AVS › AVS Forum › 3D Central › 3D Source Components › When small interaxials just don't cut it!
New Posts  All Forums:Forum Nav:

When small interaxials just don't cut it! - Page 6

post #151 of 1087
7/21/11 at 4:51pm
Sorry, no zoomimg while shooting video with that camera. So forget shooting moving people or any up close sports (like kung fu). Not versatile. Great for stills, though, and good screen.

Oh, btw - a butterfly with its wings up is flatter than a cardboard cutout in REAL LIFE. Yipes.
Reply
Reply

AVS Top Picks

post #152 of 1087
7/21/11 at 4:54pm
Thread Starter 
Speaking of canyons in 3D, my daughter and son in law took my Fuji W3 on their trip to Hawaii and stopped by some canyons near Las Vegas.
I made a slide show of some of the 3D pictures from the canyons and they are viewable canyons in 3D shot with Fuji W3.
The player gives many options for displaying in 3D including 2D.
I use checkerboard myself.
The W3 has an interaxial distance of three inches which I really like.
Too bad the video from this thing is so bad.
Reply
Reply
post #153 of 1087
7/21/11 at 4:58pm
Thread Starter 
Quote:
Originally Posted by markr041 View Post

Sorry, no zoomimg while shooting video with that camera. So forget shooting moving people or any up close sports (like kung fu). Not versatile. Great for stills, though, and good screen.

Oh, btw - a butterfly with its wings up is flatter than a cardboard cutout in REAL LIFE. Yipes.

Should I take up the challenge of shooting butterflies in 3D with my two camera rig?
I've deleted many 3D videos with butterflies in them but I'm game to shoot some more.
Reply
Reply
post #154 of 1087
7/21/11 at 5:18pm
As usual, Frank, you go above and beyond. Remarkable.

Here's what I'd like to see - assuming a rig setup similar to your previous shot - JVC on bottom, GoPro on top, Canon's flanking them:

No need for more than 10 second shots.

Shots with subjects in foreground, middle ground and background.

JVC set to wide angle, Sony set to match (or vice versa); Canon's at same framing.
JVC, Sony and Canon's set to their max wide angle settings, from the same location.
JVC, Sony and Canon's set to the same telephoto framing.
JVC, Sony and Canon's set to their max telephoto settings.

Repeat with the cameras in different locations and subjects at different distances.

I'm sure when you get this set up, other shots will suggest themselves. Anything that demonstrates the differences will be valuable for me. I'd like to see some shots with lots of color and detail, also.
Reply
Reply
post #155 of 1087
7/21/11 at 5:26pm
Quote:
Originally Posted by Frank View Post

Should I take up the challenge of shooting butterflies in 3D with my two camera rig?
I've deleted many 3D videos with butterflies in them but I'm game to shoot some more.

Maybe Mark doesn't know that with your remote rig you can get the cameras closer to wildlife than he can by carrying his Sony. Bears, deer, squirrels, birds and, I imagine, butterflies seem to have no fear of your robotic 3D system.

There's room for all kinds of 3D programming. I for one can enjoy watching programs that have no people in them at all, but I never assume that everyone else's tastes are the same as mine. Does that make me un-real?
Reply
Reply
post #156 of 1087
7/21/11 at 5:53pm
Quote:
Originally Posted by Don Landis View Post

John- did you look at some of those "experiments" he did? There appeared to be something radically wrong. After carefully studying I figured out what it was.
As you zoom in the background goes tighter ( correct) but the foreground goes wider. (incorrect) The only way this can be possible is if the camera is on a dolly and is pulled back as the lens zooms in.

The treatise is nice if only it was correct. Am I missing something?

Rather than draw scenes that support a theory, how about seeing a real world example with real cameras. I believe any test would demonstrate what Mark and I have been seeing in our shooting and what we have seen in Franks shooting.

Hey Don- his write up is right on. The results match the stereoscopic projection equations above. When he dollied the camera backwards while zooming in (to keep the subject exactly in the same place), he performed the classic 'compression' used in 2D for dramatic effect. He is increasing zoom (D from the equations above) and z (distance to camera) which reduces the relative contribution of O (camera separation) resulting in a reduction in stereo separation and a flattening of the image. If he continued to zoom and move the camera away, the image would flatten completely to 2D.

