The Foveon X3 Sensor

Ivan Glisin · 07-09-2010, 05:35 AM - **2 Likes**

Originally posted by Winder

I would like to see a monochrome K-7. No Bayer filter. The resolution of a true 14MP monochrome sensor would be amazing.

I second this. A K-7GS (grayscale) perhaps?

One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer).

However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one.

(And I still have my color filters for BW photography!)

ghelary · 07-09-2010, 06:33 AM

Originally posted by Ivan Glisin

I second this. A K-7GS (grayscale) perhaps?

One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer).

However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one.

(And I still have my color filters for BW photography!)

Yup, I've been thinking of that myself too, excepted for the sensitivity part. Basically, filter will make you loose the gain in sensitivy obtain by removing the bayer matrix. But, hey, nearly no more AA filter needed !

Rico · 07-09-2010, 06:42 AM

Originally posted by falconeye

I manually whitebalanced both images to the white tent on the left hand side and shifted colors in the K-7 sample to get X3's wrong blue sky effect (I exxagerated a little bit. In lightroom, you mostly draw the Aqua, Blue and Magenta channel sliders to -20 or -40 and push the Blue channel towards a Green hue).

The result is attached. Except for wrong colors, I cannot see a difference. However, I wonder which X3 color calibration gives such wrong colors...

Hi Falk

I have to say I set the images up pretty quickly. I changed the color mode in the image below from neutral to landscape which makes a big difference. The color adjustment was about the same. Whitebalance was kept the same too Sunlight. Funny there is a setting for original under color mode that the Sigma Pro software will not allow me to choose for whatever reason.

I think this image looks better than the first X3 image I posted:

Here is one of the first images I took with the DP2s. It is not a macro camera and it was windy but I think it demonstrates the potential of Foveon X3. There is a pop and a clarity to the image similar to Kodachrome IMHO:

falconeye · 07-09-2010, 09:38 AM

Originally posted by Rico

it demonstrates the potential of Foveon X3. There is a pop and a clarity to the image similar to Kodachrome IMHO:

I see what you mean. But we must be careful with assessments like this one.

The K-7 image has a relatively linear and accurate color perfomance while the X3 images (the ones you show) have a huge CIE deltaE error (i.e., even if white balance is correct, some colors are simply way off).

I tried to "emulate" this for the K-7 images which means to suppress some color sub channels significantly. And while the colors shift, the image gains significantly in "pop". It's like putting a filter in B&W photography.

So, what you see is an "effect". You may say it's a nice effect because it creates pop. But for sure, it delivers wrong colors too. We all know that from the days of analog film.

To tell the truth, I prefer to be in control of such effects, e.g., by using DxO film emulation package and the clarity slider.

Winder · 07-10-2010, 09:20 AM

Originally posted by Ivan Glisin

I second this. A K-7GS (grayscale) perhaps?

One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer).

However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one.

(And I still have my color filters for BW photography!)

I have seen several images from monochrome digital MF backs and the detail in textures the tonal range was very impressive.

I'm still holding out for a 5/4 format monochrome sensor......

glanglois · 07-14-2010, 12:05 PM

What are the other implications of a B&W camera?

For example, what are the results of not needing white balance, chroma NR, adjustments for hue/saturation, or color-sensitive scene modes? What is result (if any) upon AF if we're using color filters in front of the lens?

Does this make Ricoh's idea look better, swapping lens and sensor as a package?

Does this lead to camera design in which the sensor can be changed like film or a film back?

I have to go to think now ....

deejjjaaaa · 07-14-2010, 06:16 PM

Originally posted by Rico

Hi Falk

I have to say I set the images up pretty quickly.

