How to combine mono and color SPP images for sharpening control

[xpatUSA]

[...]

This puts the putative blue channel response just slightly over the green channel. In reality it's a stop down. The red channel response is also overstated here.

And reality matches Sigma's approach with the Quattro sensor, using larger pixels in lower layers to compensate for weaker signal.

A larger pixel does not "compensate for weaker signal". For a given exposure, a larger pixel has a larger capacitance proportional to the larger area therefore the voltage output does not change.

The read noise is similar, but the photon noise SNR is greater. Four times the electron count means double the SNR.

That is why the exposure does not have to be changed when switching a Foveon-based camera from hi to lo resolution (2x2 binning).

Do you think the QE numbers don't make sense because they are normalized to the amount of signal passed by the layer(s) above ?

Not sure I understand that, Jim. Sorry.

What I'm thinking is that the sensitivity curves are for the photodiodes, not the sensor. The top layer gets all the light, so QE has its normal meaning. But the next layer down only gets the light passed by the top layer. If we're talking about the QE of the photodiode in the second layer, that's one thing, and could explain the values higher than what we're seeing in the raw files. If we're talking about the QE of the second layer of the sensor, we have to multiply by the absorption spectrum of the top layer.

Agreed - I just now edited to say much the same thing.

 

[JimKasson]

[...]

This puts the putative blue channel response just slightly over the green channel. In reality it's a stop down. The red channel response is also overstated here.

And reality matches Sigma's approach with the Quattro sensor, using larger pixels in lower layers to compensate for weaker signal.

A larger pixel does not "compensate for weaker signal". For a given exposure, a larger pixel has a larger capacitance proportional to the larger area therefore the voltage output does not change.

The read noise is similar, but the photon noise SNR is greater. Four times the electron count means double the SNR.

That is why the exposure does not have to be changed when switching a Foveon-based camera from hi to lo resolution (2x2 binning).

Do you think the QE numbers don't make sense because they are normalized to the amount of signal passed by the layer(s) above ?

Not sure I understand that, Jim. Sorry.

What I'm thinking is that the sensitivity curves are for the photodiodes, not the sensor. The top layer gets all the light, so QE has its normal meaning. But the next layer down only gets the light passed by the top layer. If we're talking about the QE of the photodiode in the second layer, that's one thing, and could explain the values higher than what we're seeing in the raw files. If we're talking about the QE of the second layer of the sensor, we have to multiply by the absorption spectrum of the top layer.

Agreed - I just now edited to say much the same thing.

Great.

 

[BuhI]

Thank you for the answers, Jim.

 

[Johan Borg]

[...]

This puts the putative blue channel response just slightly over the green channel. In reality it's a stop down. The red channel response is also overstated here.

And reality matches Sigma's approach with the Quattro sensor, using larger pixels in lower layers to compensate for weaker signal.

A larger pixel does not "compensate for weaker signal". For a given exposure, a larger pixel has a larger capacitance proportional to the larger area therefore the voltage output does not change.

The read noise is similar, but the photon noise SNR is greater. Four times the electron count means double the SNR.

That is why the exposure does not have to be changed when switching a Foveon-based camera from hi to lo resolution (2x2 binning).

Do you think the QE numbers don't make sense because they are normalized to the amount of signal passed by the layer(s) above ?

Not sure I understand that, Jim. Sorry.

What I'm thinking is that the sensitivity curves are for the photodiodes, not the sensor. The top layer gets all the light, so QE has its normal meaning. But the next layer down only gets the light passed by the top layer. If we're talking about the QE of the photodiode in the second layer, that's one thing, and could explain the values higher than what we're seeing in the raw files. If we're talking about the QE of the second layer of the sensor, we have to multiply by the absorption spectrum of the top layer.

Agreed - I just now edited to say much the same thing.

Great.

Isn't it ironic that when the two of you finally agree, you're both wrong?

Just kidding, but let's for once *not* "Ignore the marketing black line" in this previously shared image:

Loading…

kronometric.org

The black Total QE line appears to be the sum of the three layers, which makes no sense at all if they don't receive the same light. Sure, the black line makes little sense to start with, but still...

 

[JimKasson]

[...]

This puts the putative blue channel response just slightly over the green channel. In reality it's a stop down. The red channel response is also overstated here.

And reality matches Sigma's approach with the Quattro sensor, using larger pixels in lower layers to compensate for weaker signal.

A larger pixel does not "compensate for weaker signal". For a given exposure, a larger pixel has a larger capacitance proportional to the larger area therefore the voltage output does not change.

