Perception, reality and a signal below the noise...

[spider-mario]

The thing that is important in my mind is the equation, which I mentioned in the original post, (https://www.dpreview.com/forums/post/64327430):

output = signal + noise

This equation has a meaning for a single measurement in the case of an external noise source - measurement noise in our case, say for e.g. read noise, irrespective of whether we can measure it or not. We know this noise is there.

But, it has no meaning for the internal noise, shot noise here. Because, the value you hold for a single measurement is the truth value for the internal state of the overall system at that time. It is the ground truth in that measurement interval. There is no notion of noise, IMHO, in a single measurement here. So we can't say that in a single measurement that the shot noise is there and we just can't measure it. The value you have is the truth value.

But, I guess the discussion is becoming more philosophical .

Indeed, because I guess the question is basically whether we consider quantum randomness as internal or not. Couldn’t we say that our multiple measurements can be across time, across space, or across possible universes?

 

[Joofa]

The thing that is important in my mind is the equation, which I mentioned in the original post, (https://www.dpreview.com/forums/post/64327430):

output = signal + noise

This equation has a meaning for a single measurement in the case of an external noise source - measurement noise in our case, say for e.g. read noise, irrespective of whether we can measure it or not. We know this noise is there.

But, it has no meaning for the internal noise, shot noise here. Because, the value you hold for a single measurement is the truth value for the internal state of the overall system at that time. It is the ground truth in that measurement interval. There is no notion of noise, IMHO, in a single measurement here. So we can't say that in a single measurement that the shot noise is there and we just can't measure it. The value you have is the truth value.

But, I guess the discussion is becoming more philosophical .

Indeed, because I guess the question is basically whether we consider quantum randomness as internal or not. Couldn’t we say that our multiple measurements can be across time, across space, or across possible universes?

The quantum randomness here (shot noise) is different from measurement noise (read noise). And, that is the gist. In a more general setting, the overall notion of internal vs external noises has important consequences in the branch of math known as stochastic calculus , which breaks an important rule of differentiation in regular calculus, such as the chain rule . At least the Ito version . Though not the Stratonovich version . But, mathematicians favor Ito's version in random processes.

And, you are right, multiple measurements across time and space (as in the ensemble of images, multiple images that is, on which everybody in this discussion seems to agree) would make sense.

--
Dj Joofa
http://www.djjoofa.com

 

[ThrillaMozilla]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

 

[Joofa]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

 

[ThrillaMozilla]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

 

[Joofa]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

 

[ThrillaMozilla]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

OK, first you tell me how many quanta were captured in each photosite.

 

[Joofa]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

OK, first you tell me how many quanta were captured in each photosite.

Isn’t that your job to figure out? You were making claims about potential sufficiency of single meaurements.

--
Dj Joofa
http://www.djjoofa.com

 

[ThrillaMozilla]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

OK, first you tell me how many quanta were captured in each photosite.

Isn’t that your job to figure out? You were making claims about potential sufficiency of single meaurements.

No, and no I wasn't.

 

[Joofa]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

OK, first you tell me how many quanta were captured in each photosite.

Isn’t that your job to figure out? You were making claims about potential sufficiency of single meaurements.

No, and no I wasn't.

Ok, I guess you had a change of opinion. Good.

 

[Erik Kaffehr]

Someone was discussing to buy Hasselblad X1D or Fujifilm GFX. One of the comments was:

'The X1D cameras just have a different look. It might be the massive color information that the camera is said to have in its firmware. It might be something in the way a "raw" file is generated. It might be the treatment of colors in the camera profile that Phocus has or that Adobe obtained from Hasselblad for Lightroom. Maybe one can wrestle an X1D raw file to look like a Fuji GFX file, but if you just start on the X1D raw in Phocus or even Lightroom, you typically have something of the camera "look" in most shots when you are done.'

That is a quite definitive statement. But do have cameras different color?

So, I downloaded DPReviews studio test images for both the X1D and the GFX 100, generated DCP Profiles for both with LumaRiver Profile Designer developed the Studio Test image with consistent exposure in Lightroom and analysed the ColorChecker colors using Babelcolor 'PatchTool'.

The color differences where like this:


The colors are split squares, I don't recall which side is wich, but they look pretty similar to me...

So, it seems both cameras are capable of producing exactly the same color.

on this chart...

That is true.

But, it is a decent indication that the sensors can produce same colors.

If my understanding is right, you used the same chart for profiling. Is so, this makes things even worse. The result you see has nothing to do with the sensors; it produces similar colors because this is what the software was designed to do, having a cheat sheet. They could have written the code so that you would get those colors from a B&W sensor as well.

