Comparing Apples to Grapefruit IQ180/D800E Crops

[marcmccalmont]

More RAWs
Marc
https://dl.dropbox.com/u/82427039/CF000681.IIQ
https://dl.dropbox.com/u/82427039/CF000685.IIQ
https://dl.dropbox.com/u/82427039/DSC_0565.NEF
https://dl.dropbox.com/u/82427039/DSC_0575.NEF

[ErikKaffehr]

Hi Marc,

Thanks for posting the "Color Checker" images. I tried to build a DNG profile for both the IQ180 and the D800 using the Xrite Passport software. The Color Checker DC you used contains also the normal color checker patches. I checked out the flower image with both the Adobe Standard profile and the one I generated, they were very similar, except in the yellows. (Before/After sample)

I also tried to generate a corresponding profile for the IQ180, but the X-rite software blow up, out of memory. Will look a bit more when back from vacation.

My impression from what I have seen this far is that the D800E seems to be an excellent value for the price if the owner strives for image quality. The IQ180 has more pixels and it shows. I assume that both equipments need excellent technique to extract maximum performance.

Nikon owners are now lucky enough to have a very affordable high resolution camera.

Thanks a lot for sharing experience and images, I just wish more MFDB owners would do that!

I hope that enjoy your flight with the D800E!

Best regards
Erik

I just started a 5 day trip when I return I'll post them
Marc

[marcmccalmont]

I hope that enjoy your flight with the D800E!

Erik
I will! if you have any other shots in mind let me know
I think the D800E is having fun too!
Marc

[ErikKaffehr]

Hi,

That's what I call a viewfinder! have you seen any nice subjects lately?

Best regards
Erik

Erik
I will! if you have any other shots in mind let me know
I think the D800E is having fun too!
Marc

[ErikKaffehr]

Hi,

Marc posted the original raw files for the CC DC shots. The DC version of Color Checker contains the original CC-patches, too, so I could make a camera calibration. Unfortunately Adobe's DNG Profile Editor would not accept the Nikon image (because of slight overexposure) and the Color Checker Passport software failed on the IQ180 file, because of size. So I made calibration with two different tools.

After calibration I adjusted the white patch to 95% and the black patch to about 20% which I think is about the normal values.

The screen dumps shows the standard CC patches. I also added a calculated color checker image from the Bruce Lindblom site. That is the correct colors. Lightroom will normally render all colors to "vivid" and add some "S-curve". I don't know how to do scene referred rendering in LR PV 2012, unfortunately. I will be back with some more analysis, using Imatest.

Best regards
Erik

[ErikKaffehr]

Update: figures remade with correct RGB


Hi!

My first try of scene referred rendition using Imatest. I'm not very experienced on this so I may have made some mistakes.

Best regards
Erik

[torger]

The color errors in deltaE indeed seem very large, much larger than the deltaE one sees when calibrating a screen for example. Perhaps some problem somewhere in the process? Or maybe cameras really are this bad?

[ErikKaffehr]

Hi!

I just updated with new figures (wrong color space was used), sorry.

Will look more into the issues in the coming days...

Update: Cameras are pretty bad. This depends on striving for pleasant color, thus increasing saturation. I tried to achive scene referred rendition, but may have failed.

Best regards
Erik

[torger]

Update: Cameras are pretty bad. This depends on striving for pleasant color, thus increasing saturation. I tried to achive scene referred rendition, but may have failed.

Dxomark too makes some check of color accuracy, and your differences between the cameras seem larger than what they come up with.

What's interesting to know is what the difference between these cameras are after profiling. Does the IQ180 actually have better color rendition or not? So far it seems the answer is yes. If a camera has issues with the color filters there can be lots of metamerism and be impossible to correct color with a profile.

It may also be the case that manufacturers have optimized the filters to be good for "important" colors, say skin tones, and let them be less accurate in other colors. It would be interesting to investigate such differences too, i e which colors are accurate and which are not.

I don't think Lightroom is a very good tool to make evaluations though. I have got the sense that Capture One cares more about accuracy and Lightroom more about a "pleasing look". If you use RawTherapee you can get truly neutral renditions if you want, even without gamma curve.

[ErikKaffehr]

Thanks for suggesting RawTherapy!

Best regards
Erik

[shadowblade]

I don't quite understand the importance of 'colour accuracy' - at least for landscape photography.

I mean, every time I put a polariser in front of my lens, or do almost any sort of postprocessing, or select a different white balance point, the colour changes anyway.

I can see how it would matter if you were designing a scanner or reproducing artwork, but for photography?

[Rob C]

I don't quite understand the importance of 'colour accuracy' - at least for landscape photography.

I mean, every time I put a polariser in front of my lens, or do almost any sort of postprocessing, or select a different white balance point, the colour changes anyway.

I can see how it would matter if you were designing a scanner or reproducing artwork, but for photography?

That's my situation re. my 'mountain' pics that are showing their noses on 'Prejudice' at the moment: they were all shot through a Nikon Polariser which I didn't remove even when shooting almost into the light for the simple reason that there was nowhere scratch-proof to stick it had it been removed. Unfortunately, some of the effects of using it appear to be different on digital to what I'd have expected with Kodachrome or Velvia, especially in skies, which don't seem to get just darker, but with an odd tone of colour - greyish - that's difficult to change and not make worse. Maybe another time - should there be one - I'll stick with the Skylight filter.

Rob C

[torger]

I don't quite understand the importance of 'colour accuracy' - at least for landscape photography.

I mean, every time I put a polariser in front of my lens, or do almost any sort of postprocessing, or select a different white balance point, the colour changes anyway.

I can see how it would matter if you were designing a scanner or reproducing artwork, but for photography?

I tend to agree. I have never worried much about color accuracy. However, if you are a portrait photographer you probably want nice skin tones, and accurate colors is a good start, I guess. It is also about how many colors that can be registered. If two slightly different colors are registered by the sensor as the same color (metamerism), then no color profiling can restore them accurately.

It is also think it is interesting to investigate this property since it may be one area that MF still has a lead. And if it can be shown that D800 is as good as the IQ180 also in color, then it is also an interesting result.

[marcmccalmont]

I tend to agree. I have never worried much about color accuracy. However, if you are a portrait photographer you probably want nice skin tones, and accurate colors is a good start, I guess. It is also about how many colors that can be registered. If two slightly different colors are registered by the sensor as the same color (metamerism), then no color profiling can restore them accurately.

It is also think it is interesting to investigate this property since it may be one area that MF still has a lead. And if it can be shown that D800 is as good as the IQ180 also in color, then it is also an interesting result.

When I first got my P30 and converted files using C1 the foliage was a bit pea green compared to my 5D/DxO so I took a day shooting my front yard with both and ended up with 2 curves the 5D green channel reduced a bit and the P30 green channel increased a bit both looked more accurate and pleasing afterwards.
Marc