CCD vs CMOS

[ErikKaffehr]

Hi,

Sony CMOS on P1 and CMOS on Sony are with some probability essentially the same. The MFD sensors are larger than the 24x36 mm sensors and that gives some advantage, at least under ideal conditions.

If files from Phase One using C1 are compared with Sony files processed in LR it should be kept in mind that C1 uses a lot of noise reduction and sharpening by default. LR/ACR have very little of each at default settings.

With the IQ3-100 MP, the sensor can produce DR in excess of 14-bits. Nikon and Sony sensors presently only use 14 (Nikon) or 13 (Sony) bits. So the IQ3-100MP does offer a small but possibly relevant advantage in DR over it's lesser brothers.

Best regards
Erik

I have been told by those who have shot CCD, CMOS on the P1 back and CMOS on the Sony's that on the P1 backs, both of the files are very nice. 

The files from the Sony's are nice, but not as robust or as nice.

[landscapephoto]

@landscapephoto: There has been an interesting test on this topic on the Reddotforum: http://www.reddotforum.com/content/2015/02/the-great-debate-ccd-vs-cmos-part-1/

It is not MF but compares the CCD Leica M9 with the CMOS Leica M240. A lot of direct comparison photos made with the same lenses.

Yes, and that is what I would have like to see for MF cameras. Only qualifier: for that comparison, all pictures were taken under the same light.

[landscapephoto]

the devil is in "compare the colours"    ...

I don't find this part to be a devil.

well, you can try for a start shot two different (pigments wise) targets chart properly, build a profile off one, do conversion of second and get witin 1-2 dE2K for all patches as measured by your spectrometer (many passes, averaged, lamp restored before each pass and drift controlled during each pass)... then may be, may be you are qualified to judge the color conversion precision for natural subjects beyond simple exclamations or adjectives... otherwise it is a matter of taste mostly.

You don't understand. I am not interested to find out which camera has better color accuracy. I am just interested to find out whether two cameras from the same maker, one with cmos and one with ccd, will give the same results.

[landscapephoto]

Sony CMOS on P1 and CMOS on Sony are with some probability essentially the same. The MFD sensors are larger than the 24x36 mm sensors and that gives some advantage, at least under ideal conditions.

Of course not. Even between two Sony cameras, you don't have the same color filter array. The same Sony cmos used in a Sony camera and in a Nikon camera give completely different results. I don't expect Phase One, Pentax or Hasselblad to use the same color filter arrays either.

[eronald]

thank you so much for sharing unprocessed raws, Synn... unless those were not actually raw files, which is what matters for the narrow subject of this topic.

For some reason "PROFESSIONAL PHOTOGRAPHERS" feel that they are giving away something valuable for free if they point a camera out the window and put the Raw file on Dropbox.

I think this is a foolproof way to distinguish a pro from an amateur: The pro always makes sure you pay

Edmund

[AlterEgo]

I am just interested to find out whether two cameras from the same maker, one with cmos and one with ccd, will give the same results.

and how you plan to eliminate color transform part from a raw converter for example ?

[AlterEgo]

Sony CMOS on P1 and CMOS on Sony are with some probability essentially the same.

but not the same CFA

from DT raws = http://postimg.org/image/bhxtviycl/

one 'd say that P1s CFA are more selective = hence better saturation with the min amount of color transform involved

[eronald]

but not the same CFA

from DT raws = http://postimg.org/image/bhxtviycl/

one 'd say that P1s CFA are more selective = hence better saturation with the min amount of color transform involved

eh? could you post the raw files somewhere so that one can look at them with rawdigger?

[ErikKaffehr]

Hi,

It is from Digital Transition's tests.

Best regards
Erik

eh? could you post the raw files somewhere so that one can look at them with rawdigger?

[ErikKaffehr]

moved here: http://forum.luminous-landscape.com/index.php?topic=110166.0

[landscapephoto]

and how you plan to eliminate color transform part from a raw converter for example ?

Why would I need to do that?

It seems that we do not understand each others. I have a CCD camera, an Hasselblad H4D-50. I like the colours I get out of it with Phocus. I don't get the same pleasing colours with, say, my Nikon.
Before I jump to, for example, an H6D-100c, I would like to know if it will give me the same colours. So I want to try the two cameras side by side on various subjects and under various lights and see if I get the same colours.

[Bart_van_der_Wolf]

It seems that we do not understand each others. I have a CCD camera, an Hasselblad H4D-50. I like the colours I get out of it with Phocus. I don't get the same pleasing colours with, say, my Nikon.

Hi,

The gist of some of the comments is, that the different rendering is not a CCD vs CMOS issue, but mostly a profiling issue. Silicon is silicon, and the effect of a chosen CFA (assuming for normal photography) is in practice of secondary importance compared to the effect of a profile. There are of course other reasons why colors differ, like lens coatings, etc., but none as important to the 'final look' as the profile.

