Sony colors - A few more insights

[stevo23]

http://www.dpreview.com/forums/thread/3914561

He showed two examples. Both taken with an A7Rii. Both processed by him using C1. Both shots were taken a few days ago (i.e., in October). The shot with the "correct" greens was taken at 2:30 in the afternoon. The shot that has greens that are "too vivid, too strong, with too much yellow in them" was taken at 5:30 pm.

He also says that most images were bad. He also mentions orange car stop lights looking yellow. Obviously, he is not talking about one shot only and he is describing what I was describing before. He clearly says that WB cannot fix it. I can assure you that this is not me with a second account but it certainly looks that way.

If the reason you posted the link is to provide evidence that Sony shooters don't know much about the golden hour, then you might have a point.

I would call that the radioactive green hour. The golden one looks different with a non-Sony camera.

If you posted it as more "evidence" of Sony cameras' uniquely getting greens wrong, please forgive me for not being particularly impressed.

Another coincidence then...

You bring this subject up at the slightest hint that could agree with you even when it doesn't even come close. Reminds me of a 5 year old little brother. Stop, ok?

I will put you in my ignore list while this thread is active, at least. When you reach the legal drinking age, send me a PM.

Aw, tit for tat. That's cute.

Don't you know that no one is interested in your theory? Seriously, this is a profile issue.

 

[stevo23]

http://www.dpreview.com/forums/thread/3914561

He showed two examples. Both taken with an A7Rii. Both processed by him using C1. Both shots were taken a few days ago (i.e., in October). The shot with the "correct" greens was taken at 2:30 in the afternoon. The shot that has greens that are "too vivid, too strong, with too much yellow in them" was taken at 5:30 pm.

He also says that most images were bad. He also mentions orange car stop lights looking yellow. Obviously, he is not talking about one shot only and he is describing what I was describing before.

All well and good, but the one and only "bad" shot he linked to is evidence to me that he doesn't appear to have a good appreciation of the significant difference to be expected when shooting in such different light. Frankly, I would say that grass the color of his reference shot would be the wrong green for a shot taken at 5:30 on an October evening. Suffice it to say that, without any other evidence to support his concerns, I don't particularly trust the conclusions he's drawn about the other claimed color issues he's seeing.

So time to discredit someone who has no interest in advancing your theory - and he doesn't even mention you. A DPR writer told you to go away once and yet you still try hard to make non-friends.

By the way, wasn't your original complaint that the Sonys failed to pick up the yellows in foliage? You guys need to get your stories straight...

He clearly says that WB cannot fix it. I can assure you that this is not me with a second account but it certainly looks that way.

Again, he appears to be fairly unsophisticated in his processing skills and is reaching out to others for help, so I wouldn't give too much weight to his WB claim.

So you use his reply which essentially doesn't support your claim as proof that your theory is correct?

 

[J A C S]

As far as I'm concerned, the ONLY thing you've demonstrated (reinforced by your response here) is that a number of photographers aren't very skilled at processing and are too quick to blame the hardware when it's really a software and wetware issue.

And that their mistakes create surprisingly similar color artifacts...

BTW, I did not say that Sony greens lack yellows - they lack yellow hues, as in small variations. You can add yellow, and I posted such conversions but then you screw everything else and you still have few variations. His shot is not too yellow, it just has wrong and unappealing radioactive greens.

Taken from the thread linked above

iPhone shot in the golden hour, http://ncmns.wordpress.com/2014/10/10/what-time-is-it-in-nature-early-signs-of-fall/

 

[naththo]

I would agree that the Adobe Standard profile is miles off for the a7rII. They may get around to fixing that at some point. Camera Standard is very nice:


Camera Standard/Adobe Standard

This guy has made some nice profiles that I use:

http://www.piraccini.net/2011/02/profili-colore-sony-a900-per-adobe-lr.html

I haven't downloaded or tried the A7RII ones, but I have used the A7 profiles and tend to like them better than the Adobe supplied ones.

Yeah the Adobe or the standard camera profile for A7 was nightmare for me. Too much green for my liking. Custom profile looks like a winner to me and also solving problem using piraccini profile as well. The custom profile looks rather close to what Canon image represent. Interesting.

