Canon 3D?

[Adam Stanislav]

Originally Posted by Adam Stanislav

You can sync two nanoFlashes, so maybe that is how they do it.

Alister is right. Now that I have my own nanoFlash I realize that the only way to sync two nanoFlashes is by syncing the cameras themselves.

[Bruce Schultz]

I'm all in for theory, it would be difficult if not impossible to move ahead technologically without it.

Here is a link to a clip I've posted on Vimeo of 3D footage shot with 2 Canon 7D cameras.

HerrSchultz 3D on Vimeo (click on description for details and watch full screen if you have a good monitor system) It's an anaglyph output so put on your red/cyan glasses to watch.

It appears to turn some theory about non-genlocked cameras a little on edge. It's not perfect by any stretch, and yet throughout all of the 7/3D tests that I've shot I've yet to see the type of artifacting that not genlocking should cause - for no-sound MOS shots of course. Granted these are short clips, but I think for the 3D enthusiast or home user this setup works pretty well.

Now that my 2 NanoFlash units have arrived I'll most likely be abandoning the 7D's for EX cams, but the 7D's have been invaluable as training tools for 3D motion photography for me.

[Adam Stanislav]

It looks nice, Bruce.

[Sareesh Sudhakaran]

If genlocking is absolutely crucial, can a lanc controller be built for the canon 7d? There are many DIY tutorials on the website.

There are a lot of 3D images and footage on the web. How does one really test these to see if they are perfect 3D? Since each eye is different, how do people on set know if their 3D is right or wrong?

One thing I've noticed in my basic 3D-making is that it is important for the eye to adjust and focus on an image as soon as possible. How does one guarantee that? In Avatar, James Cameron really played it safe with the 3D. It subtly enhances the world he was creating and that itself was a great achievement. Did he avoid any radical 3D because he had no way of truly judging its result?

[Sareesh Sudhakaran]

Originally Posted by Bruce Schultz

I recently toured a very high end 3D facility in Los Angeles. Their zoom lenses are computer locked and their proprietary software not only tracks the two zooms but adjusts the camera mounts for physical discrepancies between the cameras also. Very sophisticated and very expensive at roughly $50,000 USD rental per day per 2 camera rig. So zooming seems to be a slippery slope in 3D. I'd love to hear that it isn't though.

Wow! I was hoping to make a movie in about $250,000! If you don't mind, which 3D facility was it? Or are all of them this pricey?

What about movies like Valentine 3D or Final Destination 3D? The second was made by the PACE system but I'm sure they couldn't afford 'Avatar rates'. Here's to hoping...

[Alister Chapman]

It may be possible to build a controller for Canon cameras as they do have remote control via USB, whether it is possible to use this to bring the cameras in to sync I don't know. The Lanc controllers for the Sony cams work by turning on both cameras in sync, however after a few minutes the sync will drift off, requiring a re start. Genlock locks cameras together and they remain in sync while still connected.

There are many on set 3D monitoring solutions, most of these only work with genlocked cameras, but there are a few dual monitor and mirror type systems that will work with non genlocked cameras.

Cameron uses camera systems designed by Vince Pace, these use modified Sony HDC950 (I think) HD cameras on very expensive computer controlled side by side mounts. There is no "radical" 3D in Avatar as it's not realistic and detracts from the story turning the film into a trick show. Cameron's aim was to produce an immersive experience where the 3D makes you believe you are in the film, not merely watching it, which IMHO he succeeded in doing.

[Sareesh Sudhakaran]

Is 3D doomed to be judged only subjectively on set? Are there no objective standards or tools that can help determine an 'acceptable' range that can then be fine-tuned in post?

[Giroud Francois]

yes, some tools exist for correcting 3D after shoot, but if you budget is tight, these tools cost the hell (several thousand dollars at best).
And frankly i would not take the risk to ruin a scene that cost time and money to shoot, just because i skipped on a simple check in the field.

[Bruce Schultz]

Vince Pace says that Avatar was shot entirely on 2/3" chip cameras, specifically Sony F23, Sony F950 and Sony 1500. The last two were broken out of their normal mountings and placed in new thinner case mountings external to the recorder (in the case of the 1500) and the rest of the electronics. They were still not small enough to do closeup 3D work, so a mirror rig was designed.