Before I jump to, for example, an H6D-100c, I would like to know if it will give me the same colours. So I want to try the two cameras side by side on various subjects and under various lights and see if I get the same colours.

So, if you are using canned profiles, instead of making your own matched color rendering profiles for different cameras, I indeed suggest you compare what Hasselblad / Phocus have made of the 'Hasselblad look' and if it differs between the H4D-50 and the H6D-100c. If it differs, then it is more likely to do with the profiling, and as such can also be solved through profiling.

Cheers,
Bart

[ErikKaffehr]

Hi,

Agree with what Bart says. I would just add that it is my impression, based on what I read, that Hasselblad takes great care to achieve similar and good colour on their different backs.

Best regards
Erik

Hi,

The gist of some of the comments is, that the different rendering is not a CCD vs CMOS issue, but mostly a profiling issue. Silicon is silicon, and the effect of a chosen CFA (assuming for normal photography) is in practice of secondary importance compared to the effect of a profile. There are of course other reasons why colors differ, like lens coatings, etc., but none as important to the 'final look' as the profile.

So, if you are using canned profiles, instead of making your own matched color rendering profiles for different cameras, I indeed suggest you compare what Hasselblad / Phocus have made of the 'Hasselblad look' and if it differs between the H4D-50 and the H6D-100c. If it differs, then it is more likely to do with the profiling, and as such can also be solved through profiling.

Cheers,
Bart

[landscapephoto]

The gist of some of the comments is, that the different rendering is not a CCD vs CMOS issue, but mostly a profiling issue. Silicon is silicon, and the effect of a chosen CFA (assuming for normal photography) is in practice of secondary importance compared to the effect of a profile. There are of course other reasons why colors differ, like lens coatings, etc., but none as important to the 'final look' as the profile.

I agree that silicon is silicon and that the results should be the same if CFA, profiles and lenses are the same.

I also agree that the profile gives the final look.

I do not agree that the effects of the CFA are of secondary importance, in particular when the illuminant varies.

I am quite sensitive of the effect of lens coatings on color and find these to be very difficult to compensate with profiles, because they are not global.

So, if you are using canned profiles, instead of making your own matched color rendering profiles for different cameras, I indeed suggest you compare what Hasselblad / Phocus have made of the 'Hasselblad look' and if it differs between the H4D-50 and the H6D-100c. If it differs, then it is more likely to do with the profiling, and as such can also be solved through profiling.

I am not going to make profiles to solve a problem which I can avoid by not buying a new camera. I have a camera with colors I appreciate. If the new model does not have the same colors, I won't buy it. It is a much simpler solution.

[eronald]

Why would I need to do that?

It seems that we do not understand each others. I have a CCD camera, an Hasselblad H4D-50. I like the colours I get out of it with Phocus. I don't get the same pleasing colours with, say, my Nikon.
Before I jump to, for example, an H6D-100c, I would like to know if it will give me the same colours. So I want to try the two cameras side by side on various subjects and under various lights and see if I get the same colours.

Go for it. Only realworld trials will tell you what you need to know.
I wouldn't expect the "same" colors, if the CFAs are different, that is physically impossible. (*). But you may get colors you like.
Think of it like choosing a film - some one likes, some one doesn't.

Edmund

(*) No existing consumer sensor satisifies the Luther-Ives conditions. People who claim two different CFAs and substrates can be *perfectly* matched by a profile are delusional.

[AlterEgo]

-DNG Profiles built for for each using Adobe DNG Profile Editor

...

The idea with this was to apply the same processing chain to both images using colour profiles generated from the actual images.

and the base profile for Adobe DNG PE was ? because Adobe DNG PE starts not from scratch - hence the processing chain is the same but w/ that caveat

in C1 example above I used ICC profile = "no color correction" for both cameras, etc to make it more the same processing colorwise

[ErikKaffehr]

Hi,

I don't know.

I may repeat the experiment using DCamProf…

Best regards
Erik

and the base profile for Adobe DNG PE was ? because Adobe DNG PE starts not from scratch - hence the processing chain is the same but w/ that caveat

[rogerxnz]

Reverting to the original point of this thread and the difference between CCD and CMOS sensors, I understand there is doubt whether CMOS sensors can handle wide-angle lenses as well as CCD sensors can. I also understand tilts and shifts may be more limited with CMOS sensors than with CCDs.
Roger

[alifatemi]

I wonder if PhaseOne IQ100 CMOS got anti aliasing filter on it. Anybody please?

[Paul2660]

I don't believe any Phase CCD or CMOS back has an anti aliasing filter, for sure the CCD IQ1 and 2 did not.  Pretty sure the 100MP is the same.

Paul C