SMI (Sensitivity Metemerism Index) -- is it important?

[Mark S Abeln]

most colors come out quite right with proper WB. However, somewhat singular colors exist. These colours may fall far off the mark and get aliased with a wrong colour.
The cause is that cameras do not have a perfect implementation of human tristimulus sensititvities. This systematic colour error is different from the statistical colour errors discussed in this thread.

Here is an eaxample of what I mean: The colour checker in the image below is quite OK. But, the jacket had to be adjusted with layer and masking to the correct colour in the left section. The recorded colour coues out so far off the mark, that trying to fix this by WB adjustments and colour curves messes up everything else.

Right, I am doing a lot of masking and layers to get it right, although ideally I would like global corrections to do as much of the work as possible. My client, an interior decorator, wants specific colors accurate, and as these images will eventually be put into a nice glossy magazine, I can’t just do a rough job (but fortunately, I get paid more when that eventually comes around).

 

[Great Bustard]

Nikons with horrible colors have better SMI than Canons with excellent colors.

Let's just begin with this. I was unaware that Nikon had "horrible colors" whereas Canon had "excellent colors".

Someone had to bring it up for you

As this is the Photographic Science and Technology Forum, I am assuming this is based on some form of credible evidence, which I'd be pleased to see.

http://www.dpreview.com/forums/post/53046518

Is the example to support your claim in the link above serious or just a joke? Because if you're serious...

He is serious ok.

Then I would ask how two photos from two different scenes with different lighting, both not only using flash, but different forms of flash (built-in direct vs off-camera bounced), with unknown white balance and processing, demonstrates anything, whatsoever, about which camera has "better color".

That would be a good set of questions IMHO. May I add yet one more - how were the shots processed?

Both shots are of the same girl at two real life events with AWB on either camera. The Canon shoot made a nice memory; the whole Nikon shoot was a waste.

From where I sit, it looks like someone is misplacing the blame. Specifically, claiming that "the whole Nikon shoot was a waste" speaks volumes about the photographer, and nothing about Nikon.

Those without a blindfold will see what they need to see and I have no agenda to convince those who cannot be convinced. I bought D800 to upgrade my 5D, but instead the Nikon went back to the store.

And all those that stuck with the D800 are stupid to think the D800 was a rather nice upgrade from the 5D? The "evidence" you posted gives a rather different story.

The processing is Camera Default in ACR with no post processing, essentially the same as straight out of camera JPEGs.

Because all competent photographers shooting RAW go with the defaults.

The Canon flash was bounced off the wall, but I couldn't bounce the built-in flash in D800.

Because all competent photographers use off-camera bounced flash with one camera and direct on-camera flash with another camera.

Sorry, I am not comfortable providing raw files, because the girl is not a relative and I don't have a model release. I have nothing to prove, I simply answered a question about the basis for my opinion.

The context of your opinion is more telling than your opinion.

To make my point clear, I am not saying that Nikon is incapable of great skin tones, but it makes you work real hard for it in each and every shot while Canon does it all for you by default.

So you are saying that Nikon photographers have skills you don't? I don't have a problem with that.

As such, Nikon may be great for some pros and backyard technicians shooting resolution charts, but not for making family memories with the least effort.

Yeah, that must be it.

 

[Press Correspondent]

Nikons with horrible colors have better SMI than Canons with excellent colors.

Let's just begin with this. I was unaware that Nikon had "horrible colors" whereas Canon had "excellent colors".

Someone had to bring it up for you

As this is the Photographic Science and Technology Forum, I am assuming this is based on some form of credible evidence, which I'd be pleased to see.

http://www.dpreview.com/forums/post/53046518

Is the example to support your claim in the link above serious or just a joke? Because if you're serious...

He is serious ok.

Then I would ask how two photos from two different scenes with different lighting, both not only using flash, but different forms of flash (built-in direct vs off-camera bounced), with unknown white balance and processing, demonstrates anything, whatsoever, about which camera has "better color".

That would be a good set of questions IMHO. May I add yet one more - how were the shots processed?

Both shots are of the same girl at two real life events with AWB on either camera. The Canon shoot made a nice memory; the whole Nikon shoot was a waste.

From where I sit, it looks like someone is misplacing the blame. Specifically, claiming that "the whole Nikon shoot was a waste" speaks volumes about the photographer, and nothing about Nikon.

