Sensors

Michael Johnson, PE <cds DeleteThis @erols.com> wrote:
> measekite wrote:
>> I have read a little on the differences between a CMOS sensor and a CCD
>> sensor. The article I read is that a CCD sensor costs more to make.
>>
>> From a practical user standpoint, is one significantly better than the
>> others? Anybody know the advantages of one over the other (advantages
>> that are meaningful)?
>
> Others will give you a more detailed account of the differences than I
> can but from a user standpoint perhaps the biggest advantage the CMOS
> sensor has over a CCD sensor is lower noise levels at high ISO settings.
> CMOS sensors tend to run hotter and are not as good, or practical, for
> shooting video like the CCD based cameras. CCD sensors are smaller and,
> on a per unit basis, are typically cheaper to manufacture than CMOS
> sensors which is why most P&S cameras use them.
>

Hmm. Seems not quite right to me. CMOS is cheaper to manufacture than CCD.
I am not sure about the operating temperature differences, but I suspect CMOS
runs cooler [which would explain the reputed reduced noise levels at high
ISO]. Chip size has nothing to do with it (you mentioned the CCD is smaller
.... not true). CCD is an older technology, and it was incorporated into
video cameras early on, but I am not aware of any reasons why a CMOS sensor
could not be used [in fact, I bet Canon does use them in their cameras].

--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
 >> Stay informed about: Sensors 

Thomas T. Veldhouse wrote:
> Michael Johnson, PE <cds RemoveThis @erols.com> wrote:
>> measekite wrote:
>>> I have read a little on the differences between a CMOS sensor and a CCD
>>> sensor. The article I read is that a CCD sensor costs more to make.
>>>
>>> From a practical user standpoint, is one significantly better than the
>>> others? Anybody know the advantages of one over the other (advantages
>>> that are meaningful)?
>> Others will give you a more detailed account of the differences than I
>> can but from a user standpoint perhaps the biggest advantage the CMOS
>> sensor has over a CCD sensor is lower noise levels at high ISO settings.
>> CMOS sensors tend to run hotter and are not as good, or practical, for
>> shooting video like the CCD based cameras. CCD sensors are smaller and,
>> on a per unit basis, are typically cheaper to manufacture than CMOS
>> sensors which is why most P&S cameras use them.
>>
>
> Hmm. Seems not quite right to me. CMOS is cheaper to manufacture than CCD.
> I am not sure about the operating temperature differences, but I suspect CMOS
> runs cooler [which would explain the reputed reduced noise levels at high
> ISO]. Chip size has nothing to do with it (you mentioned the CCD is smaller
> ... not true). CCD is an older technology, and it was incorporated into
> video cameras early on, but I am not aware of any reasons why a CMOS sensor
> could not be used [in fact, I bet Canon does use them in their cameras].

CMOS may be cheaper to manufacture on an area basis but where CCD
catches up is they are smaller more fit on a wafer (or the photo chip
equivalent) so there is better economy of scale. Maybe I am wrong.
Also, I thought the reason CMOS sensors aren't ideal for video is they
can't do the rapid frame rate for extended time periods needed for
smooth video without overheating.
 >> Stay informed about: Sensors 

On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
<cds DeleteThis @erols.com> wrote:

>CCD verses CMOS sensor size in a DSLR has more to do with lens design
>than in a P&S. There is full frame or, more commonly, the 1.6x crop
>size. Whether CCD or CMOS the physical size of the sensor has to be
>more consistent to conform to the lens properties across differing
>camera models while maintaining future compatibility.

Are you sure?
Then why do EF lenses work on both APS-C sensors AND full frame
sensors?
And, on Nikon cameras, why do their non-Digital (I don't know the
appelation) work on both their APS sensor cameras and on their (old)
full frame DSLRs and in both cases (Canon and Nikon) these leneses
also work on the (obviously) full frame film cameras?
>P&S cameras have
>dedicated unchangeable lenses which allow CCD sensors to be smaller
>without any of these concerns. The sensor size in a Sony DSC-W100
>(8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>crop DSLR is roughly 338 square mm no matter what the megapixel rating.
> Out of this same area you can get 8-9 CCD type sensors.

