Why Pacific Rim will not be in 3D - Page 2

Thanks. Yep, that's me with my Hague Mini Motion Cam. A nice stabilizer but one day I hope to upgrade to a Glidecam model. Yes, the out of focus look in front of and behind the center is faked dept of field. The blur is created in one of two ways: with photoshop processing and with tilt shift lenses. These lenses are two piece lenses that allow you to tilt the glass to get the right angle of blur in camera.
But since one picture seems far more miniaturized than the other, the obvious reason is the shallow depth of field trick. Also, as I noted above, the "tilt" in tilt-shift is from the tilting of the pieces of the lens, not the tilting of the camera downward.

Here are some stereo examples I created:

Normal:


Hyper:


Normal:


Hyper:


Let me clarify what I was saying with parentheses: "The (portions of the image that are in front of and behind the point our eyes are converging on) doesn't get soft, it just doubles and splits apart." That can not be said of a simple 2D source image.
Hyperstereo in your scenario would result in an intolerable amount of parallax on the distant background

Uncharted and any game that allows you to shoot replay screenshots are great ways to learn first-hand about the consequences of your actions with a camera's configuration.

Fantastic example! Certain shots, especially around 1:53, are strongly impacted by the hyperstereo.

I would actually like to see a movie made like this-- just as an experiment. It's not realistic but it is mind-blowing to see so much depth information.

It's tilt-shift to most people.

http://commons.wikimedia.org/wiki/Category:Tilt-shift_miniature_faking

Photoshop filter imitates the tilt-shift technique without needing an expensive lens, and tilt-shift imitates a super super super wide aperture to capture the hyper shallow depth of field effect on things that are very large and very far away.

Technically that is not a binocular cue you were seeing, so no it's not a normal binocular cue. It is the monocular cue that is motion parallax. You are flying so fast that the motion parallax is able to shift the perspective temporally. No two eyes needed.
Again, with my above correction in mind, hyperstereo is not a realistic solution. And drastically changing one's artistic vision is not adapting- it's just sacrificing your vision for something different. True adapting would be to maintain your course but go about it differently. That would mean shooting more scenes with foreground objects in frame to provide at least multiple layers. He could shoot more from the surface of the creatures, like from their shoulder, showing the arm extending deeeep into the frame, then show a close-up of the point of impact. Then go inside the machines and see the human operators, and tell their smaller scale story.

The problem with that is, it would require re-writes.

I made this humorous but true illustration to show how hyperstereo affects the sense of scale of the spectator. Hyperstereo images simulate the point of view of someone who has a gigantic head, and this causes the scenery to shrink relative to the viewer's size.



But in actuality you could say the viewer's size is not growing, the city is shrinking.
Edited by cakefoo - 8/5/12 at 2:20pm

I find the discussion about miniaturization rather fascinating. It appears more of a psycho-percetion phenomena, not an actual relationship in physics. I wonder if it has to do with the fact that most people are so used to seeing 2D on TV and in the movies that when they see 3D it automatically is a visual they just aren't used to seeing as "normal" In my case I do watch 3D stereographic content far more than most people as I've been editing it for over a year+ now and this editing has me watching 3D on my Vizio at least 5-7 hours a day. Then, I may watch a 3D movie twice a week plus nearing the end of my own projects, I'm dumping a review iso render to a hard drive for viewing on my 110" screen in the evenings with my wife for her comments and reviews too. She watches 3D much less, but when she sees a problem in the images she lets me know. Usually its something that I did really wrong like a missed occlusion that I need correcting. Most of her critiques are usually with my writing where she suggests a different way of explaining something.

