Reading mode: Light Dark

Pixel Pitch Universally Defined

Go to last post
M

mgates

Leading Member
Messages
624
Reaction score
0
Location
US
Sensor Width = SW
Sensor Height = SH
Sensor Area = A = SW x SH
Megapixels w/Control Logic = MP
Silicon Control Logic Density = SCLD = A/MP
Pixel Pitch = SQRT(SCLD)

Canon 5D
Sensor Width = SW = 35.8
Sensor Height = SH = 23.9
Sensor Area = A = SW x SH = 856
Megapixels w/Control Logic = MP = 12,700,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000675
Pixel Pitch = SQRT(SCLD) = 0.0082

Canon 1D Mk II
Sensor Width = SW = 28.7
Sensor Height = SH = 19.1
Sensor Area = A = SW x SH = 548
Megapixels w/Control Logic = MP = 8,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000668
Pixel Pitch = SQRT(SCLD) = 0.0081

Nikon D200
Sensor Width = SW = 23.6
Sensor Height = SH = 15.8
Sensor Area = A = SW x SH = 373
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000373
Pixel Pitch = SQRT(SCLD) = 0.0061

Sigma SD10
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 10,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000280
Pixel Pitch = SQRT(SCLD) = 0.0053

Sigma SD14
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 14,100,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000203
Pixel Pitch = SQRT(SCLD) = 0.0045

Olympus Stylus 1000
Sensor Width = SW = 7.2
Sensor Height = SH = 5.3
Sensor Area = A = SW x SH = 38.2
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.00000382
Pixel Pitch = SQRT(SCLD) = 0.0020
 
the mp count is wrong for the SD14, it should be 4.7mp for photosites. The sensor doesn't have 14million sections, but 4.7 million stacked, therefore the pixels are 3 times wider and 9 times more area
--
long live the rebellion...XT
 
Darth_Furious said:
the mp count is wrong for the SD14, it should be 4.7mp for
photosites. The sensor doesn't have 14million sections, but 4.7
million stacked, therefore the pixels are 3 times wider and 9 times
more area
--
long live the rebellion...XT
Click to expand...
The SD14 is 14MP. It has 14 million pixels requiring electronic control logic. That is what creates noise. If it was 3CCD it would have a 7.8 micron pixel pitch. It is two silicon chips short of a 7.8 pitch.
 
"The new X3 sensor packs 14.1 million pixels, each 7.8μm in size, into a 3-

dimensional array of 2652 x 1768 x 3 pixels with an active image area measuring 20.69 x
13.79 mm."

http://www.foveon.com/article.php?a=222

--
Chunsum.

'Once the amateur's naive approach and humble
willingness to learn fades away, the creative spirit of good
photography dies with it. Every professional should remain
always in his heart an amateur.'
  • Alfred Eisenstaedt
http://www.chunsum.com
http://www.pbase.com/chunsum
http://www.pbase.com/sigmadslr/chunsum_choi
 
Chunsum J Choi said:
"The new X3 sensor packs 14.1 million pixels, each 7.8μm in
size, into a 3-
dimensional array of 2652 x 1768 x 3 pixels with an active image
area measuring 20.69 x
13.79 mm."

http://www.foveon.com/article.php?a=222
Click to expand...
VLSI circuit layering is common to both CMOS designs. VLSI layers DO NOT add silicon but do compound circuit density. .

We can apply the Foveon method to Bayer sensors. Summing silicon area of component sensors in the color pixel. Totalling MPs using the individual method. The 5D gets 12.7MP and a pixel pitch of 16.4μm. It isn't a bad method either. Foveon shouldn't use it. They get less favorable ratios.

16.4/7.8 = 2.1
8.1/4.5 = 1.8

A sanity check is MP(Pixel Pitch^2) must always equal chip die. Correct math works in all cases. Foveon's method gives Foveon three times more silicon than die. Bayer then gets four times more. If you double phyiscal silicon for every VLSI print than silicon becomes an unlimted resource for both designs.

I am sure Foveon would gladly acknowledge their error if it was pointed out to them.
 
