When small interaxials just don't cut it! - Page 3

After about 3 views I began to get the headache. Then knowing how forgetful you are, I opened up the 3D swap and swapped the right and left eye and voila' it looks OK! It's not a natural separation but with a trained eye for 3D I can live with the eye exercise. Fortunately, it isn't long or changing to other views.

Suggestion- Here's what I do, I keep a small WORD doc file with the tags on my desktop. Then when I upload a 3D to You tube I copy and paste it into the tag box so I don't forget anything or mistype it. Works for me!

Actually, I think it was only twice but it seemed like I was enduring the pain for more. I looked at it first in anaglyph and second in SBS on my vizio. Then I swapped the left right.

It seems that all YT processing reverses the left and right so that everyone uploading needs to include that tag. Maybe the guys at YT are using checkerboard and write their code with a swap built in, then the general public comes along and have to swap back.

I'd like to see it. Always interested in such direct comparisons.

Which one did you get, Frank? The one you referenced earlier? I'll be curious to see what it looks like in your rig. I've always been interested in such a device as a competitor to the traditional steady cam ( now off patent) Hint Hint!

Thanks for uploading that video, Frank. It's always interesting to see the differences in 3D cameras. I usually watch 3D first with the glasses on, and then re-watch it as I lift them up and down. It's fascinating to see the differences in the convergence and divergence of objects within the frame with different 3D rigs. I'm too new to 3D to understand exactly what's going on, but the manufacturers seem to take different approaches to it. Distant objects shot with the Sony camcorder almost always look fairly widely diverged, similar to the GoPro. Objects at similar distances with the JVC seem to converge much more. Also, the convergence/divergence with the JVC is different depending on how I initially frame. That is, if I start a shot with objects close up and then tilt up to objects further off in the background, there's greater divergence of those distant objects than if I frame on them at the start of the shot. This tells me that there are some moving elements in the optics and/or on the chips in the JVC that determine how the left/right images come together to create 3D. Obviously, that's not happening with the GoPro, with its fixed lenses. And unless I'm mistaken it's also not happening with your dual camera rig. You don't have variable convergence working on that rig yet, do you?

As I said, I'm too new to all this to draw much in the way of conclusions yet. What I do know is that, mostly, I prefer the sense of depth in the 3D video created by your dual Canon rig. The JVC doesn't provide as much. OTOH, I do like a lot of things about the JVC camcorder. If I frame carefully, the sense of depth is quite pleasing, though not as dramatic. The 3D is clean, and the depth of field is almost unbelievable. Objects from very near to very far are tack-sharp, with incredible detail. I also love the color and contrast of the JVC. I've been to Shaw's Garden a zillion times, and I'm continually impressed by how much the JVC's 3D makes me feel like I'm really there. No 2D camera has ever come close to doing that for me. So, I'm very satisfied with the JVC in those ways.

Here's what might settle some "debate" (if there is any ) about the differences in how your 3 systems perform, and specifically about the relative benefit of a wider IA distance. Put them on tripods and shoot things with settings as close as possible on all three. Set them up so that there are objects near, medium and far away. After that, throw in a shot with the Canons at their widest wide angle setting. I think I know the enhancement that's going to provide to the subject matter. I suspect it will increase the dramatic impact, just as wide angle shots have the capability of doing in 2D filmmaking. But I could be wrong. That's why I'm so keen on seeing such a comparison. If the variables are the same except for the interaxial distance, then we may be able to draw some more solid conclusions about its effect. The wide angle impact should make such a comparison even more revealing.

Almost forgot. It would be helpful to see the effect of telephoto, just like wide angle settings. That is, zoom in to the same telephoto framing with the JVC and the Canons, and we can look for differences in the sense of depth at the zoomed in range. Then, throw in a shot of the Canons at their max telephoto range.

See what I did there? I planned your weekend for you.

Actually, there is little difference in how the manufacturers do things once you see the relationships of the different designs and their methods as they decide what the purpose of the 3D stereo camera will be. One good example of this is the Bloggie3D, vs, the TD1 /TD10's and then we get to the bigger shoulder mounts from Sony and Panasonic. The Bloggie 3D uses a rather close IA but it's true intent is to be shooting up close and personal, the TD-1 and TD10's are mostly for shooting personal stuff of uthers having fun etc. The bigger cams alluded to are intended for shooting sports at greater distances. The bench rigs with wide IA's are intended for obtaining good separation of scenery on larger outdoor sets but ability to adjust for shooting up close like the TD1 / TD10's as well. More important to use same camera personality throughout the project but have the flexibility for a full range of stages.