A couple years ago I had a DLP projector and while she never saw the rainbows, I could view with a pattern that I could trigger them all the time or avoid them if I wanted. I reduced the viewing of those to a science here but never could actually say they were in my imagination or real.

cakefoo- If that were a true hyperbase image, even represented in the 2D flat world the girl's legs would be 4 legs, not 2. ( if we're talking how hyperbase reacts when too close an object is in the scene. ) I don't believe you shoot hyperbase so maybe you wouldn't know about minimum objects effects of hyperbase shooting. In my widest SB shoots, the minimum object distance that will converge is about 150 ft away for a 35mm lens and 28" IA camera separation. and yes, it will also appear larger than the normal stage I am shooting for a 28" IA scene which is 2000' to 2-3 miles. In my Bryce Canyon project, I had to toss many clips where a large object, like a fence rail some 50 ft away got into the shot. If the scene was a keeper, I would do a take 2 with care not to pan into the nearby offending object and then slide my IA to something that would converge the object at 50 ft or about 5" IA and re shoot it. In post I could then use both even side by side and never know the IA changed. Why? Because as the near object came into view, the viewers attention is drawn to it and they would not notice a flattening of the distance in the shot. Kind of like stage magic art of misdirection.

There is a company ( Element Technica IIRC )that makes a wide stereobase rig, very high end that uses a series of servo controlled auger drives that allow the shooter to change the stereobase of two RED cameras as smoothly as the zoom lens. It is quite expensive rig but can produce a seamless manipulation of 3D in a wide range of lens focal lengths and a scene change from several miles to about 5 ft. I saw it at NAB.

Sorry, didn't mean to ignore you. Been busy with family visiting, shooting a 3D pool party and trying to watch an e-bay bid today.

The first instance of the beehives rock is 1:00 into the video so it's not far. The area is discussed in more detail at about 9:55. The rock where the group of tourists are is about 35-40 tall as I recall and the one next to it is closer to 50 ft.

To any who understand the purpose of tilt shift effect-

http://www.youtube.com/watch?v=6JmSmKYxqow

Why did the shooter/editor add the Gausian blur effect at an angle across the left and right corners?

Mostly, I liked the time lapsed effect with hyperstereo but still don't get it with the blur added. That's a technique that I may just file away for now.

Don, the female viewer in that image is NOT part of the 3D image. She is supposed to represent us, the viewer. She represents where the 3D camera would be positioned. Her eyes are the camera lenses. And the lens interaxial dictates how far apart her eyes are, which increases her head size, which decreases the world size. Please re-analyze the image now with this clarification in mind.

Here is a photo I shot for the miniature effect:



And here are a couple videos I shot with interaxials of 6-9 inches, but I only did it to compensate for distance from the subjects.
.

The gradient blur is added as a cheap way of making it look like shallow depth of field:



Most people when they see that blur think that it is in response to the Z axis location of the blurred image-- the further in front of or behind the point of focus, the softer that element will appear to be.

The reason most tilt shift images are shot from an upper vantage point is because at that location the resulting image will usually have the closest objects on the bottom part of the frame and the furthest objects toward the top. This makes it easy to decide which direction to render the blur gradient to create a convincing depth of field illusion.
Edited by cakefoo - 8/5/12 at 5:40pm

Nice Football shooting! It looks like you did some slomo in post but what twin cameras did you use to shoot the high speed? While many claim that consumer camcorders shooting high speed can't be synced without genlock if you were using cameras that were not genlocked your sync was damn good. Also, there were some shots that had the disparity exceed the rule of 3% so I see even you allow this when you think it's important.

Had a few minutes to look it up and I see you used a Sanyo SH1, right? Specs say it is 300 fps with 60 fps playback at 8 Mbs. SD resolution. Still, while not HD, it was an impressive effect even if soft image. Did you manually sync the two or does the Sanyo's work with a controller?
Edited by Don Landis - 8/6/12 at 12:58pm

Thanks for the feedback No sync controller. I snap my fingers. Gets people's attention too!

Yep, hard to beat traditional sound sync. tools. I actually use a real slate and clapboard so I have a visual cue as well as sound. In a crowded public environment I use the sound only by speaking between the two cameras so as not to draw attention. Others manually fire a strobe, that works too because the pulse is so short.

I don't have red and blue glasses. You mean give each building a sense of depth from one another? Why would you want to do that? Its not how our visual system works. BUT, you COULD make that photo, and the robot photo look like you were viewing it from a helicopter, giving it its proper sense of grande scale....Holy......