Why redefine pixel pitch? It already has a perfectly standard definition. The SD10 is 9.1 microns and the SD14 is 7.8 microns.

j
mgates said:
Sensor Width = SW
Sensor Height = SH
Sensor Area = A = SW x SH
Megapixels w/Control Logic = MP
Silicon Control Logic Density = SCLD = A/MP
Pixel Pitch = SQRT(SCLD)

Canon 5D
Sensor Width = SW = 35.8
Sensor Height = SH = 23.9
Sensor Area = A = SW x SH = 856
Megapixels w/Control Logic = MP = 12,700,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000675
Pixel Pitch = SQRT(SCLD) = 0.0082

Canon 1D Mk II
Sensor Width = SW = 28.7
Sensor Height = SH = 19.1
Sensor Area = A = SW x SH = 548
Megapixels w/Control Logic = MP = 8,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000668
Pixel Pitch = SQRT(SCLD) = 0.0081

Nikon D200
Sensor Width = SW = 23.6
Sensor Height = SH = 15.8
Sensor Area = A = SW x SH = 373
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000373
Pixel Pitch = SQRT(SCLD) = 0.0061

Sigma SD10
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 10,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000280
Pixel Pitch = SQRT(SCLD) = 0.0053

Sigma SD14
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 14,100,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000203
Pixel Pitch = SQRT(SCLD) = 0.0045

Olympus Stylus 1000
Sensor Width = SW = 7.2
Sensor Height = SH = 5.3
Sensor Area = A = SW x SH = 38.2
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.00000382
Pixel Pitch = SQRT(SCLD) = 0.0020
Click to expand...
 
Just Looking said:
Why redefine pixel pitch? It already has a perfectly standard
definition. The SD10 is 9.1 microns and the SD14 is 7.8 microns.
Click to expand...
Ok 3.4MP and 4.7MP. Not 10MP and 14MP. Can't have it both ways. Both ways the 5D has 12.7MP and RGGB center-to-center pitch of 16.4. Foveon is relatively worse off.
 
It calculates Bayer Green pitch smaller than red or blue pitch. That is correct. Bayer Green has more resolution. When average pitch is determined the universal value is again reached.
mgates said:
Just Looking said:
http://www.xenics.com/Products/Glossary.php
...the distance between the center of two adjacent pixels in an array.
Click to expand...
Ok. Red to red. Green to green. Blue to blue. Again Foveon is
worse off. The math always works the same. You never get more
silicon by adding control logic.
Click to expand...
 
mgates said:
The SD14 is 14MP. It has 14 million pixels requiring electronic
control logic. That is what creates noise.
Click to expand...
Not really. It is one potential source of noise. However for the dSLRs we are talking about (including the Sigma) noise only starts to become an issue at high ISO. At this point photon count is low and shot noise starts to predominate (at least for normal exposure times - read noise etc is more significant with long exposures).

If what you say was true, pixel pitch wouldn't be important at all - just the number of logic controls. However, pixel pitch, or if you like photosite are is important. Larger photosites gather more photons for a given light intensity. Therefore they have a better SNR with regard to shot noise.

Foveon photosites are quite large, and therfore should perform quite well in this regard. On the other hand, the second and third layers receive a reduced signal that has already undergone some absorbtion. Of course Bayer photosites suffer pretty much the same thing, as each photosite recevies light that has already been filtered to remove much of the spectrum.
 
You need to think about the active area of each photosite available to receive photons, not the number of logic control elements. In this regard, a system where the photosites are stacked vertically on top of each other has a potential advantage - each RGB sensor has a larger area than an equivalent CMOS or CCD. There will be some absorption loss with each layer, but then of course with Bayer there is loss due to the CFA filter.

I think you're barking up the wrong tree, indeed in the wrong wood, with your attempt to define an imaginary "virtual" photosite for the Foveon sensor.
 
This looks like a synopsis of the reading list for Pixelscopophilia 101. The appearance of the word "logic" so frequently is a bit disconcerting, however.