For a real-world example with the TD10 at maximum zoom, see my video here at 4:56 - http://www.youtube.com/watch?v=4JzUt9V-2KU . Note the building structure in the background is almost completely flat, and the light pole which is a few hundred feet closer, is separated from the background (better 3D effect than I expected when shooting it; however it took hundreds of feet of 'z' difference to produce the projected depth difference at max zoom). If you watch until the end you'll see some dramatic 3D effects up close (looks even better playing off a BD disk on an HDTV).
Reply
Reply
post #157 of 1087
7/21/11 at 6:00pm
Quote:
Originally Posted by Frank View Post

My ideal 3D camcorder for traveling would be like my Fuji W3 except for having a built in MVC encoder for high quality 3D unlike the AVI crap the W3 puts out.

That's what I want from a manufacturer in a second gen, high quality 3D camcorder. It isn't hard to get lenses 3" apart, and in a form factor people are used to. They're just used to it in a still camera, instead of a camcorder. I love the Fuji W3 for 3D stills, but it took me only a few shots to realize that its video was a complete waste of time.
Reply
Reply
post #158 of 1087
7/21/11 at 6:08pm
Quote:
Originally Posted by JohnSchultz View Post

Hey Don- his write up is right on. The results match the stereoscopic projection equations above. When he dollied the camera backwards while zooming in (to keep the subject exactly in the same place), he performed the classic 'compression' used in 2D for dramatic effect. He is increasing zoom (D from the equations above) and z (distance to camera) which reduces the relative contribution of O (camera separation) resulting in a reduction in stereo separation and a flattening of the image. If he continued to zoom and move the camera away, the image would flatten completely to 2D.

For a real-world example with the TD10 at maximum zoom, see my video here at 4:56 - http://www.youtube.com/watch?v=4JzUt9V-2KU . Note the building structure in the background is almost completely flat, and the light pole which is a few hundred feet closer, is separated from the background (better 3D effect than I expected when shooting it; however it took hundreds of feet of 'z' difference to produce the projected depth difference at max zoom). If you watch until the end you'll see some dramatic 3D effects up close (looks even better playing off a BD disk on an HDTV).

Moving the camera in while zooming out is the classic "vertigo" shot (as in Hitchcock). I nearly got sick the first time I saw that shot. Either shot can have quite a different effect in 3D than 2D. It's fun to explore the differences and similarities shooting 3D vs 2D. I'm learning a lot.
Reply
Reply
post #159 of 1087
7/21/11 at 6:14pm
Quote:
Originally Posted by Joseph Clark View Post

Moving the camera in while zooming out is the classic "vertigo" shot (as in Hitchcock). I nearly got sick the first time I saw that shot. Either shot can have quite a different effect in 3D than 2D. It's fun to explore the differences and similarities shooting 3D vs 2D. I'm learning a lot.

Right- and in 3D objects would go from flat to round in that case.

The Fuji W3 footage can be cleaned up somewhat with Neat Video (http://www.neatvideo.com/). To make that form factor really useful (in addition to improving image quality), adding an image stabilizer as good as the TD10 would really help.
Reply
Reply
post #160 of 1087
7/21/11 at 6:19pm
Quote:
Originally Posted by JohnSchultz View Post

Right- and in 3D objects would go from flat to round in that case.

The Fuji W3 footage can be cleaned up somewhat with Neat Video (http://www.neatvideo.com/). To make that form factor really useful (in addition to improving image quality), adding an image stabilizer as good as the TD10 would really help.

Yeah, I was just thinking that it would be fun to do a 3D vertigo-type shot (move in/zoom out). I would think it would have the potential to be even more dramatic than a 2D vertigo shot, for that very reason (at least up to a certain point).
Reply
Reply
post #161 of 1087
7/21/11 at 9:00pm
Why is it that I first describe the only possible way that shot was done and then everyone wants to come in and explain to me that's what he did? What I didn't get was the use of it to derive conclusion that things go flat when you Zoom in. I didn't see it in my zoom in shots and I don't see it in Marks. I don't think the TD10 is unique and I don't think the JVC TD1 has a good enough zoom rang to matter.

Mark- I didn't realize the butterfly was 40 ft away but then I didn't see much pop on it either. You had much better negative parallax shots in that collection, but then those were distant zoomed in shots with something that popped in up close to the camera. You said we agree that near subjects of great distances are impossible to do popout and that is where I said Frank's bench system has the advantage.

I do have a question on that fixed wide angle lens you referenced. I figured using a single 0.7X wide angle adapter would not work because the lens needs to centered to not cause geometric distortion. Are you using two smaller dia wide angle adapters? How do you mount them? Pictures please.
Reply
Reply
post #162 of 1087
7/21/11 at 9:32pm
Quote:
Originally Posted by Don Landis View Post

Why is it that I first describe the only possible way that shot was done and then everyone wants to come in and explain to me that's what he did? What I didn't get was the use of it to derive conclusion that things go flat when you Zoom in. I didn't see it in my zoom in shots and I don't see it in Marks. I don't think the TD10 is unique and I don't think the JVC TD1 has a good enough zoom rang to matter.

