The roles of the camera and the raw developer in image color

[Jack Hogan]

Nikolai Vassiliev wrote: ...

There is an objective thing, called colour separation. You can do whatever you want with colour given the original has enough colour separation. The more plasticky the colour is, the harder it is to colour-correct (including the danger of getting into posterization). Providing excellent colour separation and low metameric error is what a good CFA does. The rest is software and operator skills.

There’s is the same color separation in the canon, is just that the colors are more realistic and the color balance is more natural. Have you downloaded the raws and tried to edit them yourself?

Juan,

The OP got over 70 likes, that's unusually many, re-read it carefully.

One of its main themes is that the raw converter has a lot do with how pleasing the final color turns out to be. In fact, if you read up on how color is derived from raw data and transformed to the final color space in typical photographic renderings, you will see that the differences introduced by the processing overwhelm, nay obliterate, most differences in the hardware.

So your example is a little bit like trying to tell whether the oranges used to make Tropicana orange juice are of better quality than those used to make Skipper. Once you put them through LR processing, they all taste about the same* - and then which is better just becomes a personal preference.

Of course this isn't the whole story, so now would be a good time to re-read the OP's last two paragraphs.

Jack

* This is one of the reasons why some of us don't use it as our main converter.

 

[juanmaasecas]

Nikolai Vassiliev wrote: ...

There is an objective thing, called colour separation. You can do whatever you want with colour given the original has enough colour separation. The more plasticky the colour is, the harder it is to colour-correct (including the danger of getting into posterization). Providing excellent colour separation and low metameric error is what a good CFA does. The rest is software and operator skills.

There’s is the same color separation in the canon, is just that the colors are more realistic and the color balance is more natural. Have you downloaded the raws and tried to edit them yourself?

Juan,

The OP got over 70 likes, that's unusually many, re-read it carefully.

One of its main themes is that the raw converter has a lot do with how pleasing the final color turns out to be. In fact, if you read up on how color is derived from raw data and transformed to the final color space in typical photographic renderings, you will see that the differences introduced by the processing overwhelm, nay obliterate, most differences in the hardware.

So your example is a little bit like trying to tell whether the oranges used to make Tropicana orange juice are of better quality than those used to make Skipper. Once you put them through LR processing, they all taste about the same, and then it's just a matter of personal taste which you prefer.*

Now would be a good time to re-read the OP's last two paragraphs.

Jack

* This is one of the reasons why some of us don't use it as our main converter.

You’re right, different software renders different results.

i love the output of dxo photolab, specially using a canon profile on it. But it’s so slow and I take so many pictures that it is not worth it for me.

and capture one, the initial look is much better than Lightroom, but to my taste, there’s something unnatural in the rendering that I don’t like. I don’t like how the colors change with wb, I don’t feel natural how shadows and highlights are recovered. Some people love capture one, but I just see that there’s something wrong. Of course I can’t explain what is it, and I need to try the latest version, but for me, Lightroom provides the best results, using my carefully made profiles.

what I see, the final results I see from people shooting Sony, and people shooting canon, in most of them I can tell what brand has been used. Maybe because most people use Lightroom or the default profiles? Maybe. But the general look of the pictures I see online, I like mich better the canon shooters than the Sony shooters. Of course there will be a few expert that can get really nice results (to my liking) starting from a Sony file, but most of the people photographs I see online that I like, are shoot with Canon, Nikon or Fuji. And Sony pictures have most of the time something that looks off to me. The scenes that are easier for Sony are those with very controlled lighting (studio pictures). For that kind of scene is more difficult to tell, or maybe everyone edits the pictures in a more similar way, but in day to day pictures, outdoors with different lighting, and indoors, Sony files are not as beautiful to my eyes as others.

 

[Iliah Borg]

Basically we see there that Sony+LR setup (Sony+C1 could be even more pronounced) produces more chroma variations of skin features - for glamorous aesthetic approach (wax-like skin of always-young models) Canon+LR version looked much convenient. But only this given postion - for documentary, b&w conversion this 'wax' is not always better (except the ease of retouching).

Same we can discover with different pair of cameras and plants (tree, grass etc.) greens - o remember a tread on LuLa forum where fresh D800 were compared with some 30+ megapixel MF back (P1 P30 or something like that). Medium Format shows less clarity and much less chroma uniformity so on Nikon you can't distinguish various trees and bushes breeds by green colour palette variation, only by leaves and branches shapes.