Well, of course you can.
But those sensors won't fit in a camera that uses an APS-size sensor,
like a Nikon D70s.
Using he same size (or approximately, as with nikon<>Canon, the
sensors will be the same size, whether CCD or CMOS.
>It is a
>similar effect that AMD and Intel sees from going to a 65nm process from
>a 90nm one. They get more product from the same wafer size which means
>higher yields/profits and/or lower prices per unit.

But that's not the reason they do this - it's to get faster speed
while keeping heat down. The higher yield is a happy byproduct.
--
Bill Funk
replace "g" with "a"
 >> Stay informed about: Sensors 

Michael Johnson, PE <cds.TakeThisOut@erols.com> wrote:
>
> CMOS may be cheaper to manufacture on an area basis but where CCD
> catches up is they are smaller more fit on a wafer (or the photo chip
> equivalent) so there is better economy of scale. Maybe I am wrong.
> Also, I thought the reason CMOS sensors aren't ideal for video is they
> can't do the rapid frame rate for extended time periods needed for
> smooth video without overheating.

Really? Which is bigger, the CMOS 6.3MP sensor that the Canon Digital Rebel
uses or the CCD 6.1MP sensor the Nikon D50/D70 and D70s use (among the Nikon
branded cameras ... many others use that sensor).

There are plenty of CMOS based video cameras out there. This one comes to
mind with a simple Google search.

http://www.globalsourcesdirect.com/servlet/the-250/Digital-Video-Camera/Detail

CCD, when properly cooled, can achieve higher image quality than the CMOS
counter part, and is why CCD is still used for astronomy; however, I am not
sure that there are many DSLRs around with portable liquid cooling systems in
them ...

--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
 >> Stay informed about: Sensors 

Thomas T. Veldhouse wrote:
> Michael Johnson, PE <cds DeleteThis @erols.com> wrote:
>> CMOS may be cheaper to manufacture on an area basis but where CCD
>> catches up is they are smaller more fit on a wafer (or the photo chip
>> equivalent) so there is better economy of scale. Maybe I am wrong.
>> Also, I thought the reason CMOS sensors aren't ideal for video is they
>> can't do the rapid frame rate for extended time periods needed for
>> smooth video without overheating.
>
> Really? Which is bigger, the CMOS 6.3MP sensor that the Canon Digital Rebel
> uses or the CCD 6.1MP sensor the Nikon D50/D70 and D70s use (among the Nikon
> branded cameras ... many others use that sensor).
>
> There are plenty of CMOS based video cameras out there. This one comes to
> mind with a simple Google search.
>
> http://www.globalsourcesdirect.com/servlet/the-250/Digital-Video-Camera/Detail
>
> CCD, when properly cooled, can achieve higher image quality than the CMOS
> counter part, and is why CCD is still used for astronomy; however, I am not
> sure that there are many DSLRs around with portable liquid cooling systems in
> them ...

CCD verses CMOS sensor size in a DSLR has more to do with lens design
than in a P&S. There is full frame or, more commonly, the 1.6x crop
size. Whether CCD or CMOS the physical size of the sensor has to be
more consistent to conform to the lens properties across differing
camera models while maintaining future compatibility. P&S cameras have
dedicated unchangeable lenses which allow CCD sensors to be smaller
without any of these concerns. The sensor size in a Sony DSC-W100
(8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
crop DSLR is roughly 338 square mm no matter what the megapixel rating.
Out of this same area you can get 8-9 CCD type sensors. It is a
similar effect that AMD and Intel sees from going to a 65nm process from
a 90nm one. They get more product from the same wafer size which means
higher yields/profits and/or lower prices per unit.