The read noise is similar, but the photon noise SNR is greater. Four times the electron count means double the SNR.

That is why the exposure does not have to be changed when switching a Foveon-based camera from hi to lo resolution (2x2 binning).

Do you think the QE numbers don't make sense because they are normalized to the amount of signal passed by the layer(s) above ?

Not sure I understand that, Jim. Sorry.

What I'm thinking is that the sensitivity curves are for the photodiodes, not the sensor. The top layer gets all the light, so QE has its normal meaning. But the next layer down only gets the light passed by the top layer. If we're talking about the QE of the photodiode in the second layer, that's one thing, and could explain the values higher than what we're seeing in the raw files. If we're talking about the QE of the second layer of the sensor, we have to multiply by the absorption spectrum of the top layer.

Agreed - I just now edited to say much the same thing.

Great.

Isn't it ironic that when the two of you finally agree, you're both wrong?

Just kidding, but let's for once *not* "Ignore the marketing black line" in this previously shared image:

https://kronometric.org/phot/sensor/fov/F13/f13-QE-400-to-1000nm.jpg

The black Total QE line appears to be the sum of the three layers, which makes no sense at all if they don't receive the same light.

Can't argue with that, but that doesn't make us wrong.

Sure, the black line makes little sense to start with, but still...

 

[Stillton]

In two recent threads, we determined that with the DP1 Merrill, you couldn't turn sharpening completely off in SPP in color mode, but you could in monochromatic mode.

For those who want the sharpness of the monochromatic mode for their color images, here's a technique that will do the job in Photoshop.

1. Save a color version from SPP as a TIFF.

2. With sharpening turned to minimum, save a monochromatic version from SPP as a TIFF.

3. Load both files into Photoshop.



4. Convert both files to Lab.


Both files must be 8 bits/channel or both files must be 16 bits/channel

5. In the channels control for both images, make only L visible.



6. Select all in the monoo image.



7. Copy the channel.

8. Paste the channel in the L channel of the color image.



9. Make all channels in the color image visible.

10. Convert the color image back to RGB.



11 Save the image.

It should be possible to create a script/action in Photoshop so this luminosity channel replacement is a one button operation.

 

[JimKasson]

It should be possible to create a script/action in Photoshop so this luminosity channel replacement is a one button operation.

True.

I now favor the layer luminosity blending approach over the one I proposed. It's simpler, more or less equivalent, and offers opportunities for tweaking by varying the opacity of the mono layer and using layer masks on it.

 

[saltydogstudios]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

If you cannot, wb'ing one may not achieve the same result as mapping to the other and wb'ing that.

True, but I WB'd only a neutral, and measured only a neutral.

It's a minor annoyance of Foveon sensors in my experience that you can't really change the white balance after the fact.

I always set a custom white balance in camera (when it makes sense) because it seems you're stuck with either the custom white balance from camera or a preset in SPP - but it's not like you can enter Kelvin values in SPP, which I always thought was odd. (If I'm remembering correctly).

I never quite understood why you couldn't select Kelvin values in SPP, but basically you can't shoot a neutral target and then use it to custom white balance later, you have to use the neutral target to set custom white balance in camera.

I'm reasonably sure this is also true of the DNG files produced by Quattro cameras - there's some non-reversible transformations that happen to produce the DNG. Though this has been debated in this forum - others believe that the DNG file can produce anything the X3F file could produce and that the transformation is non-destructive/completely reversible.

Sure there's more latitude in editing DNG files in something like Lightroom, but I think you're still stuck with the white balance set in camera - you could also change white balance in a TIFF or JPG in editing software, but you don't have the same leeway as a RAW file.... Though it's not like I've tested this.

I would expect both the 3 layers and the SPP output, to be completely maxed out or empty when photographing a black/white area (except close to the edge, noise etc).

If you've set up your slanted edge testing right, that won't happen. In fact, if it does happen then your slanted edge results will be inaccurate. Imatest will detect this condition and flag it for you. The image I used does not have that problem.

But it is a Foveon-sensor, who (outside Sigma) knows if one layer fills before the two others when seeing a pure white?

I looked at the raw layers, and none were saturated in the slanted edge part of the scene.

If it does not, WB'ing should not be needed with correct exposure (unless sensibility is non-linear so greyscales are not equally grey, but (dis)coloured); if it does, WB'ing would scale layers, reducing information from some of them, and/or exagerating others. If the WB-algorithm is not linear, or if the sensibility of the layers is not linear, how do we know what we are comparing when comparing a sharpness-function applied to the results?