Jim?
mgates said:
Sensor Width = SW
Sensor Height = SH
Sensor Area = A = SW x SH
Megapixels w/Control Logic = MP
Silicon Control Logic Density = SCLD = A/MP
Pixel Pitch = SQRT(SCLD)

Canon 5D
Sensor Width = SW = 35.8
Sensor Height = SH = 23.9
Sensor Area = A = SW x SH = 856
Megapixels w/Control Logic = MP = 12,700,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000675
Pixel Pitch = SQRT(SCLD) = 0.0082

Canon 1D Mk II
Sensor Width = SW = 28.7
Sensor Height = SH = 19.1
Sensor Area = A = SW x SH = 548
Megapixels w/Control Logic = MP = 8,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000668
Pixel Pitch = SQRT(SCLD) = 0.0081

Nikon D200
Sensor Width = SW = 23.6
Sensor Height = SH = 15.8
Sensor Area = A = SW x SH = 373
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000373
Pixel Pitch = SQRT(SCLD) = 0.0061

Sigma SD10
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 10,200,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000280
Pixel Pitch = SQRT(SCLD) = 0.0053

Sigma SD14
Sensor Width = SW = 20.7
Sensor Height = SH = 13.8
Sensor Area = A = SW x SH = 286
Megapixels w/Control Logic = MP = 14,100,000
Silicon Control Logic Density = SCLD = A/MP = 0.0000203
Pixel Pitch = SQRT(SCLD) = 0.0045

Olympus Stylus 1000
Sensor Width = SW = 7.2
Sensor Height = SH = 5.3
Sensor Area = A = SW x SH = 38.2
Megapixels w/Control Logic = MP = 10,000,000
Silicon Control Logic Density = SCLD = A/MP = 0.00000382
Pixel Pitch = SQRT(SCLD) = 0.0020
Click to expand...
--
Laurence

My idea of good company is the fellowship of clever, well-informed people, who have a great deal of conversation and liberality of ideas.

Jane Austen

http://www.pbase.com/lmatson/root
http://www.pbase.com/sigmadslr/root
http://www.pbase.com/cameras/sigma/sd10
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com
 
Tristan Cope said:
mgates said:
The SD14 is 14MP. It has 14 million pixels requiring electronic
control logic. That is what creates noise.
Click to expand...
Not really. It is one potential source of noise. However for the
dSLRs we are talking about (including the Sigma) noise only starts
to become an issue at high ISO. At this point photon count is low
and shot noise starts to predominate (at least for normal exposure
times - read noise etc is more significant with long exposures).

If what you say was true, pixel pitch wouldn't be important at all
  • just the number of logic controls. However, pixel pitch, or if
you like photosite are is important. Larger photosites gather more
photons for a given light intensity. Therefore they have a better
SNR with regard to shot noise.
Click to expand...
Excellent point. However if you ignore the presence of control logic Bayer red green or blue photosites become much larger.
Tristan Cope said:
Foveon photosites are quite large, and therfore should perform
quite well in this regard. On the other hand, the second and third
layers receive a reduced signal that has already undergone some
absorbtion. Of course Bayer photosites suffer pretty much the same
thing, as each photosite recevies light that has already been
filtered to remove much of the spectrum.
Click to expand...
 
Tristan Cope said:
You need to think about the active area of each photosite available
to receive photons, not the number of logic control elements. In
this regard, a system where the photosites are stacked vertically
on top of each other has a potential advantage - each RGB sensor
has a larger area than an equivalent CMOS or CCD. There will be
some absorption loss with each layer, but then of course with Bayer
there is loss due to the CFA filter.

I think you're barking up the wrong tree, indeed in the wrong wood,
with your attempt to define an imaginary "virtual" photosite for
the Foveon sensor.
Click to expand...
Foveon is claiming virtual photosites. That is a good conceptual term for this mathematical error. Since MP(Pixel Pitch^2) = 3*available silicon. If you measure virtual pixel pitch by channel Bayer ignoring total density Bayer gets MP(Pixel Pitch^2) = 4*available silicon. Red to red etc.
 
mgates,

Pitch is the physical distance from one pixel to another, either along the width or along the height ( ). The horizontal and the vertical pitch does not need to be the same; there are some examples of that.

It does not matter if we call a SD14 a 14 MP or a 4.7 MP camera, the pitch is still the same, and it is defined by the 4.7 million locations where we have detector triplets.