As for the original poster's question, P&S camera's typically use CCD
sensors. I have seen very low quality (and low mp count) P&S cameras
use CMOS sensors. The camera you referenced is one of these since I
could buy five of them for $550. I would also suspect the quality of
the video to be low resolution. If CMOS were the better choice for the
manufacturers of P&S cameras then I would expect them to be the norm and
not the rare exception. CMOS does appear to be making inroads in the
camcorder market. Personally, I think the camcorder will merge with the
P&S camera in the coming years and if this happens the CCD sensor may be
the ultimate winner.

As for the end user's considerations, CCD based cameras tend to have
higher noise levels. This is changing somewhat with the newer crop of
higher ISO, lower noise P&S cameras. P&S cameras using a CMOS sensor
are typically low end with low megapixel counts. Maybe this is because
a CCD can pack more pixel sites in the same area compared to a CMOS
chip. This promotes the design of smaller cameras while retaining high
resolution which seems to be what the P&S manufacturers have been
striving for the past several years. There are high end DSLR's using
CCD sensors. However, I don't know of any high end P&S cameras that use
CMOS sensors. Take what you will from this observation. Finally, I am
basing the heat issue of CMOS compared to CCD sensors on what I have
read here in this newsgroup. Maybe what I read was incorrect. IMO,
this is a minor point of the discussion at hand.
 >> Stay informed about: Sensors 

Bill Funk wrote:
> On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
> <cds RemoveThis @erols.com> wrote:
>
>> CCD verses CMOS sensor size in a DSLR has more to do with lens design
>> than in a P&S. There is full frame or, more commonly, the 1.6x crop
>> size. Whether CCD or CMOS the physical size of the sensor has to be
>> more consistent to conform to the lens properties across differing
>> camera models while maintaining future compatibility.
>
> Are you sure?
> Then why do EF lenses work on both APS-C sensors AND full frame
> sensors?
> And, on Nikon cameras, why do their non-Digital (I don't know the
> appelation) work on both their APS sensor cameras and on their (old)
> full frame DSLRs and in both cases (Canon and Nikon) these leneses
> also work on the (obviously) full frame film cameras?

Sensor size is roughly the same as film size for DSLR cameras. If you
fit a 35mm film camera with a 24mm film roll what would happen? There
would roughly be a 1.6x crop factor like there is with APS sensor DSLR.
Nikon or Canon could come out with a 3x crop factor DSLR but why?
Since lenses are changed on DSLRs having a sensor standard is more
important. It makes lenses operate the same on a 10D, 300D, 20D, 30D,
350D 400D etc. This isn't a concern on P&S cameras since the lens stays
with the camera and can't be changed. I'm not sure what your point is here.

>> P&S cameras have
>> dedicated unchangeable lenses which allow CCD sensors to be smaller
>> without any of these concerns. The sensor size in a Sony DSC-W100
>> (8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>> crop DSLR is roughly 338 square mm no matter what the megapixel rating.
>> Out of this same area you can get 8-9 CCD type sensors.
>
> Well, of course you can.
> But those sensors won't fit in a camera that uses an APS-size sensor,
> like a Nikon D70s.

Sure they will. You'll just have a higher crop factor. This isn't
really practical for the reasons I gave above.

> Using he same size (or approximately, as with nikon<>Canon, the
> sensors will be the same size, whether CCD or CMOS.

Once again the sensor size on a DSLR is the roughly the same for
consistency between models and lenses.

>> It is a
>> similar effect that AMD and Intel sees from going to a 65nm process from
>> a 90nm one. They get more product from the same wafer size which means
>> higher yields/profits and/or lower prices per unit.
>
> But that's not the reason they do this - it's to get faster speed
> while keeping heat down. The higher yield is a happy byproduct.

