Sigma SD10 Web Site Update

"Steven M. Scharf" wrote in message
news
Steven, good website, but you've got an error. In the section on 3-CCD
video cameras, you state that they do not have as good low-light performance
as Bayer 1-CCD video cameras because of light loss from the prism. This
isn't correct. A simple thought-picture shows why:

Assume a lens of perfect efficiency, as well as Bayer photodetector filters
and dichroic prisms of perfect efficiency. Assume the 1-CCD camera has a
single 1/2" sensor, and the 3-CCD camera has three 1/2" sensors. Assume,
also, that, unlike true Bayer, the 1-CCD camera has an equal number of R,G
and B photodetectors.

The 1-CCD camera sensor will receive 1/3 of the total light falling on it at
each photodetector, and will discard 2/3 of the total -- a red photodetector
receives 1/3 of the white light and discards the 2/3s that comprise blue and
green, a green photodetector receives 1/3 of the white light light and
discards the 2/3s that comprise red and blue, etc.. .

The 3-CCD camera doesn't depend on filtration, but on splitting white light
up into its component parts. Accordingly, no colors are discarded -- all of
the red light is directed to the red sensor, all of the green light to the
green sensor, etc. Accordingly, assuming perfect efficiency, a 3-CCD
camcorder's sensors receive 3 times more light than a 1-CCD camcorder.

In practice, this is borne out. It is true that current 1-CCD consumer
camcorders tend to have smaller sensors than 3-CCD machines, and smaller
sensors mean poorer low-light performance. However, 3-CCD camcorders, such
as Sony's VX2100 or Canon's XL1 perform far, far better under low-light
conditions than the older single CCD camcorders which had sensors the same
size, or even larger.


The Sigma SD10 Information Site has been updated. http://sigmasd10.com is
the authoritative site for unbiased information about the pros and cons of
the Sigma SD10 digital SLR.

This was a minor update.

1. A little more information was added on the cause of the color rendition
issues with the Foveon sensor.

2. Changed the 27 links of the definition of a pixel into numbered links,
rather than the entire URL.

3. Even though the R-CRV3 Li-Ion option isn't on Sigma's web site yet, I
went ahead and changed "No Poor Rechargeable Li-Ion Battery Option," to
"Poor Rechargeable Li-Ion Battery Option."

4. Added links to other reviews of the Sigma SD10.

If you find any errors or omissions on this site, please let me know. I
have
made several changes as a result of feedback from rec.photo.digital
contributors.

Disclaimers
------------
This is a non-commercial, informational site. The opinions stated on this
site are the opinions of the author, and of other contributors.

Nothing is sold on this site. No advertising is accepted.

This site is not affiliated with Sigma Corporation or Foveon Corporation.

"Steven M. Scharf" wrote in message
news
Steven, good website, but you've got an error. In the section on 3-CCD
video cameras, you state that they do not have as good low-light performance
as Bayer 1-CCD video cameras because of light loss from the prism. This
isn't correct. A simple thought-picture shows why:

Assume a lens of perfect efficiency, as well as Bayer photodetector filters
and dichroic prisms of perfect efficiency. Assume the 1-CCD camera has a
single 1/2" sensor, and the 3-CCD camera has three 1/2" sensors. Assume,
also, that, unlike true Bayer, the 1-CCD camera has an equal number of R,G
and B photodetectors.

The 1-CCD camera sensor will receive 1/3 of the total light falling on it at
each photodetector, and will discard 2/3 of the total -- a red photodetector
receives 1/3 of the white light and discards the 2/3s that comprise blue and
green, a green photodetector receives 1/3 of the white light light and
discards the 2/3s that comprise red and blue, etc.. .

The 3-CCD camera doesn't depend on filtration, but on splitting white light
up into its component parts. Accordingly, no colors are discarded -- all of
the red light is directed to the red sensor, all of the green light to the
green sensor, etc. Accordingly, assuming perfect efficiency, a 3-CCD
camcorder's sensors receive 3 times more light than a 1-CCD camcorder.

In practice, this is borne out. It is true that current 1-CCD consumer
camcorders tend to have smaller sensors than 3-CCD machines, and smaller
sensors mean poorer low-light performance. However, 3-CCD camcorders, such
as Sony's VX2100 or Canon's XL1 perform far, far better under low-light
conditions than the older single CCD camcorders which had sensors the same
size, or even larger.