Why in the @#$%#$ 23@#$%#$%, #$%@, @@#$@#$ WOULDN'T YOU WANT TO DO THAT?

OF COURSE DEL WHOEVER-GUY MADE THE WRONG DECISION!

OF COURSE!

3D can mimic ANY PERSPECTIVE, but you've got to film it right.

It should and always can, look like a.......god........damn...........window to a world, if you film it right...

Come back when you've calmed down. I refuse to respond to that babble.

When you own the rights to a movie you have the right to do with it what you want. WB decided if delToro was incapable of doing the 3D version ( by delToro's own assessment of his ability in 3D) they would just give it to someone who does know how. If the 3D is a flop, then we can say WB gave it to the wrong people. In defense of DelToro's desire to not do 3D, if he wants full control, he should not have gone to money people ( WB ) who will manipulate the film as they wish. He should keep control over his work, pay for it himself, or get money from investors who have no say in the production and distribution.

I just saw Titanic in 3D and was more entertained by the 3D version than the 2D. 3D conversions keep getting better and better. However, there were many CGI scenes that still looked fake by today's quality standards. This had nothing to do with 3D because those same scenes I felt looked fake did so whether in 3D or 2D. Of course, some of those old scenes used inverse kinematics for motion and the technology of that was nowhere near as detailed as motion tracking we do today. Crude motion will distract, as will forced perspective, to make the scene fake looking before we even get to it being converted to 3D.

The more I watch 3D in any movie and then return to 2D, the more I have trouble thinking the 2D looks fake and 3D real. I'm sure that is due to the real world is 3D and not flat 2D.

I also found in the article that 3D is now the rule in Asian theaters, not the exception. The US may have pushed the s3D art form back in the limelight, but I guess it will be the Asians who will make it mainstream. US made IMAX and 3D movies are getting the prime showtimes and causing a slump to the Chinese produced films.
Edited by Don Landis - 9/15/12 at 6:12am

More like YOUR OWN assessment of Del Toro's SUPPOSED lack of expertise.

I respect Del Toro's decision to not do 3D. Exaggerating the interaxial to get skyscraper-sized monsters to have a 3D pop would have been unnatural, while on the other hand, leaving the interaxial natural would have meant a single layer occurring deep in positive parallax, which for a film that will have large-scale creatures in the center spotlight, could mean a lot of flat shots that people will be rubbing their eyes at.
Edited by cakefoo - 9/16/12 at 1:44am

cakefoo- I find it interesting that you hang on the claim that these bizarre worlds of sci fi have to be natural. IMO, there is nothing natural about science fiction and the whole idea of being science fiction is that it CAN be UNNATURAL. That's the whole idea. It is a fictional world.

If WB can get an entertaining movie for us to watch, I'm all for it. You want to hang onto the philosophy that I should not be permitted to enjoy PR in 3D while I say bring it on and let me enjoy it.

Even though Del Toro did say he doesn't know how to do 3D of this world, I will give you this much. He doesn't want to do it but it appears this is no longer his decision. If WB can do it, then let me watch it and if entertained, I'm happy. If not, then all I can say is WB also failed to achieve what they were trying to do. It does appear that WB feels they can achieve a favorable experience in 3D to make money ( ROI) on the film. I have no problem with that.

We all know that 3D illusion is a variable experience, so you do not get to prevent me from enjoying that experience just because you have to rub your eyes when you watch it.

I disagree. No way would a city look like that in the real world. The distances from the viewing point to the nearest building is far to big to perceive 3D. Consequently in trying to make it look 3D over an unnaturally large distance makes the scene take on a miniature village look.

Since when was better stereoscopic depth perception a devlopment that occurred in a sci-fi movie? Usually TECH is evolving, but not humans. One exception would be, say in Spiderman or Captain America, putting the viewer into that character's head after they've gotten super-human abilities. He never said that. Please stop lying. All this is about is, you're just butthurt that Del Toro doesn't think your hyperstereo style is appropriate for this film.
I've not once said you can't enjoy hyperstereo. I have however defended Del Toro's natural 3D style from you, as you continually bash his stereography principles. His only two choices were shoot hyperstereo, or shoot natural. Considering the movie centers around these skyscraper-sized creatures, it would have been flat-plus-positive-parallax for a good portion of the film.