I doubt the higher yield, and lower resulting cost per sensor, is just a
byproduct but one of the main considerations in using CCD sensors in P&S
cameras. This also allows the cameras to be smaller while retaining
high resolution which seals the deal for the overwhelming majority of
current P&S cameras to use CCD sensors, IMHO, of course.
 >> Stay informed about: Sensors 

On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
<cds.RemoveThis@erols.com> wrote:

>Thomas T. Veldhouse wrote:
>> Michael Johnson, PE <cds.RemoveThis@erols.com> wrote:
>>> CMOS may be cheaper to manufacture on an area basis but where CCD
>>> catches up is they are smaller more fit on a wafer (or the photo chip
>>> equivalent) so there is better economy of scale. Maybe I am wrong.
>>> Also, I thought the reason CMOS sensors aren't ideal for video is they
>>> can't do the rapid frame rate for extended time periods needed for
>>> smooth video without overheating.
>>
>> Really? Which is bigger, the CMOS 6.3MP sensor that the Canon Digital Rebel
>> uses or the CCD 6.1MP sensor the Nikon D50/D70 and D70s use (among the Nikon
>> branded cameras ... many others use that sensor).
>>
>> There are plenty of CMOS based video cameras out there. This one comes to
>> mind with a simple Google search.
>>
>> http://www.globalsourcesdirect.com/servlet/the-250/Digital-Video-Camera/Detail
>>
>> CCD, when properly cooled, can achieve higher image quality than the CMOS
>> counter part, and is why CCD is still used for astronomy; however, I am not
>> sure that there are many DSLRs around with portable liquid cooling systems in
>> them ...
>
>CCD verses CMOS sensor size in a DSLR has more to do with lens design
>than in a P&S. There is full frame or, more commonly, the 1.6x crop
>size. Whether CCD or CMOS the physical size of the sensor has to be
>more consistent to conform to the lens properties across differing
>camera models while maintaining future compatibility. P&S cameras have
>dedicated unchangeable lenses which allow CCD sensors to be smaller
>without any of these concerns. The sensor size in a Sony DSC-W100
>(8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>crop DSLR is roughly 338 square mm no matter what the megapixel rating.
> Out of this same area you can get 8-9 CCD type sensors. It is a
>similar effect that AMD and Intel sees from going to a 65nm process from
>a 90nm one. They get more product from the same wafer size which means
>higher yields/profits and/or lower prices per unit.
>
>As for the original poster's question, P&S camera's typically use CCD
>sensors. I have seen very low quality (and low mp count) P&S cameras
>use CMOS sensors. The camera you referenced is one of these since I
>could buy five of them for $550. I would also suspect the quality of
>the video to be low resolution. If CMOS were the better choice for the
>manufacturers of P&S cameras then I would expect them to be the norm and
>not the rare exception. CMOS does appear to be making inroads in the
>camcorder market. Personally, I think the camcorder will merge with the
>P&S camera in the coming years and if this happens the CCD sensor may be
>the ultimate winner.
>
>As for the end user's considerations, CCD based cameras tend to have
>higher noise levels. This is changing somewhat with the newer crop of
>higher ISO, lower noise P&S cameras. P&S cameras using a CMOS sensor
>are typically low end with low megapixel counts. Maybe this is because
>a CCD can pack more pixel sites in the same area compared to a CMOS
>chip. This promotes the design of smaller cameras while retaining high
>resolution which seems to be what the P&S manufacturers have been
>striving for the past several years. There are high end DSLR's using
>CCD sensors. However, I don't know of any high end P&S cameras that use
>CMOS sensors. Take what you will from this observation. Finally, I am
>basing the heat issue of CMOS compared to CCD sensors on what I have
>read here in this newsgroup. Maybe what I read was incorrect. IMO,
>this is a minor point of the discussion at hand.
>


Each pixel in high end CMOS chips such as the ones Canor uses in it
DSLR have two sections, the light colecting sensor and the transistor
circuitry that keeps the noise low. This circuitry takes up silicon
real estate, as much as 70 % in some of the earlier chips. So if you
tried to make a small low noise CMOS chip for a point and shoot
camera, you'd reach the point where the pixel was all transistor
ciruitry with no room for a light sensor.

jpc
 >> Stay informed about: Sensors 

On Mon, 25 Sep 2006 19:50:07 -0400, "Michael Johnson, PE"
<cds.TakeThisOut@erols.com> wrote:

