[photo-3d] Re: 3D Camera Synchronisation

["Robert Brand" <rb@xxxxxxxxxx>]

Allan Griffin wrote: Maybe you have cameras selected which
are already within say 2ms but I don't think you have told us that.
Even so, my tests reveal that the end 'synch-spec' wanders or fluctuates
to a degree which I suspect precludes a result as close as you have stated.

Okay this calls for some finer points. I will not go into them all here,
just some most common ones that I have found.

Yes I initially selected cameras that were in sync. That is a given. Any
camera can have a sticky shutter and that is why our club services any old
stereo camera that we organise for members. It is imperitive that mechanical
imperfections and aging problems be eliminated in the selection process.
Without this you do not have any hope of sync'ing any cameras - even
coupling the shutters directly.

The cameras were selected using the method that you showed me many years ago
when I was a dirty nosed youngster to 3D, and it was a lesson that I am
exceedingly thankful for as it without it I would not have produced the
images that I have enjoyed over the years. For the sake of others, I will
detail.

Simply remove the lenses, take the backs off or open the backs fully (no
film of course) and place one camera behind the other in front of a bright
light and use an external sync cable. The settings on the camera must be
manual and starting out with a low shutter speed setting, adjust the speed
from about 1/30th and observe the image through the back of the 2 cameras
while triggering the shutters. The image should be rectangular with the same
proportions as a 35mm slide. The shutter speed is increased until the image
starts to narrow in one plane due to one shutter closing while the other is
still opening. At this point thge sync is the shutter setting +/- about 50%.
In the cameras that I have selected this was fairly reliable at 1/1000th. I
have many helicopter shots and other high speed shots taken with the cameras
with little distortion and no bracketing. Yes I know the real test is when
the blades are pointing towards the camera and the sync errors will be the
greatest and I have plenty of those too.

There are differences that cause sync to suddenly "go out" and that is
simply a matter of the cameras doing other things while you are busy
snapping a shot. One thing a modern camera does is "sniff" the light levels
at regular intervals and make calculations, the results of which are
displayed in the light meter display in the viewfinder of the KR10Ms. If the
metering in the cameras is different, they may take longer or shorter
processing times to make the calculation. Most cameras rely on someones head
blocking the light that may come flooding in through the view finder. More
reliable sync maybe achieved by blocking this light during any shot, but
especially if the sun is behind the camera especially late in the day. I
tend to block the eye piece of my left camera and simply copy the settings
from the right camera. They will match if the cameras are matched. Another
issue to watch for following on from the above is how well the light meters
are matched. If they are busy with slightly differing light inputs they may
also take different times to fire once you fire the shutter. Luckily both my
cameras seem fine.

Using the Allan Griffin method gets you close to the ideal, but there are
these other factors to consider. The method does not guarantee that the sync
will be maintained with age as mechanical differences may come into play but
wow, its a great starting point if you have access to several cameras. If
the cameras are not well sync'ed, but the delay is repeatable, then better
accuracy maybe obtained with a delay circuit, a project I have scheduled
following ISU2001 in Sydney later this year. I hope to see you there!

Regards, Robert Brand


 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/