 

[fishy wishy]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

 

[knickerhawk]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

I think you mean "loon" (singular). In my first encounter with JACS on this topic two preceding threads ago, I presented him with crops of the GretagMacbeth chart and he failed to pick the Sony one (he actually picked a Canon). He went on to dismiss the chart as almost useless. See here and the subsequent posts between us.

 

[crashpc]

One can change some colors and wb to a point he likes it. In some Cases, I like default Sony colors more, in others I don't... But you cannot change it for every image at once. With different CFA and other HW things, you simply cannot compensate for something what is missing/in excess in the data, and in another shot therd is nothing to miss, because the sceen and light is "like that".

 

[jonrobertp]

unless you actually like Sony colours .

It is easier to use canon files, both jpg and raw....than Sony ...where you need to be a computer / post processing kind of person. Canon DPP is easy...at least the older version. PS is not easy for raw files.

My experience.

 

[J A C S]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

I think you mean "loon" (singular). In my first encounter with JACS on this topic two preceding threads ago, I presented him with crops of the GretagMacbeth chart and he failed to pick the Sony one (he actually picked a Canon). He went on to dismiss the chart as almost useless. See here and the subsequent posts between us.

Actually, I said that the chart was useless several times before that; and I did not pick a chart that I said was Sony - I said it looks Sony like, which is true.

 

[fmian]

Sorry but using the studio scene here isn't really a gauge of much as far as comparisons go.
It's terribly inconsistent in many ways and shouldn't be used to make any conclusions.

 

[stevo23]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

I think you mean "loon" (singular). In my first encounter with JACS on this topic two preceding threads ago, I presented him with crops of the GretagMacbeth chart and he failed to pick the Sony one (he actually picked a Canon). He went on to dismiss the chart as almost useless. See here and the subsequent posts between us.

Actually, I said that the chart was useless several times before that; and I did not pick a chart that I said was Sony - I said it looks Sony like, which is true.

Equivocation. You got fooled - yea, you're reading it.

 

[stevo23]

unless you actually like Sony colours .

It is easier to use canon files, both jpg and raw....than Sony ...where you need to be a computer / post processing kind of person. Canon DPP is easy...at least the older version. PS is not easy for raw files.

It's all in the profiles. If you choose the right profiles, it's the other way around.

 

[fishy wishy]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

I think you mean "loon" (singular). In my first encounter with JACS on this topic two preceding threads ago, I presented him with crops of the GretagMacbeth chart and he failed to pick the Sony one (he actually picked a Canon). He went on to dismiss the chart as almost useless. See here and the subsequent posts between us.

I think it possible that there is a difference in colours mainly based on the CFA filter chosen for the sensor. And I already know how colours can shift up the sensitivity scale.

But to dismiss a color chart as a means to check and calibrate colours? That is a new one on me. I noticed the following quote:

"that chart is just a visual representation of the pretty chart that Mako posted, and of limited value. A color profiling software could just replace the colors there with the reference ones; and in fact, a B&W shot would be enough for that. But this is painting, not taking a photo."

?? It just reads like a conspiracy theory to me, with the new science of turning b&w shots into colour, too.

 

[naththo]

Sorry but using the studio scene here isn't really a gauge of much as far as comparisons go.
It's terribly inconsistent in many ways and shouldn't be used to make any conclusions.

Sony A7 skies colour looks weird compare to old Nex 7 is not which is more neutral blue while A7 shift toward cyan/blue or more so bit too green for my liking. A7 also do not display rose flowers true red sometimes which is a problem both in JPEG and RAW. ACR/Lightroom DCP Profile still end up with posterization problem in any camera including Canon especially more significant in sunset picture for example. Still a flaw in ACR/Lightroom software.

 

[J A C S]

What on earth are you loons doing? If you want to judge the colours, do it with the color chart slap bang in the center of the shot itself, which could tell you at a glance. Posting a cropped shot of green on its own in an incorrectly profiled image is not the way to prove anything.

