D600 color and contrast vs. D800 vs. 5D III
/forum/topic/1153266/0



splathrop
Registered: Feb 27, 2006
Total Posts: 527
Country: United States

Now that test images for all three cameras are available on dpreview, it might be interesting to hear comments about what raw-file Kodak color chart comparisons show. My initial impressions: both the D600 and the 5D III do a better job with color and contrast than the D800, which is probably not surprising.

The D600 is the high ISO noise champion.

The color and contrast comparison between the D600 and the 5D III is harder to evaluate. They each do better in some comparisons and worse in others. Generally, the D600 shows less saturation (at least in this test, hard to know whether settings make a difference). At times the lower saturation makes the D600 colors look more controlled, at other times, especially in middle densities, the greater saturation in the Canon colors makes them appear to separate better.

The comparisons seem to switch around a bit with changes in ISO.

Both cameras fail, or nearly fail, to resolve a few color pairs, and it tends to be the same pairs that trouble each camera, but not to the same degree. One will show a bit more trouble in dark greens, the other in deep magentas, etc. I think the Canon does just a tad better in the pastels.

I'm seeing this on a calibrated NEC 27-inch monitor. Obviously different setups will show it differently, and different eyes will see it differently.

Based on what I see in these color and noise comparisons, the D600 has moved up a notch in consideration of what this Canon user will be recommending to friends who ask about a camera for family use. But the 5D III is hanging in there as a body, and those who ask about more lens-challenging applications, especially for less-expensive longer zooms, will probably get that recommendation.

It will be interesting to see if anything worth noting happens when the Canon 6D is added to the mix.

Any comments?



jmcfadden
Registered: Oct 30, 2002
Total Posts: 30235
Country: United States

You said it best near the end. You are a Canon user and likely an engineer and both are cool

Someday someone with way more brains than myself and a likely Canon shooter too boot a doctor of sociology could parse it out but your expressed concerns are not things that get the Average Nikon shooters blood moving. It is amazing to me but if you posted this same tread over in the Canon forum it would go on for pages and pages. It will barely get a nod over here

J



SoundHound
Registered: Jan 14, 2006
Total Posts: 5318
Country: United States

It appears that the $2100 D600 goes head to head, IQ and features, with the $3500 Canon 5D MK III. The, unavailable,6D appears to compete on price but not on features. That's a big price difference!



Jay968
Registered: Mar 29, 2011
Total Posts: 89
Country: United States

I've been shooting with both D800 and D800e for about a month now and just spent 3 days comparing them to the D600. Yes the saturation from the 600 is lower but more importantly to me was the fact that compared to the D800 cameras, the D600 results tend to lack sharpness. I assume that this is due to the fact that it has a full fledged anti aliasing filter in place rather than what both D800 cameras have to offer and that I have grown spoiled by the sharpness out of the D800 bodies, but fact is fact...an anti aliasing filter is going to detract from the sharpness, and it is quite obviously the case when comparing these cameras side by side. I considered purchasing a D600 but don't think I am going to do so due to this.



splathrop
Registered: Feb 27, 2006
Total Posts: 527
Country: United States

Thanks for those. Keep them coming. So far, nobody seems to have gone to the dpreview site and actually looked at the Kodak color chart and set it up to compare these 3 bodies. It's worth a look.



bemyzeke
Registered: Jul 04, 2008
Total Posts: 384
Country: United States

Having been a Canon 5D series user for several years, and now D800e user for a couple of months, I will challenge anyone to produce a color with one camera that can not be created by the other camera.

The colors are nothing but how Raw files are interpreted. I am assuming that OP looked at the Raw version which is using Adobe's profile. Hence the reason that D600 and 5D appear so close. The built-in camera profiles that produce jpeg are not so close.

Now where you may see difference is with 16-bit raw on MFDBs, which display smoother gradation, specially in skin tones.



MayaTlab
Registered: Jul 28, 2012
Total Posts: 52
Country: France

SoundHound wrote:
It appears that the $2100 D600 goes head to head, IQ and features, with the $3500 Canon 5D MK III.


When reading specifications, then yes. When holding them in your hands or manipulating their files, you start to see differences beyond what the specifications suggested. For example, I find it easier to see the entire VF frame with my eyeglasses with the D600. Or I find the 5DIII to handle quite substantially better, etc. Overall you do feel that the 5DIII deserves a higher price, although it's true it's lacking in a few areas in comparison to the Nikon (low ISO sensor performance comes to mind).



splathrop
Registered: Feb 27, 2006
Total Posts: 527
Country: United States

I am assuming that OP looked at the Raw version which is using Adobe's profile. Hence the reason that D600 and 5D appear so close.