Those without a blindfold will see what they need to see and I have no agenda to convince those who cannot be convinced. I bought D800 to upgrade my 5D, but instead the Nikon went back to the store.

And all those that stuck with the D800 are stupid to think the D800 was a rather nice upgrade from the 5D? The "evidence" you posted gives a rather different story.

The processing is Camera Default in ACR with no post processing, essentially the same as straight out of camera JPEGs.

Because all competent photographers shooting RAW go with the defaults.

The Canon flash was bounced off the wall, but I couldn't bounce the built-in flash in D800.

Because all competent photographers use off-camera bounced flash with one camera and direct on-camera flash with another camera.

Sorry, I am not comfortable providing raw files, because the girl is not a relative and I don't have a model release. I have nothing to prove, I simply answered a question about the basis for my opinion.

The context of your opinion is more telling than your opinion.

To make my point clear, I am not saying that Nikon is incapable of great skin tones, but it makes you work real hard for it in each and every shot while Canon does it all for you by default.

So you are saying that Nikon photographers have skills you don't? I don't have a problem with that.

As such, Nikon may be great for some pros and backyard technicians shooting resolution charts, but not for making family memories with the least effort.

Yeah, that must be it.

A bunch of pointless excuses with an attitude. Like I said, those who can see will see for themselves what they need to see while those with a blind attitude will never be convinced in the obvious. No clue what you are trying to prove here. That Nikon has "great colors"? Well, everyone deserves the choice he makes. Good luck!

 

[Great Bustard]

Nikons with horrible colors have better SMI than Canons with excellent colors.

Let's just begin with this. I was unaware that Nikon had "horrible colors" whereas Canon had "excellent colors".

Someone had to bring it up for you

As this is the Photographic Science and Technology Forum, I am assuming this is based on some form of credible evidence, which I'd be pleased to see.

http://www.dpreview.com/forums/post/53046518

Is the example to support your claim in the link above serious or just a joke? Because if you're serious...

He is serious ok.

Then I would ask how two photos from two different scenes with different lighting, both not only using flash, but different forms of flash (built-in direct vs off-camera bounced), with unknown white balance and processing, demonstrates anything, whatsoever, about which camera has "better color".

That would be a good set of questions IMHO. May I add yet one more - how were the shots processed?

Both shots are of the same girl at two real life events with AWB on either camera. The Canon shoot made a nice memory; the whole Nikon shoot was a waste.

From where I sit, it looks like someone is misplacing the blame. Specifically, claiming that "the whole Nikon shoot was a waste" speaks volumes about the photographer, and nothing about Nikon.

Those without a blindfold will see what they need to see and I have no agenda to convince those who cannot be convinced. I bought D800 to upgrade my 5D, but instead the Nikon went back to the store.

And all those that stuck with the D800 are stupid to think the D800 was a rather nice upgrade from the 5D? The "evidence" you posted gives a rather different story.

The processing is Camera Default in ACR with no post processing, essentially the same as straight out of camera JPEGs.

Because all competent photographers shooting RAW go with the defaults.

The Canon flash was bounced off the wall, but I couldn't bounce the built-in flash in D800.

Because all competent photographers use off-camera bounced flash with one camera and direct on-camera flash with another camera.

Sorry, I am not comfortable providing raw files, because the girl is not a relative and I don't have a model release. I have nothing to prove, I simply answered a question about the basis for my opinion.

The context of your opinion is more telling than your opinion.

To make my point clear, I am not saying that Nikon is incapable of great skin tones, but it makes you work real hard for it in each and every shot while Canon does it all for you by default.

So you are saying that Nikon photographers have skills you don't? I don't have a problem with that.

As such, Nikon may be great for some pros and backyard technicians shooting resolution charts, but not for making family memories with the least effort.

Yeah, that must be it.

A bunch of pointless excuses with an attitude.

I was thinking the same thing.

Like I said, those who can see will see for themselves what they need to see while those with a blind attitude will never be convinced in the obvious. No clue what you are trying to prove here.