Rico, you do a typical newbie error - you raw conversion using SPP gets cyan poisoning... use SilkyPix, it will not let such cyan tint to happen.

luck1 · 07-15-2010, 11:43 AM

found this on dpreview forum, to the color shift. SD15 colors compared to SD14 and 5D MkII: Sigma SLR Talk Forum: Digital Photography Review

ddhytz · 07-16-2010, 01:40 AM

I have a DP1, a camera I carry with me most of time. Some sample shots:

100% crop

ddhytz · 07-16-2010, 01:44 AM

More sample shots

**yucatanPentax** · 02-27-2011, 12:21 AM

ddhytz,
Those are certainly nice captures. It seems the color balance isn't always off. Your shots show a fairly impressive amount of detail and seemingly accurate colors. I don't know enough about the technology to know why the colors / detail/ contrast may seem off at times and not at others. But I do like your shots!

stillshunter · 03-08-2011, 01:32 AM

Guys,

I have recently bought a DP2s as my 'back-up' compact. Can I say that I am blown away each and every time I see the outputs of the lil Sigma. To be honest, I now find myself reaching for it 9 times out of 10 over the K20D. There really is something to the Foveon files... something that is lacking from the Pentax images as well as the Nikon, and Olympus ones that preceded it. To be honest I am tempted by the SD1. Oh who am I kidding I am VERY VERY tempted!

For those who are working from specs alone, I say please defer your comments until you've actually tried the Foveon yourself. You - like me - might just be very pleasantly surprised.

robgrune · 04-20-2011, 11:11 PM

Based upon the large jpg files I downloaded from the various manufacturer websites (Leica, Panasonic, Sigma, Sony, Nikon), I can discern no difference to the images on-screen at about the A4 size, and no differences when printed at A4. At A4 size, the Foveon imager (4/3 frame) gives the best results for colour, contrast, and detail! But, going larger, to say A2 size and up, the Leica S2 (imager larger than full frame) gives the best image on screen, with Leica M9 coming second. The Foveon colour and contrast remain good, but the noise results in hard, serrated edges for any part of the image that is linear (eg an edge of a building wall, a hand rail, etc). After the Leica, in my eyes the Panasonic GH2 m4/3 gives a better image than Nikon or Sony for A2 size! To me, the results are very surprising. I would say the digital wars will be furious in 2011 and forwards. But the Foveon seems to hold the most promise, if Sigma can get the imager to full frame and improved sensitivity.

**fs999** · 04-21-2011, 12:26 AM

With the K-5 Pentax has showed us that the size of the sensor is not the matter...

stanleyk · 04-21-2011, 03:29 PM

I also use a DP1 and DP2 and plan on probably getting the SD1 when it's released. The images are really stunning when used properly. The foveon is particularly good for black and white conversions. If the SD1 had been released sooner I would have skipped the K5.

07-15-2010, 11:43 AM	#38
luck1 Forum Member Join Date: Oct 2009 Location: Pezinok Photos: Gallery Posts: 55	color shift found this on dpreview forum, to the color shift. SD15 colors compared to SD14 and 5D MkII: Sigma SLR Talk Forum: Digital Photography Review

04-20-2011, 11:11 PM	#43
robgrune Inactive Account Join Date: Apr 2011 Posts: 2	foveon is promising Based upon the large jpg files I downloaded from the various manufacturer websites (Leica, Panasonic, Sigma, Sony, Nikon), I can discern no difference to the images on-screen at about the A4 size, and no differences when printed at A4. At A4 size, the Foveon imager (4/3 frame) gives the best results for colour, contrast, and detail! But, going larger, to say A2 size and up, the Leica S2 (imager larger than full frame) gives the best image on screen, with Leica M9 coming second. The Foveon colour and contrast remain good, but the noise results in hard, serrated edges for any part of the image that is linear (eg an edge of a building wall, a hand rail, etc). After the Leica, in my eyes the Panasonic GH2 m4/3 gives a better image than Nikon or Sony for A2 size! To me, the results are very surprising. I would say the digital wars will be furious in 2011 and forwards. But the Foveon seems to hold the most promise, if Sigma can get the imager to full frame and improved sensitivity.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K-x Sony Sensor vs K-7 Samsung Sensor	karl79	Video Recording and Processing	9	09-23-2010 09:35 AM
the sensor in k-x	techmulla	Pentax DSLR Discussion	1	05-23-2010 08:05 PM
K-7 sensor vs K20d sensor	Mystic	Pentax News and Rumors	33	06-21-2009 03:01 AM
Sensor cleaning: Pec-Pads or Sensor Swabs	gadgetnu	Pentax Camera and Field Accessories	29	09-24-2007 10:52 AM
Sensor cleaning > Sensor Swab > void warranty?	Twinky	Pentax DSLR Discussion	2	07-28-2007 01:10 AM

