Canon 5Ds and 5Ds R samples at flickr

[JoeKitchen]

I will say that Sinn is pretty pro MF, but I did not get that he was talking up MF in his recent post. 

To me it was more "why not compare this new Canon to the other full frame alternatives, namely the Nikon and Sony," albeit not in the most tactful way. 

Why?  Probably because it is universally accepted that the Nikon and Sony versions are better. 

[ErikKaffehr]

Hi,

Better is an interesting term. The way I see it Nikon and to a lesser degree Sony have two key advantages, namely higher DR at low ISO and resolution up to 36 MP. The resolution advantage turns around with the 5Ds, albeit it is possible that Nikon or Sony have new high resolution cameras on market before then 5Ds is available.

But, I don't think it is proven that image quality from Nikon is higher at medium to high ISO.

I know two photographers who have both Nikon D800 and Canon 5DIII and both use both systems.

Personally I shoot Sony and P45+ so, I have no pony in that race. I shoot normally base ISO and I am quite resolution friendly and a piece of DR is always welcome.

Best regards
Erik

Why?  Probably because it is universally accepted that the Nikon and Sony versions are better. 

[Hans Kruse]

Whenever I saw this kind of arguments I always asked for RAW files for a comparison to prove why something like a D800E cannot achieve the skin tone of a 5D3 or an H4D-40's. No Canon or CCD user has ever shared any RAW files to prove it. I guess this is just common sense so no Canon or CCD user needs to prove it. 

The Bayer filters are different for different cameras and I shoot Nikon and Canon side by side and also had a Phase One IQ160 for a year. They are all different and calibration does help each camera, but they will not be the same in real photos. I shoot mostly landscapes and I got very different colors from each of these cameras. Setting them to the same WB does not solve it. I can sometimes get close by changing WB manually to approach another photo. It's not in my opinion so that one is good and others from the same scene not good. They can be good each of them in their own rendering of the same scene.

Attached are two photos that I tried to match. The first is from the Phase One IQ160 and the second Canon 5D mkIII.

[Hans Kruse]

Hi,

Better is an interesting term. The way I see it Nikon and to a lesser degree Sony have two key advantages, namely higher DR at low ISO and resolution up to 36 MP. The resolution advantage turns around with the 5Ds, albeit it is possible that Nikon or Sony have new high resolution cameras on market before then 5Ds is available.

But, I don't think it is proven that image quality from Nikon is higher at medium to high ISO.

I know two photographers who have both Nikon D800 and Canon 5DIII and both use both systems.

Personally I shoot Sony and P45+ so, I have no pony in that race. I shoot normally base ISO and I am quite resolution friendly and a piece of DR is always welcome.

Best regards
Erik

I'm probably one of them More DR is definitely welcome in landscape photography and some other genres as well. Even though the Nikons have a much higher DR there are a number of examples I have come across that would benefit from blending of exposures. Attached is a screen shot from two exposures. The clean one is from the merged result which is 0EV and the right hand one with the noise is at -3EV. In other parts of the picture serious loss of resolution and noise is the result at -3EV. The Canon would me much worse, of course. But the point is to state that even with better DR there is still many shots out there where the DR is not high enough  to get a clean and noise free picture. Clearly this is viewed at 1:1, but a very large print would reveal the lack of resolution and noise. The second attachment is the blended result.

I plan to get the Canon 5Ds R and I know I will need to blend exposures more than the Nikon, but it will be needed on both and for landscapes it's not really a big deal to do it.

[voidshatter]

The Bayer filters are different for different cameras and I shoot Nikon and Canon side by side and also had a Phase One IQ160 for a year. They are all different and calibration does help each camera, but they will not be the same in real photos. I shoot mostly landscapes and I got very different colors from each of these cameras. Setting them to the same WB does not solve it. I can sometimes get close by changing WB manually to approach another photo. It's not in my opinion so that one is good and others from the same scene not good. They can be good each of them in their own rendering of the same scene.

Attached are two photos that I tried to match. The first is from the Phase One IQ160 and the second Canon 5D mkIII.

If you shoot raw and shoot color passport then color calibration in post-processing would resuilt in nearly indistinguishable pictures between different cameras. This is true for portrait as well. For printing purposes e.g. magazines, you ought to calibrate the skin tone in accordance with some CMYK guidelines for specific races of people. My point is that you should be able to achieve the same color with virtually any decent camera that can shoot RAW.