For SD14 the horizontal pitch and the vertical pitch are the same and it is SQRT(286/4.7) um = 7.8 um.

--
Roland
http://klotjohan.mine.nu/~roland/

( ) For the 45 degree tilted Fuji thingie I don't know what they call pitch.
 
mgates said:
Chunsum J Choi said:
"The new X3 sensor packs 14.1 million pixels, each 7.8μm in
size, into a 3-
dimensional array of 2652 x 1768 x 3 pixels with an active image
area measuring 20.69 x
13.79 mm."

http://www.foveon.com/article.php?a=222
Click to expand...
VLSI circuit layering is common to both CMOS designs. VLSI layers
DO NOT add silicon but do compound circuit density. .
Click to expand...
AFAIK, the Foveon imager has 3 seperate layers of photosites , you can say it is like 3 ccd stacked together. Each layer contains 4.7Mphotosites with 7.8μm between each of them.

I hope this simple non-technical description from a non-technical person can help you see that your calculations are incorrect.
mgates said:
We can apply the Foveon method to Bayer sensors. Summing silicon
area of component sensors in the color pixel. Totalling MPs using
the individual method. The 5D gets 12.7MP and a pixel pitch of
16.4μm. It isn't a bad method either. Foveon shouldn't use
it. They get less favorable ratios.

16.4/7.8 = 2.1
8.1/4.5 = 1.8
Click to expand...
I think you are confused. like others has said, Pixel Pitch is a "PHYSICAL ATTRIBUTE" of the sensor. not something you calculate in algebra class.
mgates said:
A sanity check is MP(Pixel Pitch^2) must always equal chip die.
Correct math works in all cases. Foveon's method gives Foveon
three times more silicon than die. Bayer then gets four times
more. If you double phyiscal silicon for every VLSI print than
silicon becomes an unlimted resource for both designs.

I am sure Foveon would gladly acknowledge their error if it was
pointed out to them.
Click to expand...
I am sure they will be glad to point out the faulty points of your logic.

--
Chunsum.

'Once the amateur's naive approach and humble
willingness to learn fades away, the creative spirit of good
photography dies with it. Every professional should remain
always in his heart an amateur.'
  • Alfred Eisenstaedt
http://www.chunsum.com
http://www.pbase.com/chunsum
http://www.pbase.com/sigmadslr/chunsum_choi
 
Layers. You do understand what layers means don't you? You seem to be struggling with the concept.
mgates said:
Tristan Cope said:
You need to think about the active area of each photosite available
to receive photons, not the number of logic control elements. In
this regard, a system where the photosites are stacked vertically
on top of each other has a potential advantage - each RGB sensor
has a larger area than an equivalent CMOS or CCD. There will be
some absorption loss with each layer, but then of course with Bayer
there is loss due to the CFA filter.

I think you're barking up the wrong tree, indeed in the wrong wood,
with your attempt to define an imaginary "virtual" photosite for
the Foveon sensor.
Click to expand...
Foveon is claiming virtual photosites. That is a good conceptual
term for this mathematical error. Since MP(Pixel Pitch^2) =
3*available silicon. If you measure virtual pixel pitch by channel
Bayer ignoring total density Bayer gets MP(Pixel Pitch^2) =
4*available silicon. Red to red etc.
Click to expand...
 
mgates said:
Excellent point. However if you ignore the presence of control
logic Bayer red green or blue photosites become much larger.
Click to expand...
What we should be considering is the active sensing area of each photosite - for both systems. This means that "pixel pitch" and fill factor are important.
 
mgates said:
Sensor Width = SW
Sensor Height = SH
Sensor Area = A = SW x SH
Megapixels w/Control Logic = MP
Silicon Control Logic Density = SCLD = A/MP
Pixel Pitch = SQRT(SCLD)
Click to expand...
You are trying to state Foveon X3 sensor's pixel is the same as a Bayer sensor's pixel. We all know this is a false statement as it is well accepted that a 3.4mp X3 Foveon sensor will perform just as well or better then a 6mp Bayer sensor.

Your last three definations do not take that fact into consideration.

We should be able to expect the 4.7mp X3 sensor within the SD14 will compare well against existing 8-10mp Bayer sensors.

--
jc
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top