Color science

[webrunner5]

14 minutes ago, Mako Sports said:

Canon color on the older cameras was so much better. Like the 5Dmk2, 1DC, 7D, and T2i had awesome color.

Yeah like it or not they have changed it.

My favorite Canon color wise was the original 5D. It was a wedding shooters dream camera.

[mercer]

1 hour ago, kaylee said:

you should buy a sony rx10 ii

I did... twice. The first time didn’t work out so well, although I really liked the camera, the second time I used it for B&W.

[frontfocus]

now I know that, while I never warmed up to Sonys colors, especially skin tones, other like it. What a big surprice.

Didn‘t watch the entire video, did he fix white balance? And adjust the picture profile? 

[Mako Sports]

3 minutes ago, frontfocus said:

now I know that, while I never warmed up to Sonys colors, especially skin tones, other like it. What a big surprice.

Didn‘t watch the entire video, did he fix white balance? And adjust the picture profile? 

Tbf the Sony skin tones are getting better. I wish my A6300 produced the same colors as my Z90!

[Shirozina]

'Colour Science' is the systems used by manufacturers to interpolate missing colour information that their sensors can't capture. The Red Green and Blue colour filter arrays ( CFA's ) on sensors determine how accurate the colour capture can be on a sensor with stronger colour filters enabling better colour fidelity however stronger filters mean that less overall light is getting to the photo-sites which reduces sensitivity so manufacturers are always playing these 2 things off against each other. As photosite sensor size goes down due to smaller sensors and higher density's the situation is getting worse but obviously better algorithms are at work trying to interpolate missing colour information. On top of that if you are using an internal Y'CbCr codec where there is reduced colour information in the first place even with a 4.2.2 codec and worse a 4.2.0 codec you are left with not that much colour info to work with. RAW which is RGB is obviously better here. Now most people are not too bothered about differentiating between slightly different hues of green in foliage but our eyes and brains are programmed to be extremely sensitive to even the slightest changes in skin colour so getting interpolated colour decisions wrong in this area even by a very tiny amount will lead to huge apparent differences in skin tone quality and people liking or hating a particular brand's 'colour science'. Canon's approach seems to be one of it knows it can't get it accurate so let's make it pleasing and give everyone a nice suntan.

[austinchimp]

I think one of the problems in this discussion (which seems to pop up time and time again) is the fact that while rendition of colour clearly has a lot of scientific and technical elements behind it, what we as creators and educated consumers, are searching for is an emotional reaction. What we discus endlessly as 'colour science' is actually a euphemism for emotional connection to the image, which is far more nebulous and hard to define. Accurate colour does not necessarily equate to emotional connection to the image.

And the truth is, while virtually all modern cameras can deliver good if not great colour under the right circumstances, some cameras are able to hit that emotional sweet spot more easily. Traditionally those have been, for example, F35, D16, Canon, Alexa and more recently Fujifilm. You can get in the ballpark with Sony, but it rarely gives you that satisfying feeling of 'oh yeah'. I know I've spent many hours struggling with Sony and Panasonic footage, when I can grade a clip in a few minutes with my Digital Bolex, and it looks and feels like magic.

It's probably 70% emotional connection to and nostalgia for the look and feel of movies and tv we saw when we were young, and 20% related to modern 'cinematic' trends, and %10 related to actual colour technical accuracy. All in my humble opinion of course.

[austinchimp]

Also worth remembering that Tony Northrup's video (which I thought was very well done entertainment) is focusing on raw stills, while compressed video codecs will imprint a definite colour science due to the decisions of the manufacturer. Tony points this out in the video.

[Castorp]

It’s not a good test. I work a lot with Fujifilm jpegs and I don’t have to compare with other cameras, just with the different film simulations. I use X Raw Studio so using the cameras own engine to process the raw. The difference between f. e. Astia and Provia is subtle yet striking. Very difficult to get to those results from a flat raw file. 

Raw files aren’t created equal either. For example the raw files from the Nikon Df/D4 are amazing to begin with. 

I think many probably have this experience of sitting in front of a computer and know how long it takes them to get a result they’re happy with. I think Andrew could probably say something about this test and testify to the difference with working with different files.

That viewers don’t make the same difference isn’t strange. 

I agree with Tony that Hasselblad and Leica really deserve more credit for their colour work. 

 

[Mattias Burling]

5 hours ago, webrunner5 said:

Yeah like it or not they have changed it.

My favorite Canon color wise was the original 5D. It was a wedding shooters dream camera.

I did some side by side with the original 5D and the 6Dmkii this sumer. Virtually indistinguishable.
So they have kept their colors very well.
 

[Yurolov]

He was shooting jpeg on god knows what color profile for each camera. The test has zero relevance for shooting professionals. I give u slog2 on a a7sii and clog on a Canon, and you better believe there is a significant difference. There is a direct comparison done by that potatojet guy. Same thing, blind test. 80 percent picked canon. All of a sudden we are all miraculously wrong and Sony is really the best lol. 

[Mattias Burling]

9 hours ago, fuzzynormal said:

"we adjust the color...anyway.  As soon as you do...you invalidate the color science."

I find that to be incorrect. Even when adjusting raw, the original color science is in effect and dictates how and how much adjustments are needed. Otherwise for example all LUTs would work the same on all footage. But they don't. Anyone with two different raw capable cameras can easily try for themselves. Take a side by side. Adjust one and copy the settings to the other. It won't look the same.

And that is the key. Sure I can fiddle with images but sometimes there are a lot of them. In a project I was leading recently we where a couple of photographers that took roughly 15 000 photos. And I edited out maybe 5000 keepers that are going to be used in various marketing material. I was pretty happy that it was only one photographer that used a different brand and the rest of us all had the same color science.

If I ever for some reason would start to work full time as a photographer or filmmaker again. Spending a huge amount of time processing pictures. Then I would definitely want to minimize the time spent in post. 

[Robert Collins]

I think the basic term 'color science' is ripe for argument. The fact is there is an awful lot about 'color' that is 'subjective'. Everyone 'sees' color differently, colors (and luminosity) look different to the same person when set next door to different color and luminosity. Even our mood effects how we see colors (as in the expression 'seeing red'). And then f people say there isnt a lot of 'science' in color others read that as color doesnt matter - when it clearly does - even if color is more 'art' than 'science'.

[zerocool22]

Yeah I saw a camera comparison yesterday between canon r, a7iii, pocket 4K, nikon z7 and XT3. The colors from Canon R, XT3 and pocket4K stood out a whole lot better. And then on youtube I was seeing all these replies sony A7III is the clear winner. Maybe my eyes are different, or maybe its like the saying, "In the land of the blind, one-eye is king". 

I find color science one of the most important aspects. With most dslr's you only get 8bit, when you start changing colors around, the image gets often degraded in quality or it is very time consuming to get things right. But I have seen a trailer of a feature film shot with an A7S II and it looked pretty good, which amazes me. Wondering what settings they shot in. 

[webrunner5]

2 hours ago, Mattias Burling said:

I did some side by side with the original 5D and the 6Dmkii this sumer. Virtually indistinguishable.
So they have kept their colors very well.
 

I think you either need your eyes examined or a new Monitor if you think Canon still has the same Color Science on the 5D versus the 6D mk II. 5D on the left, 6D mk III on the right.

 

 

[Mattias Burling]

9 minutes ago, webrunner5 said:

I think you either need your eyes examined or a new Monitor if you think Canon still has the same Color Science on the 5D versus the 6D mk II.

What settings did you use in the two cameras and why haven't you lined them up?
Also, why did you change the scene in between your shots? Can you share the raw files?
My eyes and monitor are just fine thank you.

[webrunner5]

4 minutes ago, Mattias Burling said:

What settings did you use in the two cameras and why haven't you lined them up?
My eyes and monitor are just fine thank you.

They are not lined up because they have way different MP sizes. And that test is from Imaging Resources. They have used the same test for 15 years on just about every camera made. I don't think you Can see to be honest.

[Yehouda]

The relation between sensor and lense is also very important for color rendition.

[webrunner5]

1 minute ago, Yehouda said:

The relation between sensor and lense is also very important for color rendition.

They use the same lens. Jesus.

[Mattias Burling]

7 minutes ago, webrunner5 said:

They are not lined up because they have way different MP sizes. And that test is from Imaging Resources. They have used the same test for 15 years on just about every camera made. I don't think you Can see to be honest.

Thats incorrect. Its not even the same scene.. and my eyes are supposedly bad.. lol

 

Any who, this is shot (by me) with the same lens and the 5D + 6Dmii. The lights have obviously changed a tad between shots and sure, some variation is there.
But I do absolutely not see a huge difference of color tones granting me to question eyes and monitors of those that don't. DR I see a slight difference in, one is also a smidge warmer. But nothing groundbreaking. Tomato, tomato.

 

 

[Yehouda]

17 minutes ago, webrunner5 said:

They use the same lens. Jesus.

I don't say that specifically about your exemple and Canon color science evolution. I talked about relation between sensor and lense for color rendition as a general rule.