Not sure how this relates. Can you say it again in other words? Maybe suggest a different way to look.

What I based my remarks on were not direct camera-to-camera comparisons of colors and densities. That's something I would have no idea how to do.

Instead, I peered at the Kodak color chart at various ISO values, looking to see how well each camera distinguished the most similar color pairs. On my monitor (which is a fairly good one), with my eyes, none of the cameras produced perfectly distinguishable boundaries between all the color squares, at least not without wiggling the image back and forth, and maybe not even then in some cases. Perhaps you can explain to me how RAW data that produces no visible difference between two samples from the same camera is nevertheless able to do so if processed differently, which is what I think you are saying. (And of course someone with a different monitor or better eyes (mine aren't bad) might not see it the same.

I thought the D600 and the 5D III were about evenly matched with regard to how often they failed, that they had their troubles in similar areas, but that each camera had strengths and weaknesses particular to it. They both seemed to edge the D800 at higher ISOs, but the the D600 did so by a larger margin—enough to make the difference notable. The D600 also seems to be the most noise-free at high ISO.

I would like to encourage others to take a close look and report what they find, because I don't pretend at all that what I see ought to be definitive. I would like to hear from as many folks as possible.

I should add that I presume the Kodak color chart is not an adequate test of dynamic range, given that nothing about the appearance of the colors and densities on the chart would lead anyone to believe what I take to be well demonstrated: that the D800 does a much better job of holding shadow detail than the 5D III.



bemyzeke
Registered: Jul 04, 2008
Total Posts: 384
Country: United States

splathrop wrote:
I am assuming that OP looked at the Raw version which is using Adobe's profile. Hence the reason that D600 and 5D appear so close.

Not sure how this relates. Can you say it again in other words? Maybe suggest a different way to look.


All cameras in question here use Bayer array. Which means that each pixel is recording luminosity of either Red, Green, or Blue light, with pixels recording green values twice as many as each of the other color. Hence no individual pixel records full color. It is only through interpolation (some mathematical algorithm) that a final color for each pixels is achieved.

There are two factors that may impact the color output. The first is the color of filters placed over each pixel. The color filters are part of sensor assembly and the hue is optimized to achieve best color fidelity for the given sensor. Although the filters are pretty standardized, there are minute differences that can vary from camera to camera even within manufacturer (e.g. the 5D.2 had a slightly different color filter array from the original 5D).

The second factor is the technique used to convert the individual RGB luminance value into a color for each pixel site. These techniques are built onto Raw conversion programs (whether in-camera or out-of-camera). Every Raw converter has its own profile; camera manufacturers actually provide multiple profiles.

What we see on dpreview site is simply the rendition that Adobe has perfected over time. They have a certain output fidelity in mind and so they vary their algorithm to adjust for the difference in camera color filters (although the differences are minimal). So it is for this reason that color output looks so close from each camera.

On the other hand if you take jpeg output from a camera created with landscape profile and compare it to jpeg out from even the same camera, created with standard or neutral profile, the colors will be wildly different, So, in summary, the colors that we see, are simply how the Raw files are "cooked" by Raw converters.

On top of that there is also the question of color space. The color fidelity (gamut) of modern cameras well exceeds that of sRGB, and Adobe RGB space. So to get the best colors out of a given profile you want a 16-bit TIFF file in ProPhoto color space. However jpg files (those served to our browsers) are mostly only sRGB and many LCD monitors struggle even with the limited color space of sRGB.

So then the question becomes, how well are the colors mapped into the limited color space of sRGB and then how well does the monitor render sRGB colors to its own gamut (this last point being related to your calibrated monitor).

This is why I believe all cameras are able to create all kinds of colors, the limiting factor are the skills of person processing the files. The only exception that I have seen (on web) are in high level medium format backs (such as those from Hasselbad and P-One) with 16-bit Raw. However I have never really held or tested MFDB myself to be sure about this.








Kittyk
Registered: Apr 29, 2009
Total Posts: 4405
Country: Germany

the difference is so small that it should not matter. what matters to me are the difference in ergonomics and built.
for amateur it should be more camera then he will probably ever use, pro on other side should value the built and extra features more then "some more" image quality.
imo



honorerdieu
Registered: Jul 08, 2003
Total Posts: 1170
Country: United States

They're just cameras.