Hmm I don't think I can get more bare boned than Canon's own DPP. However if I use another RAW developer the results will be different - perhaps in another way, but still different for the two cameras.
Anyway, here are the images processed by DPP with the same setting except the white balance which I set to the same point on the white box below the camera.
Edit: I used neutral picture profile, hence the flatness compared to ACR where I used the "Camera default" profile. The relative difference between the 7D and 5DII is still there though.
the color looks better. still a little off because of the difference in angle, but nothing to be done about that.
i am puzzled by the difference in local contrast - i can't think of any logical reason why a FF crop should have any higher local contrast than the cropped image.
I think it's because of better color separation of the FF sensor due to a lower pixel density. High pixel density => leaks to neighboring pixels => worse color separation => lower local contrast.
denoir wrote:
Hmm I don't think I can get more bare boned than Canon's own DPP. However if I use another RAW developer the results will be different - perhaps in another way, but still different for the two cameras.
Anyway, here are the images processed by DPP with the same setting except the white balance which I set to the same point on the white box below the camera.
Edit: I used neutral picture profile, hence the flatness compared to ACR where I used the "Camera default" profile. The relative difference between the 7D and 5DII is still there though....Show more →
Kewl! Thanks for doing this! Mightily appreciated!
My own personal conclusion? By this example alone I'll never pay "the bucks" for a FF sensor. Those are so close to identical it's just silly. What little difference there is could easily be explained just by the difference in camera angles maybe. As I bring such a photo in thru ACR just a tad of any one of several sliders one way or the other could throw the results in either direction.
Of course like you say this might not be the best example - but if I were to judge just by this...
I am with Luka on this one. Perhaps the test doesn't really illustrate the full difference as a landscape shot would. If you look at my shots and compare to Cogitech's or Luka's you will see a staggering difference in depth and tonal range.
PP won't help much...
Cavewalker - my 1st M&M bokeh is favourite!
Paul, your wife is fantastically beautiful young lady. My sincere compliments!
Is it your daughter in 2nd shot?
Luka - really beautiful sunflowers...
And everyone else I didn't credit!
Helimat: Tsk, tsk..I remember all the vinyls I carry...:/
I'm currently thinking it's all PP myself. I mean you're comparing totally different shots by totally different photographers and probably with completely different processing methods. Maybe a proper/better test would show it differently but right now that's my take. Of course I could always be wrong. I probably am but just going by the evidence presented so far I can't say so.
I wouldn't claim you are wrong. Just perception and perspective which are completely individual...
I also thought it was it at some point. But the difference became really obvious recently....
Colour and contrast are correctable - no doubt about that. But the tones which are not captured by the sensor are lost for good and taken the detail with them...
I don't think it's the FF issue, it's more of the photo site density of a sensor, but it so happens that current FF sensors have more of those compared to APS-C and APS-H...
Bifurcator wrote:
My own personal conclusion? By this example alone I'll never pay "the bucks" for a FF sensor. Those are so close to identical it's just silly. What little difference there is could easily be explained just by the difference in camera angles maybe. As I bring such a photo in thru ACR just a tad of any one of several sliders one way or the other could throw the results in either direction.
Of course like you say this might not be the best example - but if I were to judge just by this...
Interestingly, if more photos show similar difference, I would be convinced to go for 7D. The difference, IMHO, is due to micro-contrast which no pp can compensate for. When comparing Canon 35mm 1.4 (at f2) vs. Zeiss 35mm f2 or Leica 35mm f2, I saw similar difference. The test photo was a white rose in reasonably bright but not harsh light. The rose petals just look flat with Canon. They are of homogeneous white, and no pp could make it match Zeiss or Leica counterpart.
Bifurcator wrote:
I'm currently thinking it's all PP myself. I mean you're comparing totally different shots by totally different photographers and probably with completely different processing methods. Maybe a proper/better test would show it differently but right now that's my take. Of course I could always be wrong. I probably am but just going by the evidence presented so far I can't say so.
It also may come down to how sensitive you are to the differences. I seem to remember a similar discussion of a Zessi vs Canon lens that we had
To me from the images posted in this threads in terms of dynamic range and tonality show a very clear difference between a 5DII, a 20D and a GH1. As for differences in PP, it's rather simple - as a rule I don't do any except sharpening and setting white balance. I use the default settings in ACR.
So following that fact we have the following options:
1) I'm mistaken by thinking there is a difference between FF, 1.6x crop and 2x crop images in this thread.
2) There is a difference as a result of sensor size
3) There is no difference but everybody but me postprocesses their images in such a way that they end up looking worse than what default settings produce
1 is certainly possible but as others have expressed that they see the same difference, I'd judge it as unlikely. 3 I think is extremely unlikely. Therefor 2 remains as the probably correct answer.