[shakes head]

I think you mean "loon" (singular). In my first encounter with JACS on this topic two preceding threads ago, I presented him with crops of the GretagMacbeth chart and he failed to pick the Sony one (he actually picked a Canon). He went on to dismiss the chart as almost useless. See here and the subsequent posts between us.

I think it possible that there is a difference in colours mainly based on the CFA filter chosen for the sensor. And I already know how colours can shift up the sensitivity scale.

But to dismiss a color chart as a means to check and calibrate colours? That is a new one on me. I noticed the following quote:

"that chart is just a visual representation of the pretty chart that Mako posted, and of limited value. A color profiling software could just replace the colors there with the reference ones; and in fact, a B&W shot would be enough for that. But this is painting, not taking a photo."

?? It just reads like a conspiracy theory to me, with the new science of turning b&w shots into colour, too.

This is what a non-linear optimization does, kinda. It may or may not exactly replace those colors because it depends on the way the optimization is done. In any case, the algorithm knows what those colors are and is trying to interpolate between them, roughly speaking.

 

[Ed the Bruce]

A question I would like the answer to before commenting on this subject is did all of the cameras used in this example you give use the same lens?

if they did then we can talk about the colors produced by the camera, if they did not then this cant be used to judge the difference between the cameras because it may be caused more by the lenses used.

I have two sony mirrorless cameras and several lenses and when I use my Zeiss glass I get very different color than when I use sony glass, and many of my older manual lenses produce very different colors

if we are going to compare the colors of different cameras the only way to do that and really know what we are judging them equally all of these cameras must be tested with the exact same lens

 

[knickerhawk]

A question I would like the answer to before commenting on this subject is did all of the cameras used in this example you give use the same lens?

if they did then we can talk about the colors produced by the camera, if they did not then this cant be used to judge the difference between the cameras because it may be caused more by the lenses used.

I have two sony mirrorless cameras and several lenses and when I use my Zeiss glass I get very different color than when I use sony glass, and many of my older manual lenses produce very different colors

if we are going to compare the colors of different cameras the only way to do that and really know what we are judging them equally all of these cameras must be tested with the exact same lens

I doubt that you'll get much disagreement with that observation about the potential impact of different lenses. The problem I see is that JACS is making a very broad - nearly categorical - claim about the deficiency of Sony colors with respect to the A series of cameras with little or no effort to isolate the possible variables (specific bodies, lighting conditions, lenses, processors, profiles, etc.) He only supports his claims with anecdotal evidence based on selective examples of posted images. He asserts that he can see the difference and expects us to accept his word for it. He asserts that the colors can't be corrected by use of appropriate color profiles and raw processing and expects us to accept his word for that too. He intimates that the Sony color problem is related to the CFAs used by Sony but provides nothing specific to support the claim and objects that tests like DXO's based on a widely used industry standard (GretagMacbeth) aren't reliable. In short, he raises questions and makes subjective and untestable assertions and ignores carefully done responses to the few times he offers testable evidence (as was the case with my first set of posts in this thread which JACS has ignored).

I grow weary of this game. I suggest that others let it go as well unless/until he offers anything other than a cavalcade of cherry-picked images and subjective opinion.

 

[J A C S]

A question I would like the answer to before commenting on this subject is did all of the cameras used in this example you give use the same lens?

if they did then we can talk about the colors produced by the camera, if they did not then this cant be used to judge the difference between the cameras because it may be caused more by the lenses used.

I have two sony mirrorless cameras and several lenses and when I use my Zeiss glass I get very different color than when I use sony glass, and many of my older manual lenses produce very different colors

if we are going to compare the colors of different cameras the only way to do that and really know what we are judging them equally all of these cameras must be tested with the exact same lens

I doubt that you'll get much disagreement with that observation about the potential impact of different lenses. The problem I see is that JACS is making a very broad - nearly categorical - claim about the deficiency of Sony colors with respect to the A series of cameras with little or no effort to isolate the possible variables (specific bodies, lighting conditions, lenses, processors, profiles, etc.)

I am just reporting what I see, not trying to make a full analysis.

He only supports his claims with anecdotal evidence based on selective examples of posted images.

That is not true, and I explained it so may times that I will not repeat it.

He asserts that he can see the difference and expects us to accept his word for it.