>Bill Funk wrote:
>> On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
>> <cds.TakeThisOut@erols.com> wrote:
>>
>>> CCD verses CMOS sensor size in a DSLR has more to do with lens design
>>> than in a P&S. There is full frame or, more commonly, the 1.6x crop
>>> size. Whether CCD or CMOS the physical size of the sensor has to be
>>> more consistent to conform to the lens properties across differing
>>> camera models while maintaining future compatibility.
>>
>> Are you sure?
>> Then why do EF lenses work on both APS-C sensors AND full frame
>> sensors?
>> And, on Nikon cameras, why do their non-Digital (I don't know the
>> appelation) work on both their APS sensor cameras and on their (old)
>> full frame DSLRs and in both cases (Canon and Nikon) these leneses
>> also work on the (obviously) full frame film cameras?
>
>Sensor size is roughly the same as film size for DSLR cameras. If you
>fit a 35mm film camera with a 24mm film roll what would happen? There
>would roughly be a 1.6x crop factor like there is with APS sensor DSLR.
> Nikon or Canon could come out with a 3x crop factor DSLR but why?
>Since lenses are changed on DSLRs having a sensor standard is more
>important. It makes lenses operate the same on a 10D, 300D, 20D, 30D,
>350D 400D etc. This isn't a concern on P&S cameras since the lens stays
>with the camera and can't be changed. I'm not sure what your point is here.

The point is this:
"CCD verses CMOS sensor size in a DSLR has more to do with lens design
than in a P&S."
Whether the sensor is CCD or a CMOS doesn't alter the choice of sensor
*size*, in a DSLR.

>
>>> P&S cameras have
>>> dedicated unchangeable lenses which allow CCD sensors to be smaller
>>> without any of these concerns. The sensor size in a Sony DSC-W100
>>> (8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>>> crop DSLR is roughly 338 square mm no matter what the megapixel rating.
>>> Out of this same area you can get 8-9 CCD type sensors.
>>
>> Well, of course you can.
>> But those sensors won't fit in a camera that uses an APS-size sensor,
>> like a Nikon D70s.
>
>Sure they will. You'll just have a higher crop factor. This isn't
>really practical for the reasons I gave above.

Yes, that *could* fit, but that destroys the point of the whole 1.6
crop factor camera.
>
>> Using he same size (or approximately, as with nikon<>Canon, the
>> sensors will be the same size, whether CCD or CMOS.
>
>Once again the sensor size on a DSLR is the roughly the same for
>consistency between models and lenses.

But it seems from what you're quoted as saying above, that you are
saying that there's a size difference inherent in CCD and CMOS
sensors.
Again:
"CCD verses CMOS sensor size in a DSLR has more to do with lens design
than in a P&S." That's confusing to me.
Why would a CCD sensor be a different size from a CMOS sensor in, say,
a Canon 30D or a Nikon D80?
>
>>> It is a
>>> similar effect that AMD and Intel sees from going to a 65nm process from
>>> a 90nm one. They get more product from the same wafer size which means
>>> higher yields/profits and/or lower prices per unit.
>>
>> But that's not the reason they do this - it's to get faster speed
>> while keeping heat down. The higher yield is a happy byproduct.
>
>I doubt the higher yield, and lower resulting cost per sensor, is just a
>byproduct but one of the main considerations in using CCD sensors in P&S
>cameras. This also allows the cameras to be smaller while retaining
>high resolution which seals the deal for the overwhelming majority of
>current P&S cameras to use CCD sensors, IMHO, of course.