I think you have that rather backwords. You said, and I quote:

http://www.dpreview.com/forums/post/53117138

The SMI shows how close the color filters in the sensor are close to the color filters in the human eye. In theory the closer the better, but in practice it does not hold true. Nikons with horrible colors have better SMI than Canons with excellent colors. If you look at the DXO numbers for SMI of popular DSLRs, there is no correlation with the practical color quality at all. You can safely ignore this measurement as completely meaningless in the practical sense.

Your "proof" said nothing about Nikon, but plenty about you.

That Nikon has "great colors"? Well, everyone deserves the choice he makes. Good luck!

I shoot Canon. I guess I missed where the colors from my photos were so much better than what those poor Nikon shooters were getting.

 

[MirekE]

and can be off by a lot. There must be a lot of constraints on the colours of the CFA concerning micro fabrication, precision, stability and transmission. Obvious problem areas are in the differentiation of the deep reds, and specially narrow pigments at the red-green transition, the yellows, and even more at the green-blue transition near 480nm. The eye has a dip in sensitivity there, seeing relatively dark blue-green. The CFA separation gets these colours much brighter and easily errs to one or the other side as compared to the eye.

I recently posted a quite striking example, (the dark green jacket)

http://www.dpreview.com/forums/post/53032750

Do you happen to have Color Munki or another spectrophotometer? It might be interesting to meter the jacket and look at the spectrum.

 

[Press Correspondent]

Nikons with horrible colors have better SMI than Canons with excellent colors.

Let's just begin with this. I was unaware that Nikon had "horrible colors" whereas Canon had "excellent colors".

Someone had to bring it up for you

As this is the Photographic Science and Technology Forum, I am assuming this is based on some form of credible evidence, which I'd be pleased to see.

http://www.dpreview.com/forums/post/53046518

Is the example to support your claim in the link above serious or just a joke? Because if you're serious...

He is serious ok.

Then I would ask how two photos from two different scenes with different lighting, both not only using flash, but different forms of flash (built-in direct vs off-camera bounced), with unknown white balance and processing, demonstrates anything, whatsoever, about which camera has "better color".

That would be a good set of questions IMHO. May I add yet one more - how were the shots processed?

Both shots are of the same girl at two real life events with AWB on either camera. The Canon shoot made a nice memory; the whole Nikon shoot was a waste.

From where I sit, it looks like someone is misplacing the blame. Specifically, claiming that "the whole Nikon shoot was a waste" speaks volumes about the photographer, and nothing about Nikon.

Those without a blindfold will see what they need to see and I have no agenda to convince those who cannot be convinced. I bought D800 to upgrade my 5D, but instead the Nikon went back to the store.

And all those that stuck with the D800 are stupid to think the D800 was a rather nice upgrade from the 5D? The "evidence" you posted gives a rather different story.

The processing is Camera Default in ACR with no post processing, essentially the same as straight out of camera JPEGs.

Because all competent photographers shooting RAW go with the defaults.

The Canon flash was bounced off the wall, but I couldn't bounce the built-in flash in D800.

Because all competent photographers use off-camera bounced flash with one camera and direct on-camera flash with another camera.

Sorry, I am not comfortable providing raw files, because the girl is not a relative and I don't have a model release. I have nothing to prove, I simply answered a question about the basis for my opinion.

The context of your opinion is more telling than your opinion.

To make my point clear, I am not saying that Nikon is incapable of great skin tones, but it makes you work real hard for it in each and every shot while Canon does it all for you by default.

So you are saying that Nikon photographers have skills you don't? I don't have a problem with that.

As such, Nikon may be great for some pros and backyard technicians shooting resolution charts, but not for making family memories with the least effort.

Yeah, that must be it.

A bunch of pointless excuses with an attitude.

I was thinking the same thing.

Like I said, those who can see will see for themselves what they need to see while those with a blind attitude will never be convinced in the obvious. No clue what you are trying to prove here.

I think you have that rather backwords. You said, and I quote:

http://www.dpreview.com/forums/post/53117138

The SMI shows how close the color filters in the sensor are close to the color filters in the human eye. In theory the closer the better, but in practice it does not hold true. Nikons with horrible colors have better SMI than Canons with excellent colors. If you look at the DXO numbers for SMI of popular DSLRs, there is no correlation with the practical color quality at all. You can safely ignore this measurement as completely meaningless in the practical sense.

Your "proof" said nothing about Nikon, but plenty about you.