Camera standard on both. Canon left, Sony right. As shot WB

As we can see, the skin alone looks OK on the Sony, but the lips and some parts of the face are too pink, and the global picture looks like it has a color filter, like it has been color graded. The Canon file looks nice and clean. Let's make the picture colder, since the pullover is too yellow in the Sony:



Again, the skin in the Sony starts looking too cold, almost with purple areas, etc.

And again, if you put both pictures together, the Sony looks like it has been colorgraded, like it has a filter, and the canon just looks neutral and beautiful



I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders:


Canon (left) vs Sony with my profile (right).

Still not as good as the canon, but the best I can get so far.

I have tested Mauricio Pirazzini profiles, Huelight profiles (I recommend you to try them), my own profiles created and tuned in Lumariver, in X-rite software, and in DNG profile editor. I have tried a lot of stuff and tested through thousands of photos in different situations. And still, in my opinion, even in RAW, Canon is just better. But as I said, I love my A73, and with my personal developed profiles I am much happier now...

You’re totally wrong. It’s not normal than the Sony, having a warmer white balance and the pullover looking too yellowish (and in general the picture itself), for example, still the face is too cold. The color balance in the Sony is off, and it’s mich more difficult to have neutral colors with the Sony while keeping a healthy looking skin.
And, topically are Sony pictures the ones we see over the internet looking too cartoonish and wax. Canon pictures most of the time have a more neutral and true to life look to them. At least in my experience, to my taste. But of course taste is very subjective.

There is an objective thing, called colour separation. You can do whatever you want with colour given the original has enough colour separation. The more plasticky the colour is, the harder it is to colour-correct (including the danger of getting into posterization). Providing excellent colour separation and low metameric error is what a good CFA does. The rest is software and operator skills.

There’s is the same color separation in the canon

To check that, take two identical shots that include a neutral grey, raw, save TIFFs from raw with nothing but appropriate Custom WB from click-on-gray and gamma applied (you can do that in RPP, select "Colorimetric gamma" in "Curve type" and "Raw RGB TIFF" from the drop-down to the right of the "Save" button), and compare the gamuts of the resulting TIFFs.

is just that the colors are more realistic and the color balance is more natural.

Not sure you understand what colour separation is, or what I'm saying - my personal preferences has nothing to do with it. I'm explaining the role of CFA, that's all.

As Dan Margulis use to say, there are no bad originals, only bad Photoshop operators - and he is one of the three original Photoshop Hall of Famers, together with Dianne Fenster and Thomas Knoll.
--

About LibRaw | LibRaw

www.libraw.org

 

[fferreres]

Basically we see there that Sony+LR setup (Sony+C1 could be even more pronounced) produces more chroma variations of skin features - for glamorous aesthetic approach (wax-like skin of always-young models) Canon+LR version looked much convenient. But only this given postion - for documentary, b&w conversion this 'wax' is not always better (except the ease of retouching).

Same we can discover with different pair of cameras and plants (tree, grass etc.) greens - o remember a tread on LuLa forum where fresh D800 were compared with some 30+ megapixel MF back (P1 P30 or something like that). Medium Format shows less clarity and much less chroma uniformity so on Nikon you can't distinguish various trees and bushes breeds by green colour palette variation, only by leaves and branches shapes.


Camera standard on both. Canon left, Sony right. As shot WB

As we can see, the skin alone looks OK on the Sony, but the lips and some parts of the face are too pink, and the global picture looks like it has a color filter, like it has been color graded. The Canon file looks nice and clean. Let's make the picture colder, since the pullover is too yellow in the Sony:



Again, the skin in the Sony starts looking too cold, almost with purple areas, etc.

And again, if you put both pictures together, the Sony looks like it has been colorgraded, like it has a filter, and the canon just looks neutral and beautiful



I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders:


Canon (left) vs Sony with my profile (right).

Still not as good as the canon, but the best I can get so far.

I have tested Mauricio Pirazzini profiles, Huelight profiles (I recommend you to try them), my own profiles created and tuned in Lumariver, in X-rite software, and in DNG profile editor. I have tried a lot of stuff and tested through thousands of photos in different situations. And still, in my opinion, even in RAW, Canon is just better. But as I said, I love my A73, and with my personal developed profiles I am much happier now...