So if we assume that there is a difference, can it be fixed for in PP? Well, assuming I'm not mistaken about the images posted here, the practical answer is no. Even if you could in theory fix it, apparently people can't be bothered or don't know how to fix it. How can we otherwise explain the images posted here where no crop sensor images display the tonality of the FF images?
Furthermore to adjust one image to look like another is a relatively simple task when you have a target image. There are many art students that can make a very decent Rembrandt reproduction. Very few can make something as good from scratch. No difference here - without a target to emulate it becomes much more difficult to get to the same result.
denoir wrote:
It also may come down to how sensitive you are to the differences.
I think that's it actually. When I look at the two images you're offering up for comparison I actually like the crop sensor image better. When scrolling through this thread I see some I'm unattracted to, some I like, and some that wow me - (all technically speaking as is the current subtopic at hand). It doesn't seem to matter what camera they were from. The ones that wow me are all to do with the scene lighting and secondarily the color theme-play. There are other things that contribute to the wow factor as well like, texture, framing, composition, and even the subject/scene itself. But whether by ignorance, insensitivity, or a greater concern for the message I seem to be completely oblivious to the differences in sensor yield. On my A2 2/3rds P&S I could see a difference in DR (between it and a FF camera) but felt that it usually wasn't anything I could not compensate for in post - most of the time. On the GH1 I can't see the differences at all. Yeah, "completely oblivious" sums it up pretty well I think.
I also think I should probably keep my oblivion in tact as the price tags associated with FF cameras give my wallet a hernia just thinking about.
Bifurcator wrote:
I also think I should probably keep my oblivion in tact as the price tags associated with FF cameras give my wallet a hernia just thinking about.
If you want to push that a bit, just enough to become suicidal, you should check out some Hasselblad H4D-60 shots. It makes the 5DII look like an el-cheapo webcam. Price tag $40,000
denoir wrote:
If you want to push that a bit, just enough to become suicidal, you should check out some Hasselblad H4D-60 shots. It makes the 5DII look like an el-cheapo webcam. Price tag $40,000
Yeah, I know you're right on that one. I have more than one friend who own and operate photo studios and I've had the pleasure of using their MF equipment on occasion. Good stuff! That was the point I was making on the previous page too. If I were a working professional based out of a studio and making the bucks I'd have me a FF for field work and a MF D-back in the studio. I do of course see the differences in scene resolve - just not DR (FF vs u4/3). I think that's what Dimity is seeing as well given his description. For MF vs. u4/3 I see the differences in DR at that point. And the two work together as well! Mathematically speaking, if you project a gradient across a FF sensor with only 4 photocites you get a 4-step reading (4 sampled points). More photocites = more steps/samples where the tonal range is more accurately sampled.
This is a contributor to DR but DR is different yet right? At least I think DR has more to do with the physical size of the sensor - and also the capabilities of the lens too. I think it's like, the difference between stretching a given tonal range over the distance of 7cm and 2cm (MF vs u4/3). Right, the given range itself being analogue, is projected with a higher fidelity end to end?
The farthest distance (diagonal) the u4/3 can have a tone range projected across it is 2.26cm for example.
a) 2/3" = 1.10cm diagonal
b) u4/3 = 2.26cm diagonal
c) 135f = 4.33cm diagonal
d) 645f = 7.50cm diagonal
I think the DR differences between a and c or b and d are humanly detectable (well, Bifurcatorly detectable anyway! ) But I (currently) have to use my imagination to see any difference in DR between the sets: (a, b), (b, c), (c, d). All of course given the same approximate number of samples (photocites slash pixels). Additionally there's more difference between a and b that any of the others (b, c), (c, d) as at this small a size the resolving capabilities of the lens becomes much more of a factor. Given a primo lens it falls back into place however.
Disclaimer:
I know I'm not using areas as most do for such discussions but linear is easier to grasp and I think more readily translatable into useful information for mental visualizations.
Very nice pictures cavewalker, cogitech and sebboh.
@ Love the cat's eye bokeh!
@sebboh First one is really nice!
@cogitech: The Rockor is in the hands of a master! Are you processing the images with Andrea? Somehow the "signature" seems very similar. (I mean this as a compliment, I love Andrea and use it a lot!)
RZetter wrote:
Very nice pictures cavewalker, cogitech and sebboh.
@ Love the cat's eye bokeh!
@sebboh First one is really nice!
@cogitech: The Rockor is in the hands of a master! Are you processing the images with Andrea? Somehow the "signature" seems very similar. (I mean this as a compliment, I love Andrea and use it a lot!)
Thanks RZetter (what is your name?)
Wow, you have a very good eye. Yes, I do use Andrea and I love it too.
