Posted by Smarty on 11/25/06 16:45

Martin,

It's quite possible that the low light sensitivity was enhanced with longer
exposure durations, but the real problem is that the sensor retains too much
"memory" of the recent past. The decay time of the image in the sensor's
storage elements (which are charge coupled / capacitive elements) is the
time it takes for the most recently captured image to dissipate and become
replaced with the next successive image. In a phosphor based display, this
same effect is referred to as "persistence", but it is most often called
"lag" in a sensor which does optical imaging like a CCD or a vidicon. The
visual result is streaking, tails, and blurs of objects which are moving. I
find this effect especially annoying. It is easy (comparatively) to make a
sensor capture more light by integrating over a longer sampling interval.
This is a trick which astronomers, photographers, and low light physicists
have used to capture very faint images, and camcorder designers are using
the same approach. The real trick is to make the sensor dump the captured
image instantly when the next one arrives with no retention of the past
frame / field. In charge coupled (C-C-D) devices, this is easier said than
done. Capacitors don't do ***instantly*** very well, since they are
inherently a storage device with a non-zero time constant.

As both you and PTravel observed, a bigger sensor is the better answer. It
is again basic physics that these will capture more light, all things being
equal. Faster lenses and quiet electronics also help, as would refrigeration
to reduce "shot noise" and thermal effects which make random electronics
move more vigorously. In still cameras, an effective technique is to briefly
close the shutter, take a picture of the sensor's instant noise background,
and then subtract it off of the prior frame, effectively cancelling most of
the noise to gain a higher effective sensitivity. For whatever reasons
(probably time budget limits in a moving frame stream), the camcorder
designers have not adopted this same strategy.

To their credit, Sony, Canon, and others have aggressively applied digital
signal processing to the captured video to extract a better signal from the
noise. This is a computationally hungry and expensive solution both in power
consumption and in chip count. Thus, better lens, sensor, and image
processing pushes up the price tag a lot.

Which is the complex way of explaining my prior comment..........you can
pick any 2 of the 3 you want: light sensitivity, clear moving images, or
cost......... (-8


Smarty











"Martin Heffels" <twentyfourthof@november.edu> wrote in message
news:hfqgm2lh3fr614ohnofqfb7ku7s1bugdji@4ax.com...
> On Sat, 25 Nov 2006 09:16:42 -0500, "Smarty" <nobody@nobody.com> wrote:
>
>>It did exactly as John published, capturing low light images
>>very well, but the motion artifacts were just awful. This defect was not
>>even slightly evident from his comparison pictures.
>
> Wouldn't that be slow-shutter, to improve sensitivity?
>
>>I guess I should not be especially surprised that Sony and others have
>>made
>>light sensitivity, clean motion images, and low price the basic 3
>>discriminants for the consumer. You can pick any 2.....but you can't get
>>all
>>3 at once!
>
> LOL. Very true.
>
> -m-
> --

[Back to original message]