Did you review the projection equations earlier in this thread? The math explains why zooming in (and objects far away) result in flat 3D. Did you see the example (flattened) TD10 footage at full zoom I posted?

Flattening while zooming in 3D is well known:
http://realvision.ae/blog/2011/04/st...zooming-in-3d/ (workaround to allow zooming without cardboarding at the expense of resolution)
http://www.dashwood3d.com/blog/begin...ereoscopic-3d/ (Lens and Focal Length selection paragraph)
Reply
Reply
post #163 of 1087
7/22/11 at 5:44am
Thread Starter 
Quote:
Originally Posted by Joseph Clark View Post

As usual, Frank, you go above and beyond. Remarkable.

Here's what I'd like to see - assuming a rig setup similar to your previous shot - JVC on bottom, GoPro on top, Canon's flanking them:

No need for more than 10 second shots.

Shots with subjects in foreground, middle ground and background.

JVC set to wide angle, Sony set to match (or vice versa); Canon's at same framing.
JVC, Sony and Canon's set to their max wide angle settings, from the same location.
JVC, Sony and Canon's set to the same telephoto framing.
JVC, Sony and Canon's set to their max telephoto settings.

Repeat with the cameras in different locations and subjects at different distances.

I'm sure when you get this set up, other shots will suggest themselves. Anything that demonstrates the differences will be valuable for me. I'd like to see some shots with lots of color and detail, also.

Will do!

Hey, I just had an idea! What do you think about this;
I order another Sony HR10 and mount both on the bench and control them with a stereoscopic LANC controller. Then in Vegas I can work with four separate video streams, (outer left, inner left, inner right, and outer right) and fade all combinations in and out with Vegas. Should make for some very interesting results.
(Best of all worlds?)
Reply
Reply
post #164 of 1087
7/22/11 at 8:10am
Thread Starter 
I'm extremely curious as to the effect of taking the four video streams from the dual Sony HR10s and at minimum zoom start with the outer left and inner left, and while zooming in, begin to fade out the inner left while simultaneously fading in the inner right. As the zoom continues and the inner left is fully faded out and the inner right is fully faded in, begin to fade out the inner right while fading in the inner right to full zoom.
Will my curiosity overcome my better judgement?

It usually does!
Reply
Reply
post #165 of 1087
7/22/11 at 12:01pm
Thread Starter 
Just as I was finishing up my newest test bench, the UPS truck pulled up with the first of my new Sony 3D camcorders. (TD-10).
Time to start preparing for Joe's tests....

Well, maybe the first of two.
Reply
Reply
post #166 of 1087
7/22/11 at 6:30pm
Thread Starter 
Test rig is ready to go for tomorrow but got to attend my great grand daughters 2nd birthday party out of town.
Weather on Sunday should be perfect for the tests.

Reply
Reply
post #167 of 1087
7/22/11 at 6:45pm
Perfect rig for the test! Really looking forward to seeing the results.
Reply
Reply
post #168 of 1087
7/23/11 at 8:10am
Thread Starter 
Why oh why did the brilliant people at Sony in all their wisdom decide to disable the composite video output when the LANC is used?
Reply
Reply
post #169 of 1087
7/24/11 at 10:01am
Thread Starter 
The tests have begun!
I had to end my initial tests early when I realized the Sony camcorder wasn't properly aligned with the others.
More machining is required and then the tests will continue this afternoon.
Reply
Reply
post #170 of 1087
7/24/11 at 11:16am
Thread Starter 
After reviewing the 3D video from the three cameras/rig I am pretty confident that a lot of people are not going to like what they see.
I'll be putting all of the clips on the Vegas timeline and switch between them.
Should be very interesting.
Will take me a while to upload everything.
Reply
Reply
post #171 of 1087
7/24/11 at 5:53pm
Quote:
Originally Posted by Frank View Post
After reviewing the 3D video from the three cameras/rig I am pretty confident that a lot of people are not going to like what they see.
I'll be putting all of the clips on the Vegas timeline and switch between them.
Should be very interesting.
Will take me a while to upload everything.
Great. Looking forward to it, no matter what it reveals. Truth is good. Thanks for doing this, Frank.
Reply
Reply
post #172 of 1087
7/24/11 at 7:52pm
Thread Starter 
Quote:
Originally Posted by Joseph Clark View Post
Great. Looking forward to it, no matter what it reveals. Truth is good. Thanks for doing this, Frank.
Be careful what you wish for.
I was expecting a significant difference in 3D but the difference in image quality was a great surprise.
Editing this is going to take me a while.
Maybe I should mail you all the raw files and let you edit them.
Reply
Reply
post #173 of 1087
7/24/11 at 8:13pm
Quote:
Originally Posted by Frank View Post
Be careful what you wish for.
I was expecting a significant difference in 3D but the difference in image quality was a great surprise.
Editing this is going to take me a while.
Maybe I should mail you all the raw files and let you edit them.
I don't take such things personally in the least. I asked for objective tests. It would silly not to be objective about the results.