07-09-2010, 05:35 AM - 2 Likes	#31
Ivan Glisin Veteran Member Join Date: Nov 2006 Location: Belgrade Posts: 656	Originally posted by Winder I would like to see a monochrome K-7. No Bayer filter. The resolution of a true 14MP monochrome sensor would be amazing. I second this. A K-7GS (grayscale) perhaps? One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer). However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one. (And I still have my color filters for BW photography!)

07-09-2010, 06:33 AM	#32
ghelary Veteran Member Join Date: Jul 2008 Location: Paris, France Photos: Gallery Posts: 613	Originally posted by Ivan Glisin I second this. A K-7GS (grayscale) perhaps? One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer). However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one. (And I still have my color filters for BW photography!) Yup, I've been thinking of that myself too, excepted for the sensitivity part. Basically, filter will make you loose the gain in sensitivy obtain by removing the bayer matrix. But, hey, nearly no more AA filter needed !

07-09-2010, 06:42 AM	#33
Rico Veteran Member Join Date: Mar 2007 Posts: 1,205	Originally posted by falconeye I manually whitebalanced both images to the white tent on the left hand side and shifted colors in the K-7 sample to get X3's wrong blue sky effect (I exxagerated a little bit. In lightroom, you mostly draw the Aqua, Blue and Magenta channel sliders to -20 or -40 and push the Blue channel towards a Green hue). The result is attached. Except for wrong colors, I cannot see a difference. However, I wonder which X3 color calibration gives such wrong colors... Hi Falk I have to say I set the images up pretty quickly. I changed the color mode in the image below from neutral to landscape which makes a big difference. The color adjustment was about the same. Whitebalance was kept the same too Sunlight. Funny there is a setting for original under color mode that the Sigma Pro software will not allow me to choose for whatever reason. I think this image looks better than the first X3 image I posted: Here is one of the first images I took with the DP2s. It is not a macro camera and it was windy but I think it demonstrates the potential of Foveon X3. There is a pop and a clarity to the image similar to Kodachrome IMHO:

07-09-2010, 09:38 AM	#34
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by Rico it demonstrates the potential of Foveon X3. There is a pop and a clarity to the image similar to Kodachrome IMHO: I see what you mean. But we must be careful with assessments like this one. The K-7 image has a relatively linear and accurate color perfomance while the X3 images (the ones you show) have a huge CIE deltaE error (i.e., even if white balance is correct, some colors are simply way off). I tried to "emulate" this for the K-7 images which means to suppress some color sub channels significantly. And while the colors shift, the image gains significantly in "pop". It's like putting a filter in B&W photography. So, what you see is an "effect". You may say it's a nice effect because it creates pop. But for sure, it delivers wrong colors too. We all know that from the days of analog film. To tell the truth, I prefer to be in control of such effects, e.g., by using DxO film emulation package and the clarity slider.