I find it very troubling that you cannot see what I am talking about. You may argue as long as you want why what you see does not prove what you think I am trying to prove, but saying that you cannot see anything wrong is either dishonest or ... I do not know, what would be the alternative?

He asserts that the colors can't be corrected by use of appropriate color profiles and raw processing and expects us to accept his word for that too.

No, I expect you to know it. I explained that many times, as well. You cannot correct lack of data by a clever algorithm but you can paint and create Mona Lisa, if you want.

He intimates that the Sony color problem is related to the CFAs used by Sony but provides nothing specific to support the claim

I am saying that it might be related to that, to refute claims, like the one you just presented, that a good profile would necessarily fix everything.

and objects that tests like DXO's based on a widely used industry standard (GretagMacbeth) aren't reliable.

They actually show differences, as I said before. They are not reliable to explain what you see because they report only a few numbers of a much more complex process (they measure functions). You cannot represent a spectral response with three numbers, and even if you have the whole spectral response, you need much more data to be able to tell how it affects an average landscape shot, for example.

In short, he raises questions and makes subjective and untestable assertions and ignores carefully done responses to the few times he offers testable evidence (as was the case with my first set of posts in this thread which JACS has ignored).

For a good reason. Those tests are useless. The spectrum (the reflectivity) of the green "leaves" could be vastly different than the greens in nature, the light, too; the hues are lacking, etc. I did not see a reason to keep discussing useless tests.

I grow weary of this game. I suggest that others let it go as well unless/until he offers anything other than a cavalcade of cherry-picked images and subjective opinion.

I said 4-5 times already that my images were not cherry picked, I explained why, and you keep repeating it. Good luck with that.

For the newcomers here, here is an anecdotal evidence straight from the source:

Loading…

www.sony.com

Look at the image right below the proud sign: "Natural results in any light".

 

[stevo23]

A question I would like the answer to before commenting on this subject is did all of the cameras used in this example you give use the same lens?

if they did then we can talk about the colors produced by the camera, if they did not then this cant be used to judge the difference between the cameras because it may be caused more by the lenses used.

I have two sony mirrorless cameras and several lenses and when I use my Zeiss glass I get very different color than when I use sony glass, and many of my older manual lenses produce very different colors

if we are going to compare the colors of different cameras the only way to do that and really know what we are judging them equally all of these cameras must be tested with the exact same lens

I doubt that you'll get much disagreement with that observation about the potential impact of different lenses. The problem I see is that JACS is making a very broad - nearly categorical - claim about the deficiency of Sony colors with respect to the A series of cameras with little or no effort to isolate the possible variables (specific bodies, lighting conditions, lenses, processors, profiles, etc.) He only supports his claims with anecdotal evidence based on selective examples of posted images. He asserts that he can see the difference and expects us to accept his word for it. He asserts that the colors can't be corrected by use of appropriate color profiles and raw processing and expects us to accept his word for that too. He intimates that the Sony color problem is related to the CFAs used by Sony but provides nothing specific to support the claim and objects that tests like DXO's based on a widely used industry standard (GretagMacbeth) aren't reliable. In short, he raises questions and makes subjective and untestable assertions and ignores carefully done responses to the few times he offers testable evidence (as was the case with my first set of posts in this thread which JACS has ignored).

I grow weary of this game. I suggest that others let it go as well unless/until he offers anything other than a cavalcade of cherry-picked images and subjective opinion.

Well said. It's essentially a circular, self-fulfilling prophecy. I think he's really just looking for attention in any form possible and he certainly gets it. He was even rebuked by a DPR admin but that didn't deter him.

 

[crashpc]

unless you actually like Sony colours .

It is easier to use canon files, both jpg and raw....than Sony ...where you need to be a computer / post processing kind of person. Canon DPP is easy...at least the older version. PS is not easy for raw files.

My experience.

It´s not always like that. I saw some comparison with 5DS recently, and I liked Sony A7R camera colors more (for the first time).

It really depends, and cannot be easily sorted out as a total rule.

I like Canon DPP colors, but the rest of the processing is not that good as with Adobe LR. So I always struggle a bit.