Do you have enough space in your Dropbox account to upload them, or is there too much data? At any rate, I'd like to have the original files. I could PM my address, or you could just wait for my disc to arrive. It should be there tomorrow or the next day.
Reply
Reply
post #174 of 1087
7/25/11 at 6:24am
Thread Starter 
Doing this test turned out to be a lot of fun and was well worth the effort.
The test was conducted in two parts;
The location was the same for both but the lighting conditions were quite different. The first series of tests were conducted in the morning when it was overcast and the afternoon tests were conducted in bright sunlight.
Both tests took place on the shore of Lake Superior at Canal Park in Duluth Minnesota.
Both the JVC and Sony were in full auto mode and the Canon HF-M32s were in shutter priority mode 1/100th sec. (They can not be in Auto mode because of the image stabilizer)
I would up shooting much more video then originally intended because the conditions were so perfect in both cases.
I don't have any usable raw files that are short enough to upload to my dropbox account.

I hope to have something to upload by this evening.

P.S.
If I had seen a test like this months ago, I could have saved a lot of money on equipment.
Reply
Reply
post #175 of 1087
7/25/11 at 8:33am
Thread Starter 
Currently rendering the first test in Windows Media format 3840 by 1080 side by side.
I'll try a BluRay ISO afterwards but am concerned about Vegas crashing.
We'll see.
Reply
Reply
post #176 of 1087
7/25/11 at 8:51am
Thread Starter 
Quote:
Originally Posted by Frank View Post

Currently rendering the first test in Windows Media format 3840 by 1080 side by side.
I'll try a BluRay ISO afterwards but am concerned about Vegas crashing.
We'll see.

Change of plans.
After seeing that the render was going to take over seven hours, I decided to render it as separate left and right tracks.
Reply
Reply
post #177 of 1087
7/25/11 at 11:16am
Quote:
Originally Posted by Frank View Post

P.S.
If I had seen a test like this months ago, I could have saved a lot of money on equipment.

Sony
Reply
Reply
post #178 of 1087
7/25/11 at 1:11pm
Thread Starter 
Managed to create an 3D Blu-Ray ISO of my first test.
Uploading now.
Reply
Reply
post #179 of 1087
7/25/11 at 8:17pm
Thread Starter 
Here is the first test and my very first try at something like this:
I expect to get better at it as I go along.
Something to keep in mind is the spacing between the Canon camcorders was excessive in some of the clips because of the mounting arrangement.
3D Blu-Ray ISO of Frank's initial 3D camcorder evaluation
Reply
Reply
post #180 of 1087
7/25/11 at 10:10pm
[SPOILER ALERT - Don't read this post if you want to watch Frank's BD ISO first]




Dual Canon HF-M32 (shooting progressive?) vs. JVC TD1 vs. Sony TD10 PQ?

In bright light I would expect:

1. Dual Canon
2. JVC (makes sense given higher bitrate)
3. Sony

In Low light:

1. Dual Canon (Sony might be close here: low light footage I shot on the TD10 turned out very good. My Canon HF11 is good in low light but not as good as Panasonic TM700 or TD10)
2. Sony (more advanced hardware+software to reduce noise vs. JVC)
3. JVC

See also: http://www.avsforum.com/avs-vb/showthread.php?t=1335637

Watching Frank's 3D BD: results are as expected for bright light. The JVC has a bit less saturation and contrast vs. Canon & Sony. JVC is sharper than Sony (higher bitrate and perhaps better glass). Sony looks more heavily image-processed (and/or compressed).

The 3D depth is as expected: at zoom 3D objects are compressed more when the cameras are closer together (follows the math).

The dual Canon's at wide angle produce hyper stereo: objects look artificially small. Use tilt-shift lenses (perhaps on 2 5Dm2s) and will have a very nice miniature effect (can also selective blur in post to achieve a similar effect).

Thanks for the demo, Frank!
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: 3D Source Components

AVS Top Picks

AVS › AVS Forum › 3D Central › 3D Source Components › When small interaxials just don't cut it!