If a maker substituted a CMOS sensor in a camera that currently uses a
CCD sensor,the sensor would be the same size, so yield would remain
the same.
--
Bill Funk
replace "g" with "a"
 >> Stay informed about: Sensors 

jpc wrote:
> On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
> <cds.DeleteThis@erols.com> wrote:
>
>> Thomas T. Veldhouse wrote:
>>> Michael Johnson, PE <cds.DeleteThis@erols.com> wrote:
>>>> CMOS may be cheaper to manufacture on an area basis but where CCD
>>>> catches up is they are smaller more fit on a wafer (or the photo chip
>>>> equivalent) so there is better economy of scale. Maybe I am wrong.
>>>> Also, I thought the reason CMOS sensors aren't ideal for video is they
>>>> can't do the rapid frame rate for extended time periods needed for
>>>> smooth video without overheating.
>>> Really? Which is bigger, the CMOS 6.3MP sensor that the Canon Digital Rebel
>>> uses or the CCD 6.1MP sensor the Nikon D50/D70 and D70s use (among the Nikon
>>> branded cameras ... many others use that sensor).
>>>
>>> There are plenty of CMOS based video cameras out there. This one comes to
>>> mind with a simple Google search.
>>>
>>> http://www.globalsourcesdirect.com/servlet/the-250/Digital-Video-Camera/Detail
>>>
>>> CCD, when properly cooled, can achieve higher image quality than the CMOS
>>> counter part, and is why CCD is still used for astronomy; however, I am not
>>> sure that there are many DSLRs around with portable liquid cooling systems in
>>> them ...
>> CCD verses CMOS sensor size in a DSLR has more to do with lens design
>> than in a P&S. There is full frame or, more commonly, the 1.6x crop
>> size. Whether CCD or CMOS the physical size of the sensor has to be
>> more consistent to conform to the lens properties across differing
>> camera models while maintaining future compatibility. P&S cameras have
>> dedicated unchangeable lenses which allow CCD sensors to be smaller
>> without any of these concerns. The sensor size in a Sony DSC-W100
>> (8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>> crop DSLR is roughly 338 square mm no matter what the megapixel rating.
>> Out of this same area you can get 8-9 CCD type sensors. It is a
>> similar effect that AMD and Intel sees from going to a 65nm process from
>> a 90nm one. They get more product from the same wafer size which means
>> higher yields/profits and/or lower prices per unit.
>>
>> As for the original poster's question, P&S camera's typically use CCD
>> sensors. I have seen very low quality (and low mp count) P&S cameras
>> use CMOS sensors. The camera you referenced is one of these since I
>> could buy five of them for $550. I would also suspect the quality of
>> the video to be low resolution. If CMOS were the better choice for the
>> manufacturers of P&S cameras then I would expect them to be the norm and
>> not the rare exception. CMOS does appear to be making inroads in the
>> camcorder market. Personally, I think the camcorder will merge with the
>> P&S camera in the coming years and if this happens the CCD sensor may be
>> the ultimate winner.
>>
>> As for the end user's considerations, CCD based cameras tend to have
>> higher noise levels. This is changing somewhat with the newer crop of
>> higher ISO, lower noise P&S cameras. P&S cameras using a CMOS sensor
>> are typically low end with low megapixel counts. Maybe this is because
>> a CCD can pack more pixel sites in the same area compared to a CMOS
>> chip. This promotes the design of smaller cameras while retaining high
>> resolution which seems to be what the P&S manufacturers have been
>> striving for the past several years. There are high end DSLR's using
>> CCD sensors. However, I don't know of any high end P&S cameras that use
>> CMOS sensors. Take what you will from this observation. Finally, I am
>> basing the heat issue of CMOS compared to CCD sensors on what I have
>> read here in this newsgroup. Maybe what I read was incorrect. IMO,
>> this is a minor point of the discussion at hand.
>>
>
>
> Each pixel in high end CMOS chips such as the ones Canor uses in it
> DSLR have two sections, the light colecting sensor and the transistor
> circuitry that keeps the noise low. This circuitry takes up silicon
> real estate, as much as 70 % in some of the earlier chips. So if you
> tried to make a small low noise CMOS chip for a point and shoot
> camera, you'd reach the point where the pixel was all transistor
> ciruitry with no room for a light sensor.