[voidshatter]

I'm probably one of them More DR is definitely welcome in landscape photography and some other genres as well. Even though the Nikons have a much higher DR there are a number of examples I have come across that would benefit from blending of exposures. Attached is a screen shot from two exposures. The clean one is from the merged result which is 0EV and the right hand one with the noise is at -3EV. In other parts of the picture serious loss of resolution and noise is the result at -3EV. The Canon would me much worse, of course. But the point is to state that even with better DR there is still many shots out there where the DR is not high enough  to get a clean and noise free picture. Clearly this is viewed at 1:1, but a very large print would reveal the lack of resolution and noise. The second attachment is the blended result.

I plan to get the Canon 5Ds R and I know I will need to blend exposures more than the Nikon, but it will be needed on both and for landscapes it's not really a big deal to do it.

Bracketing and blending is not always a viable option for me. For pixel peeping purposes alignment issues could be of a real problem. See more details here: link

[Hans Kruse]

If you shoot raw and shoot color passport then color calibration in post-processing would resuilt in nearly indistinguishable pictures between different cameras. This is true for portrait as well. For printing purposes e.g. magazines, you ought to calibrate the skin tone in accordance with some CMYK guidelines for specific races of people. My point is that you should be able to achieve the same color with virtually any decent camera that can shoot RAW.

Not the case! As I said, I calibrate all my cameras and of course shoot RAW as otherwise the calibration would not make sense using Lightroom.

[torger]

Not the case! As I said, I calibrate all my cameras and of course shoot RAW as otherwise the calibration would not make sense using Lightroom.

If you make a custom DCP profile for two different cameras under the same light condition they will indeed match very well, voidshatter is correct in that. That is if you shoot portraits in a studio with the same flashes you can make two different cameras virtually indistingiushable from eachother. There are slight differences in metamerism and color separation, but it's generally negligible, as long as a camera can separate colors a profile can correct the color into any look you want.

However, if you then take those same profiles and shoot a landscape scene with a different illuminant, that golden light in your posted image for example, and adjust white balance to taste the different cameras will modulate the color differently so they will then not match, so you are correct too.

That is to make two different cameras (different color filters on the sensor) match up you need to make a profile for each lighting condition, ie not practical in landscape, but is for studio work.

[Hans Kruse]

If you make a custom DCP profile for two different cameras under the same light condition they will indeed match very well, voidshatter is correct in that. That is if you shoot portraits in a studio with the same flashes you can make two different cameras virtually indistingiushable from eachother. There are slight differences in metamerism and color separation, but it's generally negligible, as long as a camera can separate colors a profile can correct the color into any look you want.

However, if you then take those same profiles and shoot a landscape scene with a different illuminant, that golden light in your posted image for example, and adjust white balance to taste the different cameras will modulate the color differently so they will then not match, so you are correct too.

That is to make two different cameras (different color filters on the sensor) match up you need to make a profile for each lighting condition, ie not practical in landscape, but is for studio work.

Thanks. I use one profile for all my landscape work. I have not done a lot of testing with different profiles for landscape, but have shot a couple and saw some differences. But as you mention I did not find this practical to do.

[Hans Kruse]

Bracketing and blending is not always a viable option for me. For pixel peeping purposes alignment issues could be of a real problem. See more details here: link

I recently started a thread on blending here and although I agree alignment can be an issue for the landscapes I'm shooting I have been able to use this method for all examples I had in my folders where I previously used HDR programs.

Some examples

[Bart_van_der_Wolf]

I'm probably one of them More DR is definitely welcome in landscape photography and some other genres as well. Even though the Nikons have a much higher DR there are a number of examples I have come across that would benefit from blending of exposures. Attached is a screen shot from two exposures. The clean one is from the merged result which is 0EV and the right hand one with the noise is at -3EV. In other parts of the picture serious loss of resolution and noise is the result at -3EV. The Canon would me much worse, of course. But the point is to state that even with better DR there is still many shots out there where the DR is not high enough  to get a clean and noise free picture.

Hi Hans,

A small addition, the crops you showed were not extremely underexposed, but of a rather average brightness part of the scene (although -3EV is not something normally done unless to avoid highlight clipping elsewhere in the image). Nevertheless, we will almost always benefit from collecting more photons, even in the non-shadow regions (although shadows will benefit most). That's what one can see in the crops, more than a lack of sensor DR (except for the extreme total range). I do understand that the -3EV was to cope with total scene DR.

