[barondla]

Comparing cmos and ccd color. Used the Pentax DA 16-45 lens on both the cmos sensored K-1 and ccd equiped K-10D. Both set to natural color balance, 100 iso at 45mm. Tried to take identical shots. Unfortunately, a small error crept in. One image is at f4.5 the other at f5.6 It was an overcast day, but the window light did change slightly. Hopefully that doesn't affect color much. Shot DNG Raw and converted both in Photoshop Elememts. The K-1 automatically cropped the lens to match the K-10D.

Thanks for looking,
barondla

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[VSTAR]

The difference in sensors has been discussed in many different forums. The end result is that it is very difficult to tell the difference between them if all other camera factors are similar as possible.

This discussion may help.,,,
Pentax 645D, or 645Z: CCD vs. CMOS

[Paul the Sunman]

A little more saturation in the K1 image (on my screen). Otherwise, less difference than I would have expected.

[pinholecam]

Personally, I think that CCD and the sensor tech at its time had less DR and hence more contrast.
The result of it is taken as deeper colors and more contrast (ie. 'film like')

I'd take the K1 o/p any day and add the contrast with a pre-set in LR if I wanted some close to the same look.

[ChristianRock]

It seems like the K10D image picked up a bit of a reflection on the clock.

I can see different tones in the colors... so the question is, which one is more accurate? Which one looks more like the original colors?

[Digitalis]

I didn't have to look at the Exif to tell which was which, on my calibrated monitor the colour response of the K10D is distinctive.

[Grippy]

I didn't have to look at the Exif to tell which was which, on my calibrated monitor the colour response of the K10D is distinctive.

Agreed, with my new monitor I can see the colour difference with the K10 pics, old monitor I can't
Swings and roundabouts, we could argue this all day on many different threads.

[kiwi_jono]

There are differences, but its pretty much impossible to tell how much is due to sensor and how much is to do with camera colour profiling (just because its got the same colour setting doesn't mean they will match).
The K-1 image appears a little less contrasty too.
To be honest I'm not a 100% sure which one I like best in these conditions but I am sure which camera I would prefer at higher ISO.

[Ian Stuart Forsyth]

Most of the time was see any differences in between the 2 sensor is more to do with the raw converter than anything else, more importantly the color profiles used for that conversion. More or less these profiles and the tonal curves that are applied are based on what is pleasing for that time and the taste of the developer.

When we take out the differences in the color profiles the color differences are rather small. This is not to say that you wont see any differences but most of the time its things like DR and shadow noise that you will see.





The images are crops from CCD and cmos

[Dartmoor Dave]

For those of us who choose to keep using CCD sensors so long after their supposed obsolescence, it isn't really about direct head-to-head comparison shots like this. After all, you could easily say that the differences in the example shots are so tiny that they shouldn't matter, and that with a bit of processing they'd look even more alike.

But over thousands of shots it really does matter. When you love the rendering style that CCD sensors give you straight off the camera, every CMOS shot that you have to throw bucketloads of PP at to put some oomph in it gets more irritating than the last. Why should I have to spend time in front of a computer processing photos to look almost-but-not-quite the way I want them with a CMOS, when I can have them almost exactly the way I want them straight off the camera with a CCD?

I've got a CMOS camera for those times when I need high ISO, but whenever ISO 100 is an option I use a CCD. And I don't ever print big enough that the advantage of an FF sensor would outweigh my preference for the CCD rendering style. Also, I do think it's down to more than the raw conversion profiles used. I always use my own custom profiles created exactly to make the results from my different cameras look as close as possible, but no matter what I do the difference between CCD and CMOS is always there. I can make it smaller but I can't make it go away, much as I wish I could.

Most people nowadays seem to prefer CMOS and they're absolutely entitled to hold that preference, but some of us prefer CCD and we're entitled to hold that preference too. As small as the differences might seem to be in comparison threads like this one, over thousands of shots they can end up seeming like very significant differences indeed.

[Wasp]

No problem in seeing the differences on my uncalibrated laptop screen. I liked the K10D image better and initially thought that it was from the K1. Progress, wherefore art thou?

Then again, the slight change in light may explain the change in colors.

[Rondec]

The difference is tiny and as Ian mentions, is probably mostly due to the Raw Converter used more than anything else. I liked the K10 fine back in the day and never actually liked the colors of the K20 and K7 very much (although they were better at high iso), but from the K5 on I have been happy with the color rendering of the CMOS sensors and have no particular desire to go back.

[Serkevan]

Uncalibrated (I guess, lol, never bothered to check) phone screen here: the difference is obvious on the yellows and browns, between subtle and inexistant elsewhere. Using Adobe Standard as colour profile (the K-1 image certainly looks like it) tends to have a bit lower contrast and saturation - but you do get more detail. I sometimes change to Camera Embedded/Bright/Portrait depending on whether I want some more punch, but that has a tendency to smear details in highlights.

[Digitalis]

more importantly the color profiles used for that conversion

Colour mismanagement can cause absolute havoc when doing camera comparisons too: Especially newer cameras that are released and Adobe get caught with their pants down without a reliable profile for* this can really make a mess of things especially when it comes to seeing the differences between two different sensor architectures - this can have a severe impact on initial findings when comparisons are made.

The difference is tiny and as Ian mentions, is probably mostly due to the Raw Converter used more than anything else.

Quite right, At the time In the studio I would have to light models with significantly softer light with the K10D. When I bought the K7 I could afford to use slightly harder lighting, and when the K5IIs came along, it brought dynamic range that could handle it if I needed to give the shadows/darker mid-tones a lift without introducing significant amounts of noise like it would on the K10D/K7. This trait follows with the 645D and 645Z**, my lighting choices have changed somewhat. The need to quad diffuse my softlights has been reduced as the dynamic range of the 645Z gives tremendous flexibility in how tonal ranges can be manipulated. Though I do prefer the subtle nature of the 645D's CCD rendering, Whether it is an artifact of the RAW engine or a complexity derived from the spectral bandpass of the CFA layer used with the CCD...it takes a practiced eye to spot it at times, but there is a subtle depth to colours on CCD cameras. Many of my clients have well developed colour perception so this makes the qualities of the straight out of camera output from whichever camera i'm using of considerable importance, first impressions count for a lot.



* Fortunately for me, I am capable of developing my own highly precise camera profiles.

** This change was also mirrored with the Leica S2 which made a similar change from CCD to CMOS.

[ChristianRock]

There's things that are hard to put in proper words and explain...

Take the flower under the clock for example. It seems to me quite clearly that the CCD picked up more tonality variations of the soft reddish/pinkish colors, and as a result the CMOS rendering of the petals is a bit more "flat" (please forgive me if I'm using the wrong terminology). It seems like there's some in-between tones that the CMOS doesn't pick up as easily, when talking about pinkish color tones. I can clearly see it in the greens as well.

This seems to agree with a poster in another thread where he kept swapping the same lens between two cameras (CMOS and CCD) during "pink hour" and the CMOS did not seem to pick up most of the pink tones in the clouds.