Originally Posted by Frank
In case anyone is interested I uploaded a test video I shot a couple of years ago when I was playing around with high shutter speeds.
This was shot with a shutter speed of 1/1000th.
In the first part the sync was off by about a millisecond. After that it gets closer.Fast shutter speed 3D video
It's in frame compatible side by side format as an .m2ts file.
Frank, shots of timd wildlife in 3D are indeed fascinating to watch. Normally the best we could achieve in real life with timid wildlife would be binoculars; a poor substitute for the close-up shots you have captured!
The start of Frank's file is indeed horribly out of synch and I found that part very hard to watch for that reason! I see this type of effect to a much smaller degree when using shutter glasses particularly if they are operating at only 100Hz for a 25fps source. The Left eye is exposed to a new frame 10mS before the Right eye.
I also see it very occasionally at the cinema where the Left eye is exposed to the new frame 6.94mS before the Right eye (144Hz RealD). Of course in those situations the captured content is correct but just presented to the two eyes out of phase.
I think I'm a bit more conscious of flicker than the average person. At the squash court I see the ball stroboscopically moving under the 100Hz illumination of conventional fluorescent tubes still in common use in Australian squash courts. I ask others whether they can see this and they look at me with a puzzled expresssion...
I see very noticeable flicker on my 2010 model 50" Panasonic Plasma, and others either don't see it at all, or can only just notice it if asked to look out for it.
I found it interesting to view the start of Frank's mt2s file frame by frame with de-interlacing turned off
. This allowed me to see the extent of movement in the 16.83mS between an odd field and an even field. The Left stereoscopic view leads the Right, but by how much?
At the beginning when a blu-jay flies in from the right to perch on the food bucket, the width of interlacing comb pattern artefacts (or "mouse teeth") of the bird in the Left view [and in the Right view] is about two-thirds of the width of the blu-jay's head.
However the position of the blu-jay in the Right view is displaced to the right (compared with the Left view) at a distance more than a small fraction of the width of the blu-jay's head. Even without a knowledge of the camera angles Frank used, I suspect that the Right view may have been lagging a bit more than 1mS at that point in the video. (The point I examined was 36 frames into the clip.) If the measuring device Frank was using showed 1mS, I think it may have been somehow underindicating the disparity of the capture process.
However the point is well made that this type of scene really demands a high frame rate, a short exposure time, and the best possible synchronisation. The blu-jays move their beaks, and wings, very rapidly. A timing difference of half a frame, which can arise with asynchronous cameras, would ruin such a video. I suggest however that for slow dramatic photography, even half a frame of capture mistiming, might be hard to notice.
By the way, it's remarkable how well deinterlacing can work with this demanding type of natural scene. If you pause at the 36 frames point in this file, then, depending on the deinterlacing algorithm in use, you may find that a faint second image of the bird's head is visible.
Originally Posted by Don Landis
Here's a new video I am working on using mostly ultra wide stereobase with ranges from 6" IA to 28 inches IA.
Don, if I may make a few comnents on this unofficial first cut. First, I have to agree with Joseph Clarke's favourable comment on the pacing. And there is the evocative music.
However, my eyesight is not as accommodating as others' for strong 3D, and at times I found the 3D effect a bit overpowering. (My partner is comfortable with a more pronounced 3D effect than I am!) Scenes I personally find challenging to watch are where the main subject is in the far distance but where there is also some fairly close image content at the left and right edges of the frame. Cheers.