This indeed means that exposure bracketing, whenever practical, offers a benefit (regardless the Camera DR capabilities) also in mid-tones. That also illustrates a common misconception about HDR photography. It is not only to capture a wider range of brightnesses than the sensor can handle in a single exposure, it's more about collecting as many photons as possible, in all levels of scene brightness. The rest is tonemapping ...

Cheers,
Bart

[torger]

Thanks. I use one profile for all my landscape work. I have not done a lot of testing with different profiles for landscape, but have shot a couple and saw some differences. But as you mention I did not find this practical to do.

I think that's the best method, and as far as I understand the established "best practice", you use one profile (a daylight profile) and tune white balance to taste. While accurate color is possible in a careful reproduction setup, it's impossible in landscape and it's not important either as we want to create subjective color anyway.

However the manufacturer profiles are generally huuuuugely subjective which means cameras seems to differ a lot more than they actually do, so making your own profile is a great way to "take command" of your color, and make it possible to switch both cameras and raw converters with less differences than it otherwise would be.

I think it's the better technique to start with a fairly neutral daylight profile (ie one you make yourself) and add your subjective interpretation of color in your own post-processing, ie sitting in the front seat instead of letting the manufacturer drive your color...

[Hans Kruse]

Hi Hans,

A small addition, the crops you showed were not extremely underexposed, but of a rather average brightness part of the scene (although -3EV is not something normally done unless to avoid highlight clipping elsewhere in the image). Nevertheless, we will almost always benefit from collecting more photons, even in the non-shadow regions (although shadows will benefit most). That's what one can see in the crops, more than a lack of sensor DR (except for the extreme total range). I do understand that the -3EV was to cope with total scene DR.

This indeed means that exposure bracketing, whenever practical, offers a benefit (regardless the Camera DR capabilities) also in mid-tones. That also illustrates a common misconception about HDR photography. It is not only to capture a wider range of brightnesses than the sensor can handle in a single exposure, it's more about collecting as many photons as possible, in all levels of scene brightness. The rest is tonemapping ...

Cheers,
Bart

I did not quite explain enough in that post (#103). The screen shot was from two pictures from a bracket sequence intended to protect the highlights and expose for the shadows as well. The screen shot was from Lightroom showing a compare between the merged result where what was shown was an entirely 0EV picture on the left hand side and a -3EV on the right hand side. The -3EV side had been calculated by the Lightroom Match Total Exposure function and came to 2,34. The editing of the two pictures from the bracket sequence was exactly the same with only the exposure setting in Lightroom different. The -3EV was at +1,82 and for the 0EV it was -0,52. Therefore you could see them very closely the same except for noise and loss in details. So in the merged image only the clipped parts from the 0EV was replaced by the content of the -3EV.

Again notice that this example was from the Nikon D800E. A Canon one would normally be much worse. Comparing a merged result from the Canon with a non merged result from the Nikon shows a huge advantage in IQ as it has collected so much more light (as you say). Given the method I'm now HDR software will very seldomly be used by me. Using Photoshop and the Lightroom adjustment of exposure to match two exposures I can now get an end result that looks exactly as if the camera magically could expose different strength of light differently. I'm just wondering when it would be possible to have a "shutter" per pixel so we can choose how much light to gather depending on the lighting in different parts of the image. Information from the sensor of the amount of light hitting different pixels could be fed to the RAW converter or could be calculated into what was stored in the RAW file.
That would for me be ideal landscape photography.

[Hans Kruse]

I think that's the best method, and as far as I understand the established "best practice", you use one profile (a daylight profile) and tune white balance to taste. While accurate color is possible in a careful reproduction setup, it's impossible in landscape and it's not important either as we want to create subjective color anyway.

However the manufacturer profiles are generally huuuuugely subjective which means cameras seems to differ a lot more than they actually do, so making your own profile is a great way to "take command" of your color, and make it possible to switch both cameras and raw converters with less differences than it otherwise would be.

I think it's the better technique to start with a fairly neutral daylight profile (ie one you make yourself) and add your subjective interpretation of color in your own post-processing, ie sitting in the front seat instead of letting the manufacturer drive your color...

Yes, that's basically what I do. I almost always adjust the WB for my landscape shots from the AWB coming from the camera based on what I like the picture to look like and then I adjust exposure, contrast, saturation and do local adjustments to get to the look I'm after. This depends a lot of which time of day the shot was taken and the emotional characteristic of the scene and what I'd like to say with the picture. So very subjective as you say. But I find that starting with a calibrated profile is much better than fight against a profile that looks wrong to me from the beginning. Sometimes the difference to Adobe Standard from the calibrated profile is not that much but at other times is much bigger.