I wonder if the CCD will find their way into most DSLRs if the noise
issues are resolved. The P&S cameras look to have made advances in
noise reduction in the last year but they are still not up with the
current CMOS technology. Using CCD sensors would allow for the
megapixel wars to continue (not that they need to) in the DSLR market
since more pixels can be packed into the same area relative to CMOS sensors.
 >> Stay informed about: Sensors 

On Tue, 26 Sep 2006 10:12:05 -0400, "Michael Johnson, PE"
<cds DeleteThis @erols.com> wrote:

>jpc wrote:
>> On Mon, 25 Sep 2006 17:10:04 -0400, "Michael Johnson, PE"
>> <cds DeleteThis @erols.com> wrote:
>>
>>> Thomas T. Veldhouse wrote:
>>>> Michael Johnson, PE <cds DeleteThis @erols.com> wrote:
>>>>> CMOS may be cheaper to manufacture on an area basis but where CCD
>>>>> catches up is they are smaller more fit on a wafer (or the photo chip
>>>>> equivalent) so there is better economy of scale. Maybe I am wrong.
>>>>> Also, I thought the reason CMOS sensors aren't ideal for video is they
>>>>> can't do the rapid frame rate for extended time periods needed for
>>>>> smooth video without overheating.
>>>> Really? Which is bigger, the CMOS 6.3MP sensor that the Canon Digital Rebel
>>>> uses or the CCD 6.1MP sensor the Nikon D50/D70 and D70s use (among the Nikon
>>>> branded cameras ... many others use that sensor).
>>>>
>>>> There are plenty of CMOS based video cameras out there. This one comes to
>>>> mind with a simple Google search.
>>>>
>>>> http://www.globalsourcesdirect.com/servlet/the-250/Digital-Video-Camera/Detail
>>>>
>>>> CCD, when properly cooled, can achieve higher image quality than the CMOS
>>>> counter part, and is why CCD is still used for astronomy; however, I am not
>>>> sure that there are many DSLRs around with portable liquid cooling systems in
>>>> them ...
>>> CCD verses CMOS sensor size in a DSLR has more to do with lens design
>>> than in a P&S. There is full frame or, more commonly, the 1.6x crop
>>> size. Whether CCD or CMOS the physical size of the sensor has to be
>>> more consistent to conform to the lens properties across differing
>>> camera models while maintaining future compatibility. P&S cameras have
>>> dedicated unchangeable lenses which allow CCD sensors to be smaller
>>> without any of these concerns. The sensor size in a Sony DSC-W100
>>> (8.1mp) is 7.18x5.32 mm or 38.2 square mm. The sensor area of a 1.6
>>> crop DSLR is roughly 338 square mm no matter what the megapixel rating.
>>> Out of this same area you can get 8-9 CCD type sensors. It is a
>>> similar effect that AMD and Intel sees from going to a 65nm process from
>>> a 90nm one. They get more product from the same wafer size which means
>>> higher yields/profits and/or lower prices per unit.
>>>
>>> As for the original poster's question, P&S camera's typically use CCD
>>> sensors. I have seen very low quality (and low mp count) P&S cameras
>>> use CMOS sensors. The camera you referenced is one of these since I
>>> could buy five of them for $550. I would also suspect the quality of
>>> the video to be low resolution. If CMOS were the better choice for the
>>> manufacturers of P&S cameras then I would expect them to be the norm and
>>> not the rare exception. CMOS does appear to be making inroads in the
>>> camcorder market. Personally, I think the camcorder will merge with the
>>> P&S camera in the coming years and if this happens the CCD sensor may be
>>> the ultimate winner.
>>>
>>> As for the end user's considerations, CCD based cameras tend to have
>>> higher noise levels. This is changing somewhat with the newer crop of
>>> higher ISO, lower noise P&S cameras. P&S cameras using a CMOS sensor
>>> are typically low end with low megapixel counts. Maybe this is because
>>> a CCD can pack more pixel sites in the same area compared to a CMOS
>>> chip. This promotes the design of smaller cameras while retaining high
>>> resolution which seems to be what the P&S manufacturers have been
>>> striving for the past several years. There are high end DSLR's using
>>> CCD sensors. However, I don't know of any high end P&S cameras that use
>>> CMOS sensors. Take what you will from this observation. Finally, I am
>>> basing the heat issue of CMOS compared to CCD sensors on what I have
>>> read here in this newsgroup. Maybe what I read was incorrect. IMO,
>>> this is a minor point of the discussion at hand.
>>>
>>
>>
>> Each pixel in high end CMOS chips such as the ones Canor uses in it
>> DSLR have two sections, the light colecting sensor and the transistor
>> circuitry that keeps the noise low. This circuitry takes up silicon
>> real estate, as much as 70 % in some of the earlier chips. So if you
>> tried to make a small low noise CMOS chip for a point and shoot
>> camera, you'd reach the point where the pixel was all transistor
>> ciruitry with no room for a light sensor.
>
>I wonder if the CCD will find their way into most DSLRs if the noise
>issues are resolved. The P&S cameras look to have made advances in
>noise reduction in the last year but they are still not up with the
>current CMOS technology. Using CCD sensors would allow for the
>megapixel wars to continue (not that they need to) in the DSLR market
>since more pixels can be packed into the same area relative to CMOS sensors.


