As I mentioned, the startup I'm involved with is doing a digital camera
of sorts, and to make the system work, we need compression. So we
implemented a compression algorithm (patent pending) in an FPGA chip
that will be 10x lower power than anything on the market. We are not
talking super high performance here, web cam resolution, but having
studied the entire JPEG compression algorithm, it is possible to make a
lossless version, and can be done in a reasonable amount of circuitry.
The Raw format in the E-1 is clearly straight bits with no compression
and there is no reason some raw compression could be done. The
Canon is 12 bits of data, I'm not sure what the E-1 is.
Also, having been and analog CMOS circuit designer since the late 70's,
low noise performance is a delicate art, with lots of trade-offs. CMOS
tends to be higher fundamental noise than a CCD. But with CMOS you
can do noise canceling right at the pixel and buffer it out. In CCD's you
can get better noise/pixel, but then you have to shift the charge out through
multiple stages before you hit the analog circuits that buffer the signal
to the A/D. CCD means charge coupled device. The charge wells
(often called buckets) are created with a voltage on a gate. Then by
clocking the gates high and low, you transfer the charge from one
bucket to the next to the output. That process adds additional noise
as it shifts out. The verdict is still out on which method will win in the
long run. The fact that Canon manufactures their own sensors must
mean they feel they can get the better performance with custom
processing and circuit designs.
I do think it will be possible to get better noise performance out of the
raw circuits, but in the end, it will always be the larger sensor that
captures more light per pixel that will be able to deliver the lowest
signal-to-noise. More signal always means more signal-to-noise.
The advantage here is that CMOS sensors with custom circuits on
a per pixel basis, will have the advantage with large sensors, like
the Canon full frame sensors. CCD's will improve with process
refinements, as the charge transfer efficiency and noise can be
reduced, which probably doesn't work as well with larger sensors
as the charge has to transfer a longer distance before output.
But I'm guessing here.
So the good news is that there is still room for innovation. The fact
that Canon sensor noise still out performs other cameras says
they may have guessed better on technology.
As to dust removal systems, does the E-1 send ultrasonic sound
at the sensor, or shake the sensor with ultrasonic vibration?
With a smaller sensor, will it be easier for Minolta like anti-shake
by shifting the sensor be easier and when will Olympus eventually
get to vibration reduction, or will be the same scenario as auto-focus
was with the OM's? Arriving late in the game? Now if they could
just make anti-shake subjects to photograph.... macro subject motion
reduction... I guess there will always be art to photography, no matter
what the whiz-bang is. I'll always remember the statement by Nubar
Alexanian, photography always involves 3 things, "the subject, the
photographer, and photography." The camera, film/sensor and lens is
only 1/3 of the equation.
Wayne
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|