You’re totally wrong. It’s not normal than the Sony, having a warmer white balance and the pullover looking too yellowish (and in general the picture itself), for example, still the face is too cold. The color balance in the Sony is off, and it’s mich more difficult to have neutral colors with the Sony while keeping a healthy looking skin.
And, topically are Sony pictures the ones we see over the internet looking too cartoonish and wax. Canon pictures most of the time have a more neutral and true to life look to them. At least in my experience, to my taste. But of course taste is very subjective.

There is an objective thing, called colour separation. You can do whatever you want with colour given the original has enough colour separation. The more plasticky the colour is, the harder it is to colour-correct (including the danger of getting into posterization). Providing excellent colour separation and low metameric error is what a good CFA does. The rest is software and operator skills.

There’s is the same color separation in the canon, is just that the colors are more realistic and the color balance is more natural. Have you downloaded the raws and tried to edit them yourself?
I would love to see you guys in one of this pictures, being you the model and editing them, and see what color you prefer.
I feel sometimes when we edit we do and we go for the colors we think are “right”, but when we are in the picture, we realise we prefer a different white balance, different contrast and saturation level, etc. That’s why when I’m taking pictures i lately love to appear in some pictures as well, to use myself as a “person balance”. More or less, editing a picture to my liking, then other people will look ok as well for them. If we just look at the screen and try to recover highlights, shadows, too neutral white balance, etc, then the picture looks horrible. Maybe perfect from an academic point of view, but horrible from an artistic point of view. And I like to take nice photographs, not paintings or art copy work.
please download the files and play with them. And then show your edits to someone else. No one likes seeing their face purple while the rest of the picture is warm.

I may be wrong, but Canon glass seems to produce images that look more Canon-like under a Sony, and non Canon glass in my 5D I also seems to not look that much like the Canon look. I don’t know to what extent but in reviewing images from many Adapted lenses, the lens itself seems to also be responsible for the look. My Minolta sessions in the Sony A7II look very much Minolta-like. They are pretty much unlike Zeiss glass for example, and pretty much unlike Yashica. I actually don’t like how Canon lenses render that much. It looks too serious, a little formal of sorts; to some extent I’d say the seem more SAD.

Coatings are very proprietary and good vendors like Canon will try to keep some rendering and stick to it, and a good part may reside in how the glass is coated (or maybe something more complicated)

 

[juanmaasecas]

Basically we see there that Sony+LR setup (Sony+C1 could be even more pronounced) produces more chroma variations of skin features - for glamorous aesthetic approach (wax-like skin of always-young models) Canon+LR version looked much convenient. But only this given postion - for documentary, b&w conversion this 'wax' is not always better (except the ease of retouching).

Same we can discover with different pair of cameras and plants (tree, grass etc.) greens - o remember a tread on LuLa forum where fresh D800 were compared with some 30+ megapixel MF back (P1 P30 or something like that). Medium Format shows less clarity and much less chroma uniformity so on Nikon you can't distinguish various trees and bushes breeds by green colour palette variation, only by leaves and branches shapes.


Camera standard on both. Canon left, Sony right. As shot WB

As we can see, the skin alone looks OK on the Sony, but the lips and some parts of the face are too pink, and the global picture looks like it has a color filter, like it has been color graded. The Canon file looks nice and clean. Let's make the picture colder, since the pullover is too yellow in the Sony:



Again, the skin in the Sony starts looking too cold, almost with purple areas, etc.

And again, if you put both pictures together, the Sony looks like it has been colorgraded, like it has a filter, and the canon just looks neutral and beautiful



I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders:


Canon (left) vs Sony with my profile (right).

Still not as good as the canon, but the best I can get so far.

I have tested Mauricio Pirazzini profiles, Huelight profiles (I recommend you to try them), my own profiles created and tuned in Lumariver, in X-rite software, and in DNG profile editor. I have tried a lot of stuff and tested through thousands of photos in different situations. And still, in my opinion, even in RAW, Canon is just better. But as I said, I love my A73, and with my personal developed profiles I am much happier now...