Give it a rest Don. After a month and a half, you still don't understand or respect these things and never will.
Edited by cakefoo - 9/16/12 at 1:11pm

Yes, you have defended it with great voracity. He should send you free tickets at least for all your noble to his cause effort. Me, I just hope the film will be fun and entertaining. I'm on no mission to save or destroy the director. The 3D progress in the movie industry will do just fine without director's like Del Toro if he never tries another 3D movie. I am pleased that WB has seen fit to at least do the conversion. But we do know that shooting in hyperstereo when the scene called for it would have done a better job of fixing the very problem he mentioned. Maybe they should have brought Michael Bay in as a consultant. He certainly has done a good job with large robotic creatures and buildings in 3D.

In the beginning, I stated my objection to Del Toro's claim on the flatness and how I felt it could be resolved. You shoot stereo beyond the natural 65mm but then you make a statement like above where I have to question your understanding of the capabilities of hyperstereo techniques. What I took from del Toro's interview, and it was very brief explanation, was that he hadn't considered using hyperstereo at all. His explanation described only the limitations of fixed 65mm IA stereo for 3D. Maybe he doesn't want to deal with the additional complexities of hyperstereo and I can accept that. But, there are far more ways to bake the cake to tell the story than restrict one's approach. Being open to new ways of doing things is what made guys like George Lucas what he is.

Give it a rest? Sure, why not? PR isn't going to hit the theaters until next July anyway. Meanwhile there are many new 3D movies out I haven't seen yet. Just ordered 5 more today.

The topic in the replys make me confuse.

I haven't seen the movie but I recall people saying certain shots you describe were not to their liking because they weren't realistic. Also, just because Bay did it with 30 foot robots doesn't mean it's as easy to do it with robots that are 250 feet tall. They'd have to get back further from the monsters to get their bodies in-frame, and have to therefore increase the interaxial even further. Alternatively, if they go for a natural look, they could film it from a 65mm interaxial, but then it would just be one flat layer for half the movie. I'm sure people would be demanding a refund. The last choice they have is re-writing the storyboards, re-composing so that there's fewer long-distance shots and more on-board shots from the inside of the robots, or on their shoulder as they're, say, hurling punches... or however they fight. I don't know. The problem with that option is, it just would have been a big re-write, would have delayed the film, cost money, etc.
When you're a filmmaker and your goal is to give the audience the looking-out-a-window sensation, hyperstereo is counter to that design. But just because I believe that, doesn't mean I universally disapprove of hyperstereo altogether...
Then you have been misunderstanding things since the very start of this discussion. Admittedly his English is not all that great and you are apparently struggling with it, but he says this about hyperstereo, he just doesn't call it that:

"When you look at large buildings from 300 feet away and you move a little, the buildings don't have noticeable motion parallax." He continues, "To force the parallax on robots that size (in other words, to increase the interaxial distance), you're making them miniaturized. You're making them human-scale."

To back up his claim that wide interaxials decrease the sense of scale, we look to the experts at Disney/Pixar:

Overseeing the film from a 3D technical perspective was stereoscopic supervisor Bob Whitehill. “With our work on the first two ‘Toy Story’ films, we really found a visual 3D language,” says Whitehill. “We learned that a lot of 3D has to do with the camera separation between the left and right eye. And since we’re living in a world of toys, that inter-axial separation is actually quite small. And because we want to see this world through the scale of the toys, through the scale of Woody, we learned how to set the distance between those cameras—the left and right eye cameras—based on Woody’s size. When we got to ‘Toy Story 3,’ it was very easy to create that sense of scale. For example, with ‘Up,’ you would default to a 2.25-inch camera separation, whereas with ‘Toy Story,’ we were defaulting to a 1/3-inch separation. It made a huge difference in getting the toy’s point of view and giving a greater sense of scale.” (source)

So tiny interaxials increase the sense of scale, therefore larger interaxials decrease it.