DV specifications - "effective pixel resolution" ?
David Konieczynski
What is the effective pixel rating on many camcorders?
I am evaluating the differences between the Sony DCR-TRV11 vs. DCR-TRV20.
The Sony manual states the following specifications:
Image Device
1/4 type CCD
DCR-TRV11:
Approx. 680,000 pixels
(Effective: Approx. 340,000 pixels)
DCR-TRV20:
Approx. 1,070,000 pixels
(Effective (still): 1,000,000 pixels)
(Effective (moving): 690,000 pixels)
I understand that the maximum DV resolution for video is approximately
720x480, about 340,000 pixels. Why aren't the "effective pixels" for the
two cameras the same (i.e. 340,000 pixels)? Does the DCR-TRV20 "over
sample," and average multiple pixels to derive a better picture?
Any insights are greatly appreciated.
Thank you,
David
Paul Tauger
"David Konieczynski" <dk...@worldnet.att.net> wrote in message
news:tPYW5.18742$0r2.5...@bgtnsc07-news.ops.worldnet.att.net...
> Dear all -
>
> What is the effective pixel rating on many camcorders?
>
> I am evaluating the differences between the Sony DCR-TRV11 vs. DCR-TRV20.
> The Sony manual states the following specifications:
>
> Image Device
> 1/4 type CCD
>
> DCR-TRV11:
> Approx. 680,000 pixels
> (Effective: Approx. 340,000 pixels)
>
> DCR-TRV20:
> Approx. 1,070,000 pixels
> (Effective (still): 1,000,000 pixels)
> (Effective (moving): 690,000 pixels)
>
> I understand that the maximum DV resolution for video is approximately
> 720x480, about 340,000 pixels. Why aren't the "effective pixels" for the
> two cameras the same (i.e. 340,000 pixels)?
They are. The difference in the number of pixels is relevant only to still
imaging. There are, however, other differences between the two cameras. Do
a deja search on "Lux ratings are meaningless" for a short report on
differences in performance.
John P. Beale
: DCR-TRV11: (Effective: Approx. 340,000 pixels)
: DCR-TRV20: (Effective (moving): 690,000 pixels)
: I understand that the maximum DV resolution for video is approximately
: 720x480, about 340,000 pixels. Why aren't the "effective pixels" for the
: two cameras the same (i.e. 340,000 pixels)? Does the DCR-TRV20 "over
: sample," and average multiple pixels to derive a better picture?
I don't fully understand this myself, but here is one interesting fact:
typical single-chip CCD cameras have a matrix of color filters over the
pixels like this:
R G R G R G R G
G B G B G B G B ...
R G R G R G R G
...
where R=red, G=green, B=blue. The camera forms an intensity signal using
data from each pixel, but gets color information in each channel only from
alternate pixels, and alternate rows for red and blue. They use more
green-sensitive pixels than red or blue because the eye is more sensitive
to green light. (btw. some cameras use complementary color filters.) This
color filter matrix amounts to color sub-sampling of the image. DV video
uses a different color sub-sampling strategy, so increasing number of CCD
pixels probably can improve image quality as recorded as a DV bitmap on the
tape, and it certainly can reduce the chance for unnecessary color
artifacts on hard edges and fine detail. The analog video from the camera
is (I think) not sub-sampled in the DV sense, but of course it is
sub-sampled off the CCD as mentioned above, and the analog chroma signal
also has some finite bandwidth.
3-CCD cameras use entirely separate CCD chips for red, green, and blue so
they can get more accurate color images.
I don't know if this helps anyone, but that's all I know about it anyway.
-john
Paul Tauger
"John P. Beale" <be...@shell3.ba.best.com> wrote in message
news:90jen9$2nvl$1...@nntp1.ba.best.com...> : DCR-TRV11: (Effective: Approx. 340,000 pixels)
> : DCR-TRV20: (Effective (moving): 690,000 pixels)
> : I understand that the maximum DV resolution for video is approximately
> : 720x480, about 340,000 pixels. Why aren't the "effective pixels" for
the
> : two cameras the same (i.e. 340,000 pixels)? Does the DCR-TRV20 "over
> : sample," and average multiple pixels to derive a better picture?
>
> typical single-chip CCD cameras have a matrix of color filters over the
> pixels like this:
>
> R G R G R G R G
> G B G B G B G B ...
> R G R G R G R G
> ...
>
> where R=red, G=green, B=blue. The camera forms an intensity signal using
> data from each pixel, but gets color information in each channel only from
> alternate pixels, and alternate rows for red and blue. They use more
> green-sensitive pixels than red or blue because the eye is more sensitive
> to green light. (btw. some cameras use complementary color filters.) This
> color filter matrix amounts to color sub-sampling of the image. DV video
> uses a different color sub-sampling strategy, so increasing number of CCD
> pixels probably can improve image quality as recorded as a DV bitmap on
the
> tape, and it certainly can reduce the chance for unnecessary color
> artifacts on hard edges and fine detail. The analog video from the camera
> is (I think) not sub-sampled in the DV sense, but of course it is
> sub-sampled off the CCD as mentioned above, and the analog chroma signal
> also has some finite bandwidth.
Hmmm. This is interesting. I've already reported on the differences in
color saturation and low-light performance between the TRV-11 and TRV-20. I
wonder if this is attributable to the sub-sampling you've described. This
would mean that, though motion video resolution isn't enhanced by the
additional pixels, image quality is, nonetheless, improved.
It would certainly explain what I've observed empirically: the TRV-11 and
TRV-20 are not the same camera, and differ in other ways than just the still
imaging resolution.
Howard Ring
The statement for the TRV11 indicates that 680,000 pixels are used
to generate a stabilized image of 340,000 pixels. The statement for
the TRV20 indicates that 1,000,000 pixels are used for still picture
mode, while 690,000 pixels are used for video mode, and leaves unsaid
the 690,000 to 340,000 reduction of electronic image stabilization.
Just a guess mind you.
Howard
David Konieczynski
As an update, I posed my questions to Sony's technical sales department
yesterday. After holding on the phone for more than 20 minutes, "John"
offered only that some image resolution was lost to electronic image
stabilization (despite the camcorders' optical image stabilization). He was
unable to comment on any differences between the video quality of the -11
versus -20.
I purchased the -11 from Circuit City with a 30-day no-questions-asked,
no-restocking-fee return policy. After one week's use, I have come to
realize that the camcorder does suffer significant image degradation with
"orange-yellow" casting and notable graininess when shooting indoors with
artificial lighting. My test video was of my 10-month daughter crawling on
a white carpet at 6am (dark outside), with all room lights on. The same
subject / room at 12 pm (with ambient light coming through windows), yielded
excellent color rendition and clarity.
Based on the comments of John and Paul, I have decided to return the -11 and
try the -20. I'll report on my experiences in a few weeks.
Happy holidays,
David Konieczynski
Howard Ring
(DCR-TRV20) uses optical image stabilization. It claims to use
Super SteadyShot, but I cannot find a good (more than two
sentence) explanation of how it works. The information on the
DCR-TRV900 claims to use Optical Super SteadyShot however.
Howard
Paul Tauger
"Howard Ring" <ri...@NOSPAmediaone.net> wrote in message
news:3A2FDCAA...@NOSPAmediaone.net...
The TRV-20 uses electronic stablization, whereas the TRV-900 uses optical
stablization. The latter uses mechanical stabilizers to move the lens
around, whereas the former keeps the imaged centered by analyzing the image
on the CCD, and then using that portion of the CCD's image which corresponds
to the "centered" image.
>
> Howard