That Nikon has "great colors"? Well, everyone deserves the choice he makes. Good luck!

I shoot Canon. I guess I missed where the colors from my photos were so much better than what those poor Nikon shooters were getting.

Not sure why the subject of poor Nikon skin tones is so touchy for you. It is a well known fact pointed out by many on this forum and self evident on the first try. The Sony sensor in D800e was attractive, but after the release of A7r, I no longer have any interest in Nikon. So let me summarize my view of Nikon along with all your defense: I don't care.

 

[bobn2]

So let me summarize my view of Nikon along with all your defense: I don't care.

Clearly no, since if you didn't care you wouldn't even bother to comment. However, you did choose to comment so clearly you care.

 

[Bernard Delley]

and can be off by a lot. There must be a lot of constraints on the colours of the CFA concerning micro fabrication, precision, stability and transmission. Obvious problem areas are in the differentiation of the deep reds, and specially narrow pigments at the red-green transition, the yellows, and even more at the green-blue transition near 480nm. The eye has a dip in sensitivity there, seeing relatively dark blue-green. The CFA separation gets these colours much brighter and easily errs to one or the other side as compared to the eye.

I recently posted a quite striking example, (the dark green jacket)

http://www.dpreview.com/forums/post/53032750

Do you happen to have Color Munki or another spectrophotometer? It might be interesting to meter the jacket and look at the spectrum.

unfortunately, I do not own such equipment. Googling after Color Munki, I found an interesting page on spectrophotometers: http://www.xrite.com/spectrophotometer

In the end, I might ask my scientist friends to measure it. They are super well equipped to do this. But, I would need to cut out a little piece of the jacket and I hate to think of the working hour that might easily passed with this. Especially, as my spectroscopist friend said " you already know how this property of the color can arise".

 

[Wayne Larmon]

I think you have that rather backwords. You said, and I quote:

http://www.dpreview.com/forums/post/53117138

The SMI shows how close the color filters in the sensor are close to the color filters in the human eye. In theory the closer the better, but in practice it does not hold true. Nikons with horrible colors have better SMI than Canons with excellent colors. If you look at the DXO numbers for SMI of popular DSLRs, there is no correlation with the practical color quality at all. You can safely ignore this measurement as completely meaningless in the practical sense.

Your "proof" said nothing about Nikon, but plenty about you.

That Nikon has "great colors"? Well, everyone deserves the choice he makes. Good luck!

I shoot Canon. I guess I missed where the colors from my photos were so much better than what those poor Nikon shooters were getting.

Not sure why the subject of poor Nikon skin tones is so touchy for you. It is a well known fact pointed out by many on this forum and self evident on the first try. The Sony sensor in D800e was attractive, but after the release of A7r, I no longer have any interest in Nikon. So let me summarize my view of Nikon along with all your defense: I don't care.

There is a thread going on in Luminious Landscape that might have some additional information. It starts as a discussion about a LL article on the development of the Phase One IQ250 CMOS medium format back. In particular about the struggle that Phase One's chief color scientist (not sure of his actual position) has gone through to make usable profiles for all the gnarly CMOS DLSR sensors that Capture One supports. And how he used that knowledge to guide the development of the IQ250 sensor. (So Phase One wouldn't make the same mistakes that DSLR designers made.) I'm summarizing the thrust of the article, not vouching for its accuracy.

Which is interesting but isn't the part I wanted to point to. This post that shows CFA spectrum plots of a Canon and a Nikon sensor is more germane to this discussion. Along with the analysis in this post.

The point being, I think, that there is more overlap in the red and green channels in Canon filters, with the result being that colors that are somewhat similar to Caucasian skin tones get mapped into displaying as what are perceived as pleasant Caucasian skin tones under a wider variety of illuminants. Other manufacturers, including Nikon, apparently didn't choose this path.

I don't know if this is a valid theory, but I thought that it might be worth looking at.

Wayne

 

[Press Correspondent]

So let me summarize my view of Nikon along with all your defense: I don't care.

Clearly no, since if you didn't care you wouldn't even bother to comment. However, you did choose to comment so clearly you care.

You wish, but Sony killed Nikon for me.

 

[bobn2]

So let me summarize my view of Nikon along with all your defense: I don't care.

Clearly no, since if you didn't care you wouldn't even bother to comment. However, you did choose to comment so clearly you care.