The genius of Avatar as I see it is that Cameron chose to marry focus and convergence on the same plane for most of the film. If you pay close attention you will notice that most of the action, especially dialogue scenes take place near the camera and not in the mid or deep background. This seemed to have the twin effects of easy audience focusing and not having but a few semi-jarring cuts to deep focus/convergence shots.

[Sareesh Sudhakaran]

Assuming one shoots with long lenses even for close-ups and two-shots (if there's enough set space), why would anyone want to shoot on a beam-splitter?

I agree 20 feet is important ground for ENG and documentaries, but for feature films, can't there be a workaround just for side-by-side? What's missing here?

[Nick Hiltgen]

You're going to lose a lot of your 3d effect shooting with long lenses, for close up work. It your subjects will appear to be kind of card board cut outs instead of feeling actually 3d. I think the mirrored rig really is your best option. For what it's worth if you don't want to pay 50k a day you could rent a beam splitter and an experienced stereographer for significantly less. Further I don't know that Zoom lenses really are your Ideal choice as you are dealing with significantly more elements of glass.

3d is definitely not doomed to be judged subjectively on set in fact I would say that's probably the worst way to do it as most likely what you're subjectively judging off of is a screen no where near the same size as what your material will be projected on. There are a few programs out there and 3d calculaters, Florian the inventor/designer of the p+s technik rig is supposedly releasing a 3d calculator for that contraption, and there is another program (calculator) called stereoscopic master that will let you enter in all of your parameters to design your shots effectively.

On the canon front I think you might have far more effective 3d if you use a camera with a smaller sensor size (as that's closer to the iris of your eye) however that won't give you the 2d movie look that a lot of people like.

[Sareesh Sudhakaran]

I understand what you mean by the 'forced' focus that long lens gives - but isn't that what Avatar has predominantly? It had very safe 3D where your eyes had to focus on the screen and nothing else. Judging by its success, I guess revolutionary 3d is less welcome. If we use video lenses instead of film lenses, we will get more depth of field for the same frame, so close-ups zoomed on on those lenses will have more depth and wide-distortion in them, wouldn't it?

I guess judging 3d and using the 3d calculators is like cinematography using film and a light meter. I'm sure one day there will be a standard for this. It's difficult to judge on set because there are so many ways in which 3D can be viewed, plus the monitor you use on set won't really tell you how it's going to look like on another screen (cinema or HDTV).

Would the Canon Rebel 2Ti have a smaller sense than 5d? Otherwise we might be better with SI2Ks or Reds with Video lenses.

[Bruce Schultz]

Excuse me if this is a Simpson's "Doh" moment but, I have noticed that the 3D "effect" changes based on different sizes of displays. When I do a convergence of 2- 3D streams in FCP on the 23" computer screen and 23" external HDTV to where it looks good, then when I make a 1080P BluRay and watch the same shots on a 50" HDTV it doesn't have the same look, it generally looks flatter.

Stereoscopic's calculator allows you to define the final viewing screen size, but I don't see any changes in the calculation by switching from HDTV 42" size to theater size, so I'm not sure what to make of that.

[Nick Hiltgen]

Bruce you're absolutely right on both accounts. I've been told you can actually take 3d that works on the big screen and bring it down to the small screen without much difficulty the problem is when you do the reverse it cause too much divergence. which is most likely what you're seeing with your work going from your home editing system to the best buy monitor.

I hope to play around with the stereoscopic calculator some and see what kind of results I can get. (I.e. see if I agree with their findings...)

Sareesh, I think maybe I'm misunderstanding you or you me. While yes long lenses will give you a shallower depth of field, as far as 3d goes it also really does make the charachters seem like they're cardboard cut outs, it appears to be layers of paper instead of an immersive 3d enviornement. Kinda like if you look at something through binoculars.

I think pretty much if you use any lens over let's say a 50mm (in motion picture 35mm terms, divide by 2.5 to get 2/3"-20mm 5.6 or so for 1/3" equivalent 9mm) you'll start to get more binocular vision. To me this isn't really groundbreaking 3d, but then again I'm not really sure what is.

[Sareesh Sudhakaran]

Can we really achieve a sort-of standard 3D viewing experience if we use a 3D calculator - when viewed on screens of different size and technology (internet, TV, HDTV and cinema screens)?