07-10-2010, 09:20 AM	#35
Winder Veteran Member Join Date: Apr 2010 Location: Tennessee Posts: 6,617	Originally posted by Ivan Glisin I second this. A K-7GS (grayscale) perhaps? One thing I keep thinking about regarding this design is how much sensitivity improve and resolving power would improve. Remove Bayer filter and each photosite (in fact, now a true pixel in a grayscale sensor) would get much more light, at least two stops is my guess (depends on how narrow are RGB photosite filters), and since there is no interpolation, resolving power would easily double, which is in fact already demonstrated with Foveon X3 vs. Bayer pattern sensors, (4.6MP X3 comparable 10MP Bayer). However, the native base ISO would go to ISO 400 (or more) from ISO 100 assuming the same sensor. And here is the thing: why the same sensor? A much more efficient sensor could be made for, say, ISO 50-3200 with SIGNIFICANTLY better signal to noise ratio. A dream machine for superclean and supersharp BW images! I think there is a market for such camera, and I would certainly consider getting one. (And I still have my color filters for BW photography!) I have seen several images from monochrome digital MF backs and the detail in textures the tonal range was very impressive. I'm still holding out for a 5/4 format monochrome sensor......

07-14-2010, 12:05 PM	#36
glanglois Veteran Member Join Date: Aug 2007 Location: Madison, Wis., USA Posts: 1,506	What are the other implications of a B&W camera? For example, what are the results of not needing white balance, chroma NR, adjustments for hue/saturation, or color-sensitive scene modes? What is result (if any) upon AF if we're using color filters in front of the lens? Does this make Ricoh's idea look better, swapping lens and sensor as a package? Does this lead to camera design in which the sensor can be changed like film or a film back? I have to go to think now ....

07-14-2010, 06:16 PM	#37
deejjjaaaa Veteran Member Join Date: May 2007 Location: steel city / rust belt Posts: 2,043	Originally posted by Rico Hi Falk I have to say I set the images up pretty quickly. Rico, you do a typical newbie error - you raw conversion using SPP gets cyan poisoning... use SilkyPix, it will not let such cyan tint to happen.

07-16-2010, 01:40 AM	#39
ddhytz Veteran Member Join Date: Sep 2007 Location: Melbourne Posts: 602	I have a DP1, a camera I carry with me most of time. Some sample shots: 100% crop

07-16-2010, 01:44 AM	#40
ddhytz Veteran Member Join Date: Sep 2007 Location: Melbourne Posts: 602	More sample shots

02-27-2011, 12:21 AM	#41
yucatanPentax Loyal Site Supporter Join Date: Dec 2009 Location: Dallas / Yucatan Posts: 1,841	ddhytz, Those are certainly nice captures. It seems the color balance isn't always off. Your shots show a fairly impressive amount of detail and seemingly accurate colors. I don't know enough about the technology to know why the colors / detail/ contrast may seem off at times and not at others. But I do like your shots!

03-08-2011, 01:32 AM	#42
stillshunter Veteran Member Join Date: Nov 2009 Location: Southern Tablelands NSW Photos: Gallery \| Albums Posts: 406	Guys, I have recently bought a DP2s as my 'back-up' compact. Can I say that I am blown away each and every time I see the outputs of the lil Sigma. To be honest, I now find myself reaching for it 9 times out of 10 over the K20D. There really is something to the Foveon files... something that is lacking from the Pentax images as well as the Nikon, and Olympus ones that preceded it. To be honest I am tempted by the SD1. Oh who am I kidding I am VERY VERY tempted! For those who are working from specs alone, I say please defer your comments until you've actually tried the Foveon yourself. You - like me - might just be very pleasantly surprised.

04-21-2011, 12:26 AM	#44
fs999 Loyal Site Supporter Join Date: Jul 2008 Location: Luxembourg Photos: Gallery Posts: 8,645	With the K-5 Pentax has showed us that the size of the sensor is not the matter...

04-21-2011, 03:29 PM	#45
stanleyk Veteran Member Join Date: Mar 2009 Location: Taylor, Texas Posts: 1,018	I also use a DP1 and DP2 and plan on probably getting the SD1 when it's released. The images are really stunning when used properly. The foveon is particularly good for black and white conversions. If the SD1 had been released sooner I would have skipped the K5.