The Sigma SD10 Information Site has been updated. http://sigmasd10.com is
the authoritative site for unbiased information about the pros and cons of
the Sigma SD10 digital SLR.

This was a minor update.

1. A little more information was added on the cause of the color rendition
issues with the Foveon sensor.

2. Changed the 27 links of the definition of a pixel into numbered links,
rather than the entire URL.

3. Even though the R-CRV3 Li-Ion option isn't on Sigma's web site yet, I
went ahead and changed "No Poor Rechargeable Li-Ion Battery Option," to
"Poor Rechargeable Li-Ion Battery Option."

4. Added links to other reviews of the Sigma SD10.

If you find any errors or omissions on this site, please let me know. I
have
made several changes as a result of feedback from rec.photo.digital
contributors.

Disclaimers
------------
This is a non-commercial, informational site. The opinions stated on this
site are the opinions of the author, and of other contributors.

Nothing is sold on this site. No advertising is accepted.

This site is not affiliated with Sigma Corporation or Foveon Corporation.

On 9 Jul 2004 19:33:56 -0700, (Georgette
Preddy) wrote:



He seems to have forgot to post this controlled comparison, wher the
SD9 shows more than double the color resolution of the 10D, the SD9 is
slight higher resolution than the 10D with the Sigma shooting in
medium resolution...
http://www.pbase.com/canon10d_sigmasd9

Did you steal those pics to Steve

On 9 Jul 2004 19:33:56 -0700, (Georgette
Preddy) wrote:



He seems to have forgot to post this controlled comparison, wher the
SD9 shows more than double the color resolution of the 10D, the SD9 is
slight higher resolution than the 10D with the Sigma shooting in
medium resolution...
http://www.pbase.com/canon10d_sigmasd9

Did you steal those pics to Steve

In article , bagal
wrote:

No, the reason Foveon is creating shock waves is because maintaining a one
to one correspondence seems to minimise artefacting and degradation.

The only shock waves they're causing is by the poor souls who were
duped into buying them and then seeing the crappy skin tones and
colors.

In article , bagal
wrote:

No, the reason Foveon is creating shock waves is because maintaining a one
to one correspondence seems to minimise artefacting and degradation.

The only shock waves they're causing is by the poor souls who were
duped into buying them and then seeing the crappy skin tones and
colors.

"Zebedee" wrote in message news:40eee231$0$29050
I do like the Foveon sensor but that's the only thing I like about this
camera.

The theory behind the Foveon sensor is fine. Having three sensors per pixel
is an excellent way to improve color resolution, and such a system was used
successfully on the initial Foveon camera, and is used on some mid-range to
higher end camcorders (three sensors, and a prism). But the camcorders do a
smart thing, they shift pixels to increase the spatial resolution.

The problem with the Foveon sensor is that silicon color separation is a
very inexact science. This necessitates the extensive post-processing from
RAW data, but even this is not sufficient, which is why there are all the
complaints about the color problems.

The Bayer type of sensor uses a much better method of color separation.

I dislike tremendously the tri-sensor arrays of the current digital
cameras because of the way they can cause moire problems and because
all images are interpolated and are therefore inherantly unsharp. What
we get (crudely) when we buy a 3mp camera is a 1mp camera with
interpolation.

Fortunately, the human eye does just fine with interpolated color, it's the
spatial information that is most important. And of course the color of each
pixel from a Foveon sensor is also calculated and corrected.

Technically you are incorrect. With a 3 megapixel camera, you get a 3
megapixel camera, with the color of each pixel determined by the eight
surrounding pixels (as well as the color of the pixel itself). You actually
have 0.75M red, 0.75 blue, and 1.5M green pixels. With a Foveon 3.43
megapixel sensor, you have 3.43 megapixels, each comprised of there
photodetectors.

Read the press release, Foveon press release
(http://foveon.com/press_sigma.html): "The Foveon X3 image sensor in the
Sigma SD9 contains over 10.2 million color photodetectors, which are
organized in 3 layers within the sensor to form 3.54 million full-color
pixels."

"Zebedee" wrote in message news:40eee231$0$29050
I do like the Foveon sensor but that's the only thing I like about this
camera.

