They are used to create a bigger image by employing mutliple projectors. Their images overlap (seamlessly) to tile the image across the screen. I'm sure that with the correct settings you could actually get your plan to work. |
|
|
And don't judge CGI on film! Watch the movie on a high-end digtal projector, and see what is being done with the technology. |
|
|
Hmmm... so I guess digital projection isn't so jumpy. Thanks for the annotation. |
|
|
First off, I need a little clairafication.... Are you suggesting projecting the character by a projector and the backround by another? And the projector projecting the character, is only projecting the character, and not anything around it, like a cut-out? In other words, every projector produces a rectangular image. Everything digital will be very dark gray for black. That means even if the projector is outputting a image of the character, it would still have the rest of the its space to fill. |
|
|
As for the matrix driver, it is a professionally coded unit that is programmed specific to the application. With that said, it could drastically simplify your idea by digitally projecting the characters from a collection of projectors. There are also x-y articulation mounts out there that could then move the projectors like you discribe. |
|
|
Your idea is possible in my opinion. The problem is it wouldn't work that well with pre-exsitig movies. It would be easy to code a new movie to the format....... |
|
|
It's not just the number of frames. GCI can
look jumpy if there is no motion-blur
rendered into the frames. The same effect
can be see in live action if a short shutter-
speed is used to record each frame of film.
This is about cost, too. Motion-blur is
computer intensive to render. |
|
|
Ah. I diverge here, but as an old-school monochrome film photo guy, I find digital simulated blurring and depth of field effects highly annoying. |
|
|
Digital imagery is an entirely seperate art form, and I've had to train myself to use a different set of skills to appreciate it with. |
|
|
evilpenguin, yes I am suggesting projecting the character by a projector and the background by another, but multiple characters could be projected with the same projector if they all have relatively the same motions at the same time. And the projector that projects the character would not project any non-moving object around the character. |
|
|
"it could drastically simplify your idea by digitally projecting the characters from a collection of projectors. There are also x-y articulation mounts out there that could then move the projectors like you discribe." Is that supposed to different from my idea in some way? |
|
|
//Is that supposed to different from my idea in some way?// what are you saying here? |
|
|
I meant: 'Is that supposed to be different from my idea in some way?' |
|
|
Okay, and no. What I said is supposed to merit your plan. I think that its not only possible, but buildable with off the shelf parts. 'They' make mounts to move projectors and with the lower cost of projectors currently the setup your describe is obtainable. |
|
|
In other words, good work sir, what now? |
|
|
//what now?// well, we just need to start creating the media... |
|
|
I've seen the same jumpiness when viewing a live-action video that was shot in bright sunlight (fast shutter speed). Higher FPS would make it better (1080i60 is smoother but not as sharp during motion, vs 1080p30 which is sharper, but always looks jittery to me during fast motion), but it's the motion blur that really sets it apart. Pause a fast-action scene that looks good to you, and do the same with one that looks jerky. You'll see that the jerky one is crisp and focused, while the smooth one has significant blur in the direction of motion. |
|
|
While motion blur can be accurately simulated, it is computationally intensive, as it is essentially the weighted averaging of several fully-rendered frames. |
|
|
As the processing power available continues to increase, you'll see effective motion blur become more common in CGI films. |
|
|
I know what you are talking about, about the live action video. But live action video can't be fixed with this idea. |
|
|
I wonder if controlling multiple 'objects' with multiple projectors would be less computationally intensive than creating accurate motion blur in 3D. |
|
|
Movie film projectors flash each frame twice (16mm and 35mm) or three times (8mm and Super 8). When panning over a subject with sharp high-contrast details, this will result in an annoying double image. Motion blur can help alleviate this effect. |
|
|
Flashing the picture more often might also help to reduce the double-vision artifacts (instead of two strong images, there would be, e.g., four weaker ones). Since the film must be advanced during one of the dark intervals between flashes, however, it may be difficult to increase the pace of flashes without overstressing the film. |
|
|