There are a couple things going on that give large pixel CMOS sensors
an edge. The first is that the transitor circuitry in the individual
CMOS pixels does a better job of cleaning up what I call "left behind
noise" That's the 1 to maybe up to 12 photoelectrons that aren't
cleaned out in the idividual CCD pixels when the CCD circuitry resets
everything just before the shutter opens. A few years ago when I last
looked into the problem that was a major source of camera noise. I
wouldn't be surprised if one of the main reason CCD sensors have
improved recently is that the chip people have come up with a better
way of zeroing out their sensors.

A far more importart reason for using large sensors is that over most
of the dynamic range of even small sensors, modern cameras are photon
shot noise limited. That means the noise is in the light not the
camera and is caused by the random way Mama Nature toses light into
your lens. There a few tricks you can play to cut down on that noise
in specific exposure regions--using a sensor with a cyan, magenta, and
yellow color mask rather than a red green and blue mask like Nikon,
Olympus and maybe a few other did four or five years ago--but the only
good way to cut down on the noise is to collect more photoelectons in
a large sensor. That way you end up with a better statistical averge
and less noise.

jpc
 >> Stay informed about: Sensors 

Michael Johnson, PE <cds.DeleteThis@erols.com> wrote:
>
> I doubt the higher yield, and lower resulting cost per sensor, is just a
> byproduct but one of the main considerations in using CCD sensors in P&S
> cameras. This also allows the cameras to be smaller while retaining
> high resolution which seals the deal for the overwhelming majority of
> current P&S cameras to use CCD sensors, IMHO, of course.

Tell Canon that ... they use CMOS in their point and shoot cameras. I am
quite confident that they are doing it cheaper than Nikon is for their point
and shoot camera.

--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
 >> Stay informed about: Sensors 

Thomas T. Veldhouse wrote:
> Michael Johnson, PE <cds.TakeThisOut@erols.com> wrote:
>
>>I doubt the higher yield, and lower resulting cost per sensor, is just a
>>byproduct but one of the main considerations in using CCD sensors in P&S
>>cameras. This also allows the cameras to be smaller while retaining
>>high resolution which seals the deal for the overwhelming majority of
>>current P&S cameras to use CCD sensors, IMHO, of course.
>
>
> Tell Canon that ... they use CMOS in their point and shoot cameras. I am
> quite confident that they are doing it cheaper than Nikon is for their point
> and shoot camera.
>
Which ones? I just picked up a S3IS and it's got a CCD, according to the
online specs..

--
I do *NOT* have a short attention sp...(Oooh!! shiny!!)
 >> Stay informed about: Sensors 

"Thomas T. Veldhouse" <veldy71.DeleteThis@yahoo.com> writes:
> Tell Canon that ... they use CMOS in their point and shoot cameras.

Which models?
 >> Stay informed about: Sensors