The theory behind the Foveon sensor is fine. Having three sensors per pixel
is an excellent way to improve color resolution, and such a system was used
successfully on the initial Foveon camera, and is used on some mid-range to
higher end camcorders (three sensors, and a prism). But the camcorders do a
smart thing, they shift pixels to increase the spatial resolution.

The problem with the Foveon sensor is that silicon color separation is a
very inexact science. This necessitates the extensive post-processing from
RAW data, but even this is not sufficient, which is why there are all the
complaints about the color problems.

The Bayer type of sensor uses a much better method of color separation.

I dislike tremendously the tri-sensor arrays of the current digital
cameras because of the way they can cause moire problems and because
all images are interpolated and are therefore inherantly unsharp. What
we get (crudely) when we buy a 3mp camera is a 1mp camera with
interpolation.

Fortunately, the human eye does just fine with interpolated color, it's the
spatial information that is most important. And of course the color of each
pixel from a Foveon sensor is also calculated and corrected.

Technically you are incorrect. With a 3 megapixel camera, you get a 3
megapixel camera, with the color of each pixel determined by the eight
surrounding pixels (as well as the color of the pixel itself). You actually
have 0.75M red, 0.75 blue, and 1.5M green pixels. With a Foveon 3.43
megapixel sensor, you have 3.43 megapixels, each comprised of there
photodetectors.

Read the press release, Foveon press release
(http://foveon.com/press_sigma.html): "The Foveon X3 image sensor in the
Sigma SD9 contains over 10.2 million color photodetectors, which are
organized in 3 layers within the sensor to form 3.54 million full-color
pixels."

"PTRAVEL" wrote in message
m...

"Steven M. Scharf" wrote in message
news
Steven, good website, but you've got an error. In the section on 3-CCD
video cameras, you state that they do not have as good low-light
performance
as Bayer 1-CCD video cameras because of light loss from the prism. This
isn't correct. A simple thought-picture shows why:

snip

I've just been shopping for camcorders, and I was surprised to read that the
3-CCD models had poorer low-light performance than the single CCD models--it
seems counter-intuitive until you realize that the absorbtion losses through
the prism are non-trivial. Foveon had the same issue in their original
camera, but it was a studio camera that was not used with poor lighting, so
it wasn't a big issue.

I was comparing 1-CCD with 3-CCD within the same price range. When you get
to the higher end 3-CCD models, with the larger sensors, it is indeed true
that their low-light performance is better.

I was looking at models in the $600 price range. The reviews of the
Panasonic GS70/GS120, strongly criticized the low light performance, while
the reviews of the Canon Optura Xi, stated that the low light performance
was good (not great). Plus, the Optura Xi has optical image stabilization,
which is far better than digital image stabilization. I wasn't looking at
the level of the Canon XL1, or the other prosumer models. There are two new
3-CCD consumer models coming from Sony and Pansonic, with larger sensors,
that probably will have as good low light performance as something like the
Optura Xi, but they will be much more expensive.

"PTRAVEL" wrote in message
m...

"Steven M. Scharf" wrote in message
news
Steven, good website, but you've got an error. In the section on 3-CCD
video cameras, you state that they do not have as good low-light
performance
as Bayer 1-CCD video cameras because of light loss from the prism. This
isn't correct. A simple thought-picture shows why:

snip

I've just been shopping for camcorders, and I was surprised to read that the
3-CCD models had poorer low-light performance than the single CCD models--it
seems counter-intuitive until you realize that the absorbtion losses through
the prism are non-trivial. Foveon had the same issue in their original
camera, but it was a studio camera that was not used with poor lighting, so
it wasn't a big issue.

I was comparing 1-CCD with 3-CCD within the same price range. When you get
to the higher end 3-CCD models, with the larger sensors, it is indeed true
that their low-light performance is better.

I was looking at models in the $600 price range. The reviews of the
Panasonic GS70/GS120, strongly criticized the low light performance, while
the reviews of the Canon Optura Xi, stated that the low light performance
was good (not great). Plus, the Optura Xi has optical image stabilization,
which is far better than digital image stabilization. I wasn't looking at
the level of the Canon XL1, or the other prosumer models. There are two new
3-CCD consumer models coming from Sony and Pansonic, with larger sensors,
that probably will have as good low light performance as something like the
Optura Xi, but they will be much more expensive.