I am not saying you cannot do it, I am saying that I do not understand why you can do it.

In fact, what I am really saying is I know too little, and want to know too much

And if that is so, whitebalancing the 3 layers directly and whitebalancing the (r-g-b-like, I suppose) internal representation in SPP may not give the same rounding errors/precision. But I am not sure - while your method is laid out, the internals of SPP are not, it may well work with raw layers internally until the last step, and then the difference between the methods is likely small; and I do not know how much is lost in precision inside SPP (and if all filters/adjustments are always applied), and how much is lost if you save intermediate files in your process).

When I apply the compromise matrix, I use 64-bit floating point. Plenty of precision to go around.

And that is the only thing you do? And then save it in 16-bit (or above)?

Yep.

Then I'll suffice with that.

(The background of all that being I would like the SPP process and the comparison process to be as similar as possible, so the difference in results can be attributed SPPs intentionally changed behavior, and not to differences due to discrete computation messing up. I think I forgot to mention that in the start of my former post - sorry)

--
"no one should have a camera that can't play Candy Crush Saga."
https://www.instagram.com/sodiumstudio/
Camera JPG Portrait Shootout http://sodium.nyc/blog/2020/05/camera-jpg-portrait

 

[JimKasson]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

 

[xpatUSA]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

If you cannot, wb'ing one may not achieve the same result as mapping to the other and wb'ing that.

True, but I WB'd only a neutral, and measured only a neutral.

It's a minor annoyance of Foveon sensors in my experience that you can't really change the white balance after the fact.

I always set a custom white balance in camera (when it makes sense) because it seems you're stuck with either the custom white balance from camera or a preset in SPP - but it's not like you can enter Kelvin values in SPP, which I always thought was odd. (If I'm remembering correctly).

Apropos of which, this might be of interest:

Loading…

kronometric.org

 

[saltydogstudios]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

The photos at the top of this thread are good enough for my purposes.

https://www.dpreview.com/forums/thread/4723273

Thanks for the offer - if anyone else is interested I can let them chime in.

 

[xpatUSA]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

The photos at the top of this thread are good enough for my purposes.

https://www.dpreview.com/forums/thread/4723273

Thanks for the offer - if anyone else is interested I can let them chime in.

I've measured a few Foveons. Here's one for an SD9 (original F7 sensor):



FWIW

 

[JimKasson]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

The photos at the top of this thread are good enough for my purposes.

https://www.dpreview.com/forums/thread/4723273

I don't think they say anything at all about the accuracy of Foveon colors.

Thanks for the offer - if anyone else is interested I can let them chime in.

 

[saltydogstudios]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

The photos at the top of this thread are good enough for my purposes.

https://www.dpreview.com/forums/thread/4723273

I don't think they say anything at all about the accuracy of Foveon colors.

For my purposes, accuracy isn't too interesting to me - at the end of the day I just want to take photos that I like. A camera can be less accurate but have a more pleasing output.

I have a theoretical interest in - say - how the dyes in a CFA contribute to color interpretation. And I have suspicions on how each manufacturer tunes the dyes to arrive at that camera's colors.

But I don't know what I would do with the accuracy data. I'm not doing fine art reproduction.

If you have any links to research I can read up on - especially comparing different cameras or how the dyes in the CFA contribute to color interpretation I'd be interested in reading up.

I suspect Nikon tuned their dyes to - say - produce pleasing skin tones in natural light situations (based on personal experience), and Canon maybe tuned their dyes for more realistic sky colors (based on rumors I read), and there's a bit more color variation in skin tones in Canon cameras as a result (which I prefer for studio photography as a base to start editing from).

But honestly I have no idea what their intention was, what their dyes actually were designed to do and how that affects the final image beyond just my personal experience of working with these cameras.

So - without benchmarks/comps to other cameras I'm not sure a "Foveon color accuracy" study would mean much to me, and even if I had those comps - I'm still going to reach for the cameras I reach for in various situations based on personal experience of working with those cameras/files.

Thanks for the offer - if anyone else is interested I can let them chime in.

 

[JimKasson]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

The photos at the top of this thread are good enough for my purposes.

https://www.dpreview.com/forums/thread/4723273

I don't think they say anything at all about the accuracy of Foveon colors.

For my purposes, accuracy isn't too interesting to me - at the end of the day I just want to take photos that I like. A camera can be less accurate but have a more pleasing output.

