RE: Videao cameras and quality of image

[Tim Sullivan x5830 <sullivan@kenyon.edu>]

Ludwik Kowalski quotes me:

> I'm not an expert on Video Point, but we do a lot of video at Kenyon (and
> I am co-author of PhysVis, Kenyon's freeware program getting kinematic data 
> from captured video clips, so have some expertise in digital video.) 

> I can't think how the problem can be ascribed to Video Point itself, rather 
> the problem may be the current state of video. VHS tape is the worst (about
> 200x512 resolution). SVHS is better at about 400x512. Direct camera out is 
> even better (limited by the CCD pixel resolution and quality of optics).

And then writes:

Is it true then that by recording a falling object with the VHS-camera tilted 
by 90 degrees one gets much better vertical resolution (512) than from the
same nontilted camera?

I respond:

The 512 figure in quotes of TV resolution refer to the number of scan lines in
a standard NTSC signal. So this resolution is the resolution in the vertical
direction, since the scan lines run horizontally. So rotating the camera by 90
degrees would lose you resolution for vertical motion according to these
considerations. I qulaify this opinion with two comments. First, I have not
made a systematic study of this. Second, I am unsure whether the 512 lines
represent "real" resolution. It certainly represents an upper limit, but I
can imagine systems that wouldn't achieve this limit. 


Ludwik goes on to ask:

Are you saying that it is better to capture (digitize) clips directly than
storing them on tape and digitizing from tape? Do you know the typical CCD
pixel resolution of an old camcorder, for example, a large HVS Magnavox 
manufactured 5 years ago, or so. How does this compare with today's small
camcorders?

I respond:

I don't have any information on older camcorders vs newer ones in terms of
resolution. (CCD chips are now standard and they are superior to their 
predecessors for this application in having no warm up variations in "zoom", 
among other things. CCDs have been standard for awhile, so I doubt a camera
manufactured 5 years ago wouldn't use CCDs anyway.)

Depending on camera design, you are almost surely better off capturing directly
from the camera signal than from tape in terms of resolution. VHS specs at only
200 "pixels" across the width of the screen. SVHS is supposed to be about 400
and even that is not considered to be "broadcast quality", i.e., suitable for
broadcast by a TV network. So good camera output is superior to the tape media
available for the home market. But that leaves room for makers of home
camcorders to make some choices about the size and resolution of the CCD chips
they use in their units. So whether a particular camera output is superior to
VHS, I can't really say.


But having opined so much about resolution, it is unclear how much of an impact
it has in practice in the physics lab. Most objects I can think of are larger
than one pixel and the eye can do a pretty good job of finding the centroid of
the object to mark. If you do curve fitting on the results, then you further
smooth any fluctuations due to pixellation. We found that if students use the 
mouse to mark points there is a resulting jitter of more than one pixel. (To
reduce this, we added the ability to use arrow keys to more precisely position
the cursor to our PhysVis program.) So it isn't clear how much money and time
are justified to improve resolution given these other considerations.

Tim Sullivan
sullivan@kenyon.edu