If producing the same colors was so simple, Adobe would have done it, right, in one of their profiles instead of this?

Hi,

Yes and no. One thing is that I have actually made a large set of tests.

But, the page you refer to has three fundamental errors:

• The first is that photographs of skin color at best provide a metameric match. A combination of dyes that provide similar stimulus to the skin tones, but don't have a spectral match.
• The second is that DPReview has found out that they portrait prints are subject to significant fading. So, they change between shooting occasions.
• The third is that I presume that they are glossy and subject to glare, depending on positioning of the camera and lights.

A way to improve my comparison would be to use the normal ColorChecker as a 'learning set' and use another reference as 'evaluation set'. Such an evaluation set could be ColorChecker GS.

Unfortunately, I don't have access to a good set of ColorChecker SG shots Hasselblad X1D

A photograph of a portrait print is not a photograph of skin colours. The faces on the DPR test setup are worse than useless -- they are actively misleading.

The two skin colour squares on the Color Checker are probably closer. They seem to be mixed with iron oxide pigments with spectra resembling those of real skin. Colour prints are mixtures of YMCK dyes and have quite different spectra.

Here is a plot of reflectance of the 'light skin' compared with the average of a significant numbers of samples from my hands and forehead.




Note that there is some similarity, but the curves diverge above 600 nm. The 'ColorChecker SG' has several skin patches and they don't that rise above 600 nm.

The difference does not matter for calibration/evaluation as calibration just seeks parameters that reproduces the learning set.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[ThrillaMozilla]

But, If you still think that a single image does. Then please measure the shot noise in the single cat image that I posted earlier.

Please no more wording. Just report a number. If you can't do that without having multiple images, then that negates your assertion about shot noise in a single image.

Don't be ridiculous. I don't have access to your camera, and even if I did, I would not going to do the work needed to determine the number of quanta in each pixel. But you could do that.

Well, you are wiggling your way out of the situation you put yourself in. But it is OK as long as you understand what is being said here.

Well, OK, fine. Send me your camera. I'll make the necessary measurements but I get to keep it as payment.

Let me summarize the situation and be done with it:

Joofa claims that: "shot noise has no meaning or definition for a single measurement".

You don’t need my camera. You just need a single image, and it was provided to you. You should be able to download it.

OK, first you tell me how many quanta were captured in each photosite.

Isn’t that your job to figure out? You were making claims about potential sufficiency of single meaurements.

No, and no I wasn't.

Ok, I guess you had a change of opinion. Good.

I don't know what you're thinking, but on a forum that works on rapid, often informal communication, it's not necessary to treat every perceived inaccuracy as if it requires a hostile attack. You have a bad attitude.

I'm not going to discuss this any more with you.

 

[J A C S]

Here is a plot of reflectance of the 'light skin' compared with the average of a significant numbers of samples from my hands and forehead.


Note that there is some similarity, but the curves diverge above 600 nm. The 'ColorChecker SG' has several skin patches and they don't that rise above 600 nm.

The difference does not matter for calibration/evaluation as calibration just seeks parameters that reproduces the learning set.

Just by eyeballing, the CC "skin tone" should look reddish next to your hand. Is that what you see?

 

[Erik Kaffehr]

Here is a plot of reflectance of the 'light skin' compared with the average of a significant numbers of samples from my hands and forehead.


Note that there is some similarity, but the curves diverge above 600 nm. The 'ColorChecker SG' has several skin patches and they don't that rise above 600 nm.

The difference does not matter for calibration/evaluation as calibration just seeks parameters that reproduces the learning set.

Just by eyeballing, the CC "skin tone" should look reddish next to your hand. Is that what you see?

Cannot say, backside of the hand is probably less red, inside is more reddish.


These were all my skin samples.


And those are the spectrum plots. The low curve, I probably lifted the spectro during measurement.

What I see is that there is a sharp rise around 580 nm, both on the color checker and my skin samples. There is also a bump around 500 nm on the ColorChecker. There is a bump around 500 nm on my skin measurements, but it is broader.

There are bunch of extra 'skin samples' on the ColorChecker SG, that don't have the rise above 600 nm.





So, the ColorChecker light skin is more reddish than the sample I used from the ColorChecker SG. This is the only ColorChecker SG sample that shows the 500 nm hump, outside 'the original ColorChecker patches' that are part of the SG.

I think the 500 nm hump is a characteristic of caucasian skin. It seems absent from other skin plots I have seen, but I don't have seen a lot.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles