Color Science: Canon vs Sony

A few weeks ago I had the opportunity to play with my Canon 6D and a Sony A7RII with a Sigma adapter.


I shot these photos within moments of each other. The settings on both were:
1/250, f/2.8, ISO 125.
The images were shot with a Canon 70-200 2.8L II IS. They were shot RAW, and the temp/tint match was done in ACR.

The images were visually matched for color temp and tint and no other adjustments were made to color tones, or anything.
Oddly, I noticed the exposure on the Sony was (originally) slightly darker (like .10 stop)--possibly from the adapter? Anyway, I tweaked that too.

Anyway. First pic is just the two images, side by side, straight out of camera with only color temp/tint visual matching.

The second photo is both images, side by side as I would edit them for a portrait shoot thru ACR then in PS, minimal matching--more just "make it look good."

In both cases of "match" and then "real edit" I did not spend much time or energy on matching or editing... I was not agonizing over these for hours.

The left image is A and the right image is B


So. Which is which? Which do you like?


I have to admit, it challenged my conception of color science, and of response tonality in RAW and in PS.
The "which do you like better" vids on YT are hugely problematic because they never visually match images: they assume that 5500K on one camera is the SAME as 5500K on another camera.
To match these images in ACR visually means that one was at 7100K and the other was at 6200K. A 32" inseam isn't the same between any brand of jeans in the world, and apparently neither is 7100K between cam brands.
The point of my little experiment vs the YouTuber's comparisons was to match things as closely as possible for the "starting point" and then to see if I could get a result I was happy with, assuming the starting point was the same.
The other goal was to not spend much time agonizing over my efforts to match the shots. In other words, if I jump to Sony, I don't want a 10% increase in editing time because of color science.

I am very much interested in replicating this test with an A7iii or an A7RIII because I hear the color response is even more improved.
In any case, call me intrigued and rather impressed at how close they came out (at least in this one instance.)

The biggest differences I can see (and there isn't much difference) is in the retention of detail in the highlights and in the shadows the hair above his right eye and the bright spot on his left shoulder. There is also a slight difference in the luminance of the two... A images have a bit more color present in his shirt. Skin tones match pretty well..... and proves that in most situations both cameras will do just fine. With out seeing the actual, it would be hard to tell which of these two rendered better. Thanks for posting.

The interesting thing about A vs. B comparisons, is that even when the two cameras being compared are both from the same manufacturer, the photos are not identical in appearance!

Im still going to shout canons color science is the best. Just be sure to tell me which picture was shot with canon so that i know which picture i like better

Do such comparison say anything about the camera or is it all down to the camera profile of the raw converter?
Just look at all the LR vs. C1 comparisons (e.g. https://youtu.be/AZdzc​aaqaeY?t=1m45s) where they show the clear difference between two raw converters for the same raw file. And the differences are also in my experience way bigger, then what I see above.

drmaxx wrote in post #18765248
Do such comparison say anything about the camera or is it all down to the camera profile of the raw converter?
Just look at all the LR vs. C1 comparisons (e.g. https://youtu.be/AZdzc​aaqaeY?t=1m45s) where they show the clear difference between two raw converters for the same raw file. And the differences are also in my experience way bigger, then what I see above.

BOTH...the choice of camera model matters in any color science comparison -- due perhaps to differences in sensors between models,
AND...the choice of RAW convertor matters in the interpretation of what the sensor recorded.

And just because one RAW convertor is employed for different models of one brand also ASSumes that the RAW conversion is done 'as the manufacturer intended' for each different model sensor!
...what accounts for the variances seen in Post 3 photos ? ? ?

Here an example of LR vs C1 - my point beeing that the difference are considerable just based on the applied camera profile:

since there seems to be some question...


the idea behind this exercise is if Joe Photog has a Canon or a Sony and he wants to jump into the other brand, what does it really mean for him in terms of skin tones?
How does his workflow change? Is it going to be a nightmare just getting good colors out when shooting in RAW?
If a guy just switches brands, can he get good skin tone results with his current work flow? What are the real life implications of jumping from one brand to another?
Is it some huge ordeal? Or is it no big deal?
Those are the questions I wanted to answer with the most simple test I could do in the little time I had.

Now if my exercise had been about software I would have used one camera in C1 vs ACR/PS and A/B'd those results.
If it had been about a lens, I would have used one body and two lenses and done the A/B that way.

But this is about the current Big Kvetch: "Sony Skin Tones suck!"

My starting point to that was, "well maybe, but let's see. Let's get these photos close, note what it took to get there, and then edit them like I would do, and step back and see the results."

My findings are that it took a lot less work than I thought, and the Sony ended up better than I thought it would, and that there's a lot less difference than all the big A/B comparison vids online.
YMMV, and in different light, I would expect different results--which I did not get to test. Yet.

Addendum: I would expect to find variances between camera models, but I would expect to find the most variance between camera BRANDS.
The goal here is not to say "there's no difference--you can just color balance it out", but rather to explore what the difference really is when you're not trying to make a drastic comparison for clickbait.

I think the "matched" shots still have the tint a little off; in the skintones A leans a little green and B a little magenta. This makes me think A is Sony and B is Canon, but it could easily be your adjustments making it look that way.

Overall I like B a little better in the "matched" comparison (due to the skintone tint) and A a little better in the "processed" comparison. But they're close enough I'd be happy with either camera.

I tried really hard to tune that "flat out of camera" set to be close...it seemed like the pink/green balance was one way in photoshop, and another when I uploaded it.
So yes the pink on the right can certainly be reduced--it looks closer in PS, but more pink here in the browser...so yes, you are absolutely right about that.
And I felt the same way: I am happy with both "edited" versions.

Rogue,

I just find shooting using AWB Sony seems to appear cooler in Kelvin.

The difficulty in this type of discussion is that different light scenes will also change the behavior of the image.

I've taken photos of cars outdoors for a local VW show for just under a decade. Overcast/cloudy days with no sun can make any photo drab and miserable looking. However on extremely hot sunny days the photos can change dramatically. Seems when I shoot canon the AWB will take on a warmer look when I import RAW to LR. Sony seems to have cooler white balance after importing.

As an new adapter to the Sony world I found batch filing not as easy as my Canon files for events work. I've had discussions with a friend of mine (Old POTN members may recognize Nick-san)that's shot Canon to Nikon ...then back to Canon .....Now Sony A9. He's also indicated a difference in his workflow moving to Sony. Sony Raw files definitely manipulate differently than Canon RAWS.

I purchased the first Sony A7iii that hit my local area back in April 12, 2018. If I compare editing Fuji files the work involved is almost effortless in achieving the look I want. Seems Fuji's AWB hits a pleasant sweet spot in many lighting environments. Editing Canon I find my workflow go smoothly but a little more work in PP. Sony RAW files in LR I still find to this day more effort in making the images similar in kelvin. Sony does seem to be cooler in many cases and I will need to slide the kelvin slider a tad more than Canon in many cases. Fuji AWB really seems to hit the target very well in many cases.

In a beautiful warm day in the sun fuji almost hits this calm eye candy look almost SOOC with minor tweaks in RAW. This has to do with mix of how the camera renders in this type of light and partly Fuji colours IMO. Sony does seem to have more contrast and colours would need some more warmth in post. SOOC raw has a more aggressive look but possibly due to how the camera's AWB decides.

Now that I am shooting more Sony I'm almost considering to shoot grey card to see if this changes my LR workflow. I'm having a lot of fun experimenting at the moment...when I have time.

I went to dpreview.com and looked at past reviews and testing of two Sony and two Canon FF cameras.
dpreview shoots a 'standard' scene with cameras under review, for comparison purposes... a still scene with resolution targets and color gamuts, and four reproductions of 4 different models of different ethnic backgrouds.

Viewing the same (Asian) face in the standard scene using their direct comparison tool, I found that in general, when viewing a RAW shot and the standard viewing software used by dpreview, the Sony cameras came out with notably warmer and redder facial tones than the Canon... an Asian face lacking in yellow tones is not expected! (BTW, being Asian myself, I can readily make that statement about accuracy of skin tones...I only go 'red' when sunburned!)

Wilt wrote in post #18766349
I went to dpreview.com and looked at past reviews and testing of two Sony and two Canon FF cameras.
dpreview shoots a 'standard' scene with cameras under review, for comparison purposes... a still scene with resolution targets and color gamuts, and four reproductions of 4 different models of different ethnic backgrouds.

Viewing the same (Asian) face in the standard scene using their direct comparison tool, I found that in general, when viewing a RAW shot and the standard viewing software used by dpreview, the Sony cameras came out with notably warmer and redder facial tones than the Canon... an Asian face lacking in yellow tones is not expected! (BTW, being Asian myself, I can readily make that statement about accuracy of skin tones...I only go 'red' when sunburned!)

Being Asian having my High end sake and single malt scotch gets me in the red

AlanU wrote in post #18766365
Being Asian having my High end sake and single malt scotch gets me in the red

Ueno...did I guess right?

It sounds like you react to alcohol like my cousin! My wife and I went to his wedding, and at the reception each table toasted him and his wife. Nearing the end of the circulation he was apparently in distress even though he was feeling no pain, from the appearance of his face. Shortage of alcohol dehydrogenase afflicts many Asians, but not me. I drove my Best Man and good friend home after my 'bachelor bash' because they were unable.

Wilt wrote in post #18766380
Ueno...did I guess right? That is the proper way but they added a "Y" for some odd reason. Name commonly found incorrect in Hawaii too.

It sounds like you react to alcohol like my cousin! My wife and I went to his wedding, and at the reception each table toasted him and his wife. Nearing the end of the circulation he was apparently in distress even though he was feeling no pain, from the appearance of his face. Shortage of alcohol dehydrogenase afflicts many Asians, but not me. I drove my Best Man and good friend home after my 'bachelor bash' because they were unable.

You're a blessed man to not get red