A more interesting notion might be to design a projector that would flash continuously at 48hz but could selectively advance or not advance the film with each flash. Portions of a film involving large broad pans could thus be projected at 48fps while the bulk of the film was at 24fps. |
|
|
This idea really isn't about film... |
|
|
Projectors flash multiple times per frame so that we don't notice the dark frames so much. We'd notice 24 dark frames per second as a horrible flicker. With 48 or 72 dark frames per second, we see it only as a slightly dimmer image. Those of us who are particularly sensitive still notice the flicker (I'm sensitive up to about 100 hz), but it's not as bad as 24. |
|
|
As for multiple objects, most CGI films are done in several layers already (background, individual character, effects), with motion blur being applied only to each portion as necessary. |
|
|
With that approach, there are shortcuts which provide a reasonable approximation. The most common is to perform a directional blur on each layer in accordance with a differencing pass performed against the previous frame. If this is done by an animation package that tracks the object model rather than just the output image, this method can provide fairly good results at very low computational cost. |
|
|
Projecting each character with a separate projector would make the character's motion across the screen smoother, but wouldn't help at all with the jumpiness associated with the local motion of the character (moving limbs, etc...), if the issue is still that of low frame rate combined with sharp rendering and fast motion. As another downside, there would be the additional effect of a black mask behind the character which would not move along exactly with the separately projected character. This mask would suffer the same jerky motion that you find so objectionable, making your solution unworkable. |
|
|
I've been thinking about things that would not make this idea work as I originally imagined it to, and [Freefall] talked about all of those things I've been thinking about. |
|
|
This idea would still work if none of the projections overlap (if it had no background for example). |
|
|
I'm sure this idea could be used for a different purpose. |
|
|
I thought conventional film shutters only ran at about 32 frames per second. That's not too far off from 24; would it really cost that much more to render the background at 32? For that matter, why would you need to render the background at anything more than 32, since in most instances the background, being what it is, isn't usually required to change quickly? |
|
|
If I'm understanding this idea correctly, the idea is to use separate systems to render different parts of the movie independantly, and thus sharply reduce the horsepower required to display everything simultaneously. To bring it all together, show backgrounds, foregrounds, and characters (or however you want to divvy it up) all simultaneously via three separate systems onto one screen, and record that. |
|
|
CGI can be made to 100 or even 1000 FPS, but since no projector nor TV/monitor is that fast (yet) it wouldn't do much good. Sometimes the jerkiness is wanted in a scene for effect. |
|
|
[elhigh] The idea is mostly to improve the motion picture quality, theoretically for approximately the same amount of 'man-hours' worked to create the movie. I'm not worried about reducing "the horsepower required to display everything simultaneously.", and the number of projectors required mostly depends on the movie and the amount of money to spend. |
|
|
[jutta] if I understand your interpretation correctly, I believe that is how the earliest DVDs were produced. It produces some unsettling effects. |
|
|
I am by no means an early adopter - can't afford it. But I will buy used equipment from early adopters when the price comes down. I got an early Star Trek TNG DVD set about two years after they were introduced; by that time the quality of new DVDshad improved. Watching the older version was enlightening - parts of Picard's face would move, then the rest would follow. It's subtle, but noticeable. |
|
|
No offense, BJS, but have you seen a live action film lately? |
|
|
Like, you know, with people, and panning movements? |
|
|
Ok, I just got back from Beowulf. If seeing this is what triggered this idea, then I can understand why you brought it up. It's not a limitation of CGI, especially not 3D CGI. |
|
|
I think this is just a budgetary thing. |
|
|
Beowulf cut a lot of corners. |
|
|
It looked like another sequel to Shrek. It probably would have been improved if Angalina Jolie and her prosthetic hips were replaced with a wisecracking donkey. |
|
|
elhigh, what you describe could also have been compression artifacts. An early lower bandwidth of the medium might have called for stronger compression, which would more visibly focus on the areas of greater movement - talking faces - over dark, largely unmoving, larger areas - bodies. |
|
| |