When small interaxials just don't cut it! - Page 8

How much does the small version of that stabilizer weigh, Frank?

Yes, thanks. I'm anxious to see the results of your newest experiments.

I'm not sure what to make of the latest test footage, Frank. The Sony seems more contrasty (in a good way) than both the JVC and the Canon's. The JVC looks overexposed. Most surprisingly, to me, the Canon's look softer than either of the others. Was there a focus issue? The JVC doesn't look over-processed to me. It looks flat and a little dull. I haven't shot in a setting like that (so it's kinda apples and oranges), but that sort of result isn't typical of the kind of recordings I get. I'm scratching my head a bit over these results.

Which codecs were you using to do the JVC conversion to left/right pairs? I ran across something today that might explain the odd results you were seeing in Vegas (barring a Vegas bug) if you were using the Matrox codecs.

Great. I'll look forward to that.

There's always the low tech way-- A clapboard.

Do you also rely on zoom control being synced between the two TD10's? I think this is a much tougher issue than start time for the clips and you will need to eyeball that for precise match. Then tweak it in the 3D stereoscopic effect in Vegas. If I get into this, I was thinking that I would need to do everything manual anyway but your work with the sync cable and LANC is something I'm watching on your reports.

If I ever do get a second TD10 and bench system I was planning to just use a clapboard. I don't consider it an issue as I used to use that all the time when shooting 4 cameras where only 3 were chase locked to time code. BTW- the new NEXcam will be adding timecode to the recorded file but I don't think it offers chase lock. When the salesman told me that I said so what good is the timecode? Either way Vegas can sync up the clips easily with the audio too.

I am quite surprised and amazed the zoom lenses sync up with Lanc. I think I would have put money on that that it would not work.

Hey, Don and Frank,

Just to make sure you guys have not gone off the deep end using two 3D cameras, can you clarify what is the point of two 3D cameras?

1. I certainly see the value of two 3D cameras on the scene with two operators - two viewpoints (however, my spouse does not take direction (she flunked cooking class in HS because she would not follow the written recipe)).

2. Is it that you can use one camera's 3D for close shots and then for distant shots use the combined left (right) views from the two cameras? This way you effectively have two (actually three) interaxial distances to choose from in post and do not have to shift the spacing of cameras in the middle of a shoot. Or is it something else - videos for spiders?

Mark- Frank has his own agenda. I understand much of it but his logic for importance to some of the available options he can get with his work escapes me.

In my case your #1 is the obvious one and also includes one operator with two TD-10's side by side with one left rolllng on full wide while the second is operated by me going in for closeup shots that I later intercut in the edit.

The second advantage of having two TD-10's is when I come to a location I want to do of wide open scenery and need the wide IA plus the wide angle too. Now I would need to mount both TD-10's on a bench like Franks Aluminum tube and ( in my case) put each in 2D mode and just shoot both like a conventional 2 cam left eye - right eye system. Then do the usual work flow for that system in Vegas with pairing the two files and having the advantage of both.

Frank's work is my lead on this and between what he is doing ( far more than what I plan to do) plus my studies from other doing twin 2D cameras, I hope to develop a very simple to transport and quick to set up twin camera bench for shooting in the public environment. What Frank is doing is more of a science experiment. What I want to do is have a production solution.


I have had thoughts of buying a second 3D cam as the JVC because I do feel Joe's example and Bravia3D work demonstrates an obvious PQ advantage of the JVC over the Sony, but I prefer the tools that Sony offers more, so have decided not to get JVC as a second so I can have the bench system and the #1 two cam shooting angles by owning two TD10's.

I would have moved ahead with those plans but as my luck would have it a buyer for one of my retired broadcast cameras bailed on me Friday because of the stock market crash. ( his excuse) So, my plans to buy a second TD10 and a new editing computer is on hold for now. I may still move forward with building a bench but not sure when I will be buying another TD10.


Hey Frank- Just had a question for you- Have you tried to match up a JVC and a TD10 image for pairing using Sony's 3D calibration? If I build a bench I may try to experiment with the TD 10 and my old SR12. I wonder if that is possible or how difficult it would be.

Why these idiots from sony not implemented SBS recording mode in 1080p60 at 28Mbps?! I think quality will be better tnan interlaced MVC.

Thanks for doing this unique experiment. I wonder if anyone else in the world has thought to do this kind of comparison. Very interesting.

Before we all chime in, I'd be interested in hearing what conclusions you drew from the experiment.

BTW, the best image was of the JVC and Sony from behind at the start of the test. Was that shot with your Fuji W3?

AVS has been down all day because of a hardware failure in Texas. Nice to have it back up again.

I was wondering what the trouble was with AVS. Satelliteguys was also down so they must also use the same server service.

I've been busy with other non 3D projects lately but will download the files Frank and have a look as soon as I can. I will be traveling & shooting3D Friday and Saturday. Editing projects are stacking up fast. I think this will be my last shooting trip for awhile though.

I am quite curious to your results as it will shape my plans for a bench system for the near term.

Yes, I love the W3 for stills. It's native IA is wider than my own eyes. Your shot at the start of the test has me excited to do some 3D Blu-rays from my stills. That hadn't really occurred to me. D'oh!!! (Been too busy with the JVC.)

I came to the same conclusion. The widest IA of left lens JVC and right lens Sony gives the greatest sense of depth for me, too. I'm shocked at how close you were able to get the images to line up. You did this without any adjustments in Vegas?

Frank- Amazing tests. I learned quite a bit!

First the negatives- Please learn to place your graphics text so there are no collisions. You can shrink the titles down and move them off to the side if you don't want to add the 3D effect to place them on top of everything.

Now for the good part- Overall I liked the Sony for overall range out to about 50 ft. The image quality was near identical to the JVC but at distance the Sony did pull the small pilot boat in front of the building just a bit more than the JVC in the Z axis.

The dual cameras regardless of which lens was used performed much better when the subjects were in a range of about 15 ft and beyond. Not surprised here. The first shot of people walking in front of the camera looked good on both the Sony and JVC but was unnerving with the twin cams. So, for close range work, the twin cams just doesn't work.

Using the two left cams the video looked good on distant shots and the pilot boat was well separated from the building behind as well as good separation between the bridge and the Ferry boat? off in the distance. Neither the Sony nor the JVC produced separation between the Bridge and the Ferry Boat. With the dual cameras the street lights were well separated but looked flat with both the JVC and the Sony.

I believe the JVC has a slight edge on the Sony for IA distance but it appears the wider angle lens more than makes up for that giving the Sony a better range of effective 3D especially at distance. Not sure why but that's how I saw it.

Pros and cons of which of the 4 cams to use. In my opinion the only advantage of using #5 JVC left with Sony Right is that gave you the most IA for the distant shots in this particular test setup. However what you sacrificed is your Sony had to be put in 3D mode to achieve a right camera active recording. This sacrifices the optional 2D camera settings and restricts the Sony's 2D options. I would guess that you could achieve the same Z depth range with these cameras by moving the IA out to that of the #5 but use the two cams in 2D mode. In my opinion this would yield the best overall image of distant shots. For close up shots, switch one of the cams to 3D and shoot close in. This is how I had planned to build a bench system. i.e. Use two TD10's and run them both in 2D mode.

I'm very impressed you were able to achieve such accurate alignment between two completely different lens and camera algorithms. I don't know how difficult that was both in the field shoot and also in post to pair the two video files but you did an excellent job. Like I said, my only criticism actually had nothing to do with the purpose of the test. Thank you for pulling this off with such precision.


Other observations- The motion blur was noticeable on the pans for both cams. Did you have image stabilizer on or off for those cameras?