All the above mystery can be learned by reading one of my favorite books 3D Movie Making by Bernard Mendiburu. He offers explanations on the whole process which contrary to popular belief is more science and has been reduced to formula than art. The book is loaded with drawings to explain the relationships between the focal length and the IO designs of the cameras for a particular shooting stage. That stage has a width, length and a height and a point of distance from the camera. The book does a good job of using drawings to explain these relationships but it falls short when demonstrating who the relationships change as the stage changes. So, to overcome this shortfall, he includes a DVD with the book that uses 3Dstudio Max wireframes of a twin camera and a stage in 3D where there is an explanation of what happens when the stage changes dimensions, the distance to the stage changes and the focal length of the lenses change and the interocular or spacing between the cameras changes to compensate. By doing this in real time you can see how it all works in harmony. He shows where stepping outside the limits in his drawing will produce changes to the eyeball angle in viewing and produce eye fatigue, or not. It is a pretty good collection of explanations, when paper book form falls short. There are other books too but they are more specialized in a particular area and the cost of them is about 6 times what this one is because to cover all Mendiburu covers takes like 6 volumes and it lacks the DVD. The book is available through Amazon. I've read it cover to cover 3 times and watched the DVD twice now. Some of the concepts and relationships are not that easy to grasp.

I'll check that out. Right now, I'm learning as I shoot. For me, that's an even more valuable way to find out how things work, and what should and shouldn't be done. I'm making plenty of mistakes, but every shot that doesn't work teaches me something just as valuable as one that does.

Looks good Frank. I can see the two camera rig advantage for a Bryce Canyon project some day.

Sure. I know the GoPro can't zoom, so the only fair way is to use it as the control camera. That is, set it up and then frame the others as close as possible to the GoPro. The idea is to remove the other variables, so that the difference in interaxial distance is the main one we'll be seeing. It won't be exact, to be sure, but it should be revealing. The other zoom tests should show the differences between the JVC and the dual Canon rig. One thing I remember from your earlier tests was the flattening effect the JVC had on objects it was zoomed in on. But one thing I'd really like to see is a comparison shot of the Canon and the JVC in the same place, and with both zoomed out as far as possible. One drawback of consumer cameras in general is that they don't go very wide. The JVC is even more restricted than typical 2D consumer video cameras when shooting in 3D mode (about a 42mm equivalent). What I'm curious about is the possible added impact the Canons' greater wide angle zoom might give the 3D image.

Joe- When I need to get good z axis depth from my TD10 on a distant shot I zoom in and the dept stereo begins to appear ( within limits) But the advantage I am seeing with Franks tests are that his wide IA combined with wide angle gives much better z axis stage size and can produce depth at distances so that the wide angles give good depth that the close IA like the JVC and Sony don't get. Frank has me convinced that to shoot good stereo 3D of large scenes like mountain ranges and places like Bryce Canyon, the need for both wide angle and depth, the only way is with the wide IA. But if I want to shoot a distance of 150 ft. to a stage of dancers, zooming in can achieve a good enough stereo effect for most people.

I don't have one at full wide angle of this scene but here is a difference between my opening shot which was about 20% zoomed in and the second one at full 10x zoom. red/cyan glasses to view the depth of near middle and far objects.

I think I know what I'm going to see if Frank performs those tests, but I'd like to have a few controlled clips to demonstrate the differences. Frank's in the fairly unique position of having those three systems at his disposal.

I've watched Frank's clips over the last several months, and the wider IA and the wider wide angle shots create 3D with a special look and feel. I've likened it to the difference I felt watching my JVC RS40 for the first time. I've been a CRT, DLP and LCD owner, but the RS40 was my first experience with a great LCoS projector. The tremendous native contrast ratio felt so very "right" to me. I was in love. It's pretty much the way I feel about the JVC TD1, in terms of its picture quality, if not its ability to create the most dramatic 3D I've ever seen. The TD1 has the ability to capture images with tremendous contrast, detail and deep, rich color. In that sense, it's the perfect match for my JVC RS40. I hope JVC comes out with a camera next year that has a wider IA, a broader zoom range and a wider wide angle lens. If they threw in a servo zoom with LANC, I'd be in 3D heaven.

Thanks for doing the test, but I'm just as interested in the gyro. Looking forward to hearing all about it.

The people who make the Blackbird stabilizer have been working on a gyro device for over 3 years. They came up with the Blackbird on their way to a full blown gyro stabilizer. I haven't heard how close they are to achieving it.

Wow. That's a great configuration to do the test. Can't wait to see the results!!!

How did it go with the gyro?

I look forward to it. I felt like a total dufus with the Blackbird stabilizer when I took it out for a spin a while back. Steadicam-type devices are not the easiest things in the world to master, and it had been a long time since I'd used the Steadicam JR on a regular basis. Then I got wrapped up in editing and put the BB aside. I know full well, when I get back to it and start to feel comfortable, it's going to be a whole new ballgame in terms of the way I want to shoot. A gyro stabilizer sounds like it has so much potential to create great shots.