I have a theoretical interest in - say - how the dyes in a CFA contribute to color interpretation. And I have suspicions on how each manufacturer tunes the dyes to arrive at that camera's colors.

But I don't know what I would do with the accuracy data. I'm not doing fine art reproduction.

If you have any links to research I can read up on - especially comparing different cameras or how the dyes in the CFA contribute to color interpretation I'd be interested in reading up.

I suspect Nikon tuned their dyes to - say - produce pleasing skin tones in natural light situations (based on personal experience), and Canon maybe tuned their dyes for more realistic sky colors (based on rumors I read), and there's a bit more color variation in skin tones in Canon cameras as a result (which I prefer for studio photography as a base to start editing from).

But honestly I have no idea what their intention was, what their dyes actually were designed to do and how that affects the final image beyond just my personal experience of working with these cameras.

So - without benchmarks/comps to other cameras I'm not sure a "Foveon color accuracy" study would mean much to me, and even if I had those comps - I'm still going to reach for the cameras I reach for in various situations based on personal experience of working with those cameras/files.

Thanks for the offer - if anyone else is interested I can let them chime in.

I can supply benchmarks.

One thing you should consider: cameras that are less accurate tend to have more capture metameric error:

Loading…

blog.kasson.com

 

[DavidMillier]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply. It will take a morning (or maybe a whole day) to do the work, and I don't want to waste my time if nobody's interested.

Yes please

 

[xpatUSA]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply.

Curves I posted are for two sensor models, F7 and F13, except the one marked as "Merrill" which was extrapolated somehow from F13 data.

F7 has two versions - one for the SD9 without micro-lenses and one for the SD10 with micro-lenses. I've read that the F13 for one or more DP models has radially offset micro-lenses.

F13 is for all subsequent cameras up to but not including the Merrill models. There are no published Sigma/Foveon curves to my knowledge for the Merrill F20 sensor. Here's a Marketing one for the Quattro which claims to be the same as for the Merrill:

Obviously without the hot mirror and normalized to the top layer.

Yes please

 

[xpatUSA]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

Early Sigma cameras prior to the Merrills have a camera-to-XYZ 3x3 matrix which passed to the raw file as meta-data.

A very early one taken from a paper:

XYZ white ref is probably D55 but they don't actually say.

 

[JimKasson]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

That is correct. The Foveon sensors do not pass the Luther-Ives criterion.

But is it better than CFA sensors? Foveons seem better than at least Nikons at yellow and violet - there have been a couple of threads about this here and in the Nikon Z forum.

We've talked a lot about sharpening, but not much about color accuracy. In general I think color isn't talked about enough in the camera community.

Would you like me to do an analysis of the accuracy of Foveon colors? Thanks to Ted, I now have the sensitivity curves, but I'm not sure for what camera they apply.

Curves I posted are for two sensor models, F7 and F13, except the one marked as "Merrill" which was extrapolated somehow from F13 data.

F7 has two versions - one for the SD9 without micro-lenses and one for the SD10 with micro-lenses. I've read that the F13 for one or more DP models has radially offset micro-lenses.

F13 is for all subsequent cameras up to but not including the Merrill models. There are no published Sigma/Foveon curves to my knowledge for the Merrill F20 sensor. Here's a Marketing one for the Quattro which claims to be the same as for the Merrill:

Obviously without the hot mirror and normalized to the top layer.

Yes please

Thanks, Ted. That helps, assuming the sensor stack is the same.

--
https://blog.kasson.com

 

[JimKasson]

And this (not your reaching a higher level, but the large overlaps of layer-response) is the reason I claimed that the colours are not (purely) additive in one of the now filled-up threads:

Look at the 1931 CIE XYZ curves. They are purely additive, and there is considerable overlap.

Thanks for responding, and for having the patience, Jim. Maybe I am formulating something else than what I am after in an attempt to simplify my homemade english. Let me retry, sorry if tedious:

It is not clear to me that one can get to the 1931 CIE-curves by doing a linear transformation of the responses of the foveon-chip.

Early Sigma cameras prior to the Merrills have a camera-to-XYZ 3x3 matrix which passed to the raw file as meta-data.

A very early one taken from a paper:

XYZ white ref is probably D55 but they don't actually say.

Good. But I'll bet you dollars to doughnuts you don't get 1931 CIE XYZ curves if you use a monochromator on the camera and pass the results through that matrix.

--

the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.

Photography meets digital computer technology. Photography wins -- most of the time.

blog.kasson.com