You’re totally wrong. It’s not normal than the Sony, having a warmer white balance and the pullover looking too yellowish (and in general the picture itself), for example, still the face is too cold. The color balance in the Sony is off, and it’s mich more difficult to have neutral colors with the Sony while keeping a healthy looking skin.
And, topically are Sony pictures the ones we see over the internet looking too cartoonish and wax. Canon pictures most of the time have a more neutral and true to life look to them. At least in my experience, to my taste. But of course taste is very subjective.

There is an objective thing, called colour separation. You can do whatever you want with colour given the original has enough colour separation. The more plasticky the colour is, the harder it is to colour-correct (including the danger of getting into posterization). Providing excellent colour separation and low metameric error is what a good CFA does. The rest is software and operator skills.

There’s is the same color separation in the canon, is just that the colors are more realistic and the color balance is more natural. Have you downloaded the raws and tried to edit them yourself?
I would love to see you guys in one of this pictures, being you the model and editing them, and see what color you prefer.
I feel sometimes when we edit we do and we go for the colors we think are “right”, but when we are in the picture, we realise we prefer a different white balance, different contrast and saturation level, etc. That’s why when I’m taking pictures i lately love to appear in some pictures as well, to use myself as a “person balance”. More or less, editing a picture to my liking, then other people will look ok as well for them. If we just look at the screen and try to recover highlights, shadows, too neutral white balance, etc, then the picture looks horrible. Maybe perfect from an academic point of view, but horrible from an artistic point of view. And I like to take nice photographs, not paintings or art copy work.
please download the files and play with them. And then show your edits to someone else. No one likes seeing their face purple while the rest of the picture is warm.

I may be wrong, but Canon glass seems to produce images that look more Canon-like under a Sony, and non Canon glass in my 5D I also seems to not look that much like the Canon look. I don’t know to what extent but in reviewing images from many Adapted lenses, the lens itself seems to also be responsible for the look. My Minolta sessions in the Sony A7II look very much Minolta-like. They are pretty much unlike Zeiss glass for example, and pretty much unlike Yashica. I actually don’t like how Canon lenses render that much. It looks too serious, a little formal of sorts. Coatings are very proprietary and good vendors like Canon will try to keep some rendering and stick to it, and a good part may reside in how the glass is coated.

I agree that Canon glass affects, and the most I think is the contrast curve (and rolloff), and how the contrast affects colors. It's not just about getting an accurate matrix, but about how the colors change. I've run my sample photos In Capture one and in DxO PhotoLab using different profiles, and I just like the Canon files better, and when you have a better starting point as a reference, the end result will be better. It's how our brain works. We think something looks good if we don't have a reference, but having the canon reference besides the Sony, the Canon looks better to me. And tweaking white balance, profile and colors, I can get something very similar to the Canon from the Sony, but, what happens if we don't have that reference? This is why I think I like better what most Canon photographers pictures look like: they just have a better neutral looking reference.

I LOVE my Canon 70-200 BTW.

 

[fferreres]

It’s a gorgeous rendering. Canon had the advantage of DPP since day 1. Lightroom did not exist at the time. Canon established how it should look with DPP and it is likely vendors like Lightroom had to to do a better finery ed job with CR2 and CR3 files to establish themselves to replaced DPP. Canon glass that I have (about 5 lenses) has some fingerprint but probably, combined with DPP Canon had a better plan to establish a consistent look and rendering where as Sony was/is more all over the place. Not sure if am saying nonsense, but I still have DPP and at least in my case, I’d not accept much deviation from what DPP would get me.

 

[Erik Kaffehr]

I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders.

Hi,

Thanks for the invitation, but I don't know where to find the files...

Best regards

Erik

 

[juanmaasecas]

I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders.

Hi,

Thanks for the invitation, but I don't know where to find the files...

Best regards

Erik

MEGA

MEGA provides free cloud storage with convenient and powerful always-on privacy. Claim your free 20GB now

mega.nz

 

[Erik Kaffehr]

I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders.

Hi,

Thanks for the invitation, but I don't know where to find the files...

Best regards

Erik

https://mega.nz/#F!RIRhFSjJ!LrFqofyr5TxGrWQJ2dadXQ

Hi,

Thanks a lot! I downloaded the images. Here are quick and dirty processed interpretations:

https://echophoto.smugmug.com/Canon-vs-sony/

For the whole figure I used the shoe as reference white point, used Lumariver profiles and adjusted exposure.