You wish, but Sony killed Nikon for me.

Still commenting, still caring.

 

[Usee]

So let me summarize my view of Nikon along with all your defense: I don't care.

Clearly no, since if you didn't care you wouldn't even bother to comment. However, you did choose to comment so clearly you care.

You wish, but Sony killed Nikon for me.

Still commenting, still caring.

--
Bob

If you drive through the dark and then you see a light,

wouldn't you be inclined to look at it if it's bright enough?

-

Could you sleep, if someone is permanently yelling?

-

Would you say, that you care, in those situations?

-

I prefer to read and write arguments to a theme.

If I'm annoyed because of the lack of arguments...


...you really want to look into the sun, or a you forced to?

...I give a hint - like in this case.

-

Do I care?

--
Envy is the highest form of recognition.
-
Stop to run, start to think.
-
Think twice - that doubles the fun!
-
Your world is as big, as Your mind.
-
Avoid to have only one point of view!
-
U see?

 

[Usee]

I think you have that rather backwords. You said, and I quote:

http://www.dpreview.com/forums/post/53117138

The SMI shows how close the color filters in the sensor are close to the color filters in the human eye. In theory the closer the better, but in practice it does not hold true. Nikons with horrible colors have better SMI than Canons with excellent colors. If you look at the DXO numbers for SMI of popular DSLRs, there is no correlation with the practical color quality at all. You can safely ignore this measurement as completely meaningless in the practical sense.

Your "proof" said nothing about Nikon, but plenty about you.

That Nikon has "great colors"? Well, everyone deserves the choice he makes. Good luck!

I shoot Canon. I guess I missed where the colors from my photos were so much better than what those poor Nikon shooters were getting.

Not sure why the subject of poor Nikon skin tones is so touchy for you. It is a well known fact pointed out by many on this forum and self evident on the first try. The Sony sensor in D800e was attractive, but after the release of A7r, I no longer have any interest in Nikon. So let me summarize my view of Nikon along with all your defense: I don't care.

There is a thread going on in Luminious Landscape that might have some additional information. It starts as a discussion about a LL article on the development of the Phase One IQ250 CMOS medium format back. In particular about the struggle that Phase One's chief color scientist (not sure of his actual position) has gone through to make usable profiles for all the gnarly CMOS DLSR sensors that Capture One supports. And how he used that knowledge to guide the development of the IQ250 sensor. (So Phase One wouldn't make the same mistakes that DSLR designers made.) I'm summarizing the thrust of the article, not vouching for its accuracy.

Which is interesting but isn't the part I wanted to point to. This post that shows CFA spectrum plots of a Canon and a Nikon sensor is more germane to this discussion. Along with the analysis in this post.

The point being, I think, that there is more overlap in the red and green channels in Canon filters, with the result being that colors that are somewhat similar to Caucasian skin tones get mapped into displaying as what are perceived as pleasant Caucasian skin tones under a wider variety of illuminants. Other manufacturers, including Nikon, apparently didn't choose this path.

I don't know if this is a valid theory, but I thought that it might be worth looking at.

Wayne

Yes, that is a interesting argument, which would also fit to the claim that a Foveon based camera can deliver nice transitions due to the great (color) overlap of it's three layers - if used with suited software.

There are always many variables in the game and a unsuited software can destroy everything a sensor provided.

However...

...if a sensor can't deliver, the software also can't except it is filling the missing gradations itself, which is also possible, but not really what I'm looking for.

 

[The Davinator]

Nikons with horrible colors have better SMI than Canons with excellent colors.

Let's just begin with this. I was unaware that Nikon had "horrible colors" whereas Canon had "excellent colors".

Someone had to bring it up for you

As this is the Photographic Science and Technology Forum, I am assuming this is based on some form of credible evidence, which I'd be pleased to see.

http://www.dpreview.com/forums/post/53046518

Is the example to support your claim in the link above serious or just a joke? Because if you're serious...

He is serious ok.

Then I would ask how two photos from two different scenes with different lighting, both not only using flash, but different forms of flash (built-in direct vs off-camera bounced), with unknown white balance and processing, demonstrates anything, whatsoever, about which camera has "better color".

I pointed this out before as well. Nothing can be ascertained by such a flawed comparison. It must be a joke.