For the close in portrait I did use Adobe standard. Lacking a good WB reference is not easy, but I recalled that skin tones used to be around #,15,15 in Lab coordinates.

The 'Temp' slider affects 'b' coordinate and the 'Tint' slider operates mostly on 'a' axis.

For the 'close in' portrait I used Adobe standard.

The subject is tricky, as it is mixed light, The face is in shadow so it is mostly illuminated by clouds and blue sky. Background is sunlit.

Best regards

Erik

 

[juanmaasecas]

I invite you to play with the files. It is so much more difficult to get right pictures with the Sony after you see what the canon with one simple click can do. However in this case, my profile/preset does wonders.

Hi,

Thanks for the invitation, but I don't know where to find the files...

Best regards

Erik

https://mega.nz/#F!RIRhFSjJ!LrFqofyr5TxGrWQJ2dadXQ

Hi,

Thanks a lot! I downloaded the images. Here are quick and dirty processed interpretations:

https://echophoto.smugmug.com/Canon-vs-sony/

For the whole figure I used the shoe as reference white point, used Lumariver profiles and adjusted exposure.

For the close in portrait I did use Adobe standard. Lacking a good WB reference is not easy, but I recalled that skin tones used to be around #,15,15 in Lab coordinates.

The 'Temp' slider affects 'b' coordinate and the 'Tint' slider operates mostly on 'a' axis.

For the 'close in' portrait I used Adobe standard.

The subject is tricky, as it is mixed light, The face is in shadow so it is mostly illuminated by clouds and blue sky. Background is sunlit.

Best regards

Erik

Thank you so much for your time.

The white balance should be the same in both pictures, they were taken just a few minutes apart from each others.

For sure the lighting situation is very tricky, as it's a backlit photo and it's not near golden hour... So even if you try to use the sneaker as a WB target, what part of the sneaker?

Anyways, I like my interpretation with my own profiles more than your edits with Lumariver profiles/adobe standard, and your WB of choice (every picture has a different look).

 

[JimKasson]

I agree that Canon glass affects, and the most I think is the contrast curve (and rolloff), and how the contrast affects colors. It's not just about getting an accurate matrix, but about how the colors change.

First off, with sophisticated profiles, it's not just about getting a matrix, but also about getting a LUT.

Second, accuracy across the whole color gamut automatically makes the gradations right, assuming adequate precision.

Jim

--
https://blog.kasson.com

 

[ahaslett]

Great post - should pop up every time color is debated!

Color is also a matter of personal taste - and that taste usually drift during a lifetime (also as the eye lens becomes less transparent and more yellow by age - young people usually see shorter wavelenghts in the blue spectrum than oldies - and so on). So my ideal color palette was most likely different 20 or 40 years ago.

Minolta back then made some noise about using two different greens for their Bayer matrix - to better differentiate different hues of green (after all that is where the sensitivity of the eye peak).

Yes - color is complex. And very interesting. Always want to know more about the topic...

The way that colours are perceived also

Not also. That element is present in all color perception.

has a cognitive processing element. For example the well-known effect of size of colour decks used for decorative selection.

Right. The spatial effects are important, and less well understood than the spectral effects. It's not even clear that superposition applies. That's one reason there are 2-degree and 10-degree standard observers.

Jim

We did some research at Dulux using different sizes and colours of colour decks as against a painted “room”. We also calibrated colours as seen on a PC monitor against subjective perception.

Apparently the academic doing the work wasn’t convinced of the value, so my colleague sent him into the room, asked him outside the room to pick the colour from a colour card and sent him back into the room with the card. His reaction “that’s amazing”.

One of my other colleagues used printed images to study affective responses to fragrances. Colour was an important factor as I recall, which is hardly surprising.

In the end, it’s all about reactions to images.

The old saw is that half the neurons in your brain are involved with vision. That makes color science a deep area to mine. We are in no danger of understanding it all anytime soon. But we can level the playing field do make limited progress, mostly by holding user conditions constant. We've come a long way in the last 30 years wrt device independent color, but we've got a long way to go.

Jim

As D-day approaches, I just wanted to thank you for all the lessons I have learned from you!

All the best

Andrew