Interesting (to me at least) comparison of MFD and 35mm

[David Watson]

I was asked by another forum member what I thought of the H3D-22 compared to the Leica M9 having owned both.  My reply was to send some test shots taken by both cameras of my barn door.  I have also used this same subject for tests of a Nikon D3X and now my H4D60.   It is interesting just how good all 4 cameras are and how well both the D3X and particularly the M9 stands up to the test but you can make your own minds up with these full frame jpegs.

[David Watson]

Details of the preceding images

[ced]

David I find the 35's show their limits very quickly in these results.
Thanks for posting.

[henrikfoto]

I think the most suprising is to see how little the differense is between Hasselblad 60mp and Hasselblad 22.

Maybe Hasselblad are doing a better job with Kodak?

[ondebanks]

Thanks for posting the comparison.

From worst to best in IQ, I rank them: D3X, M9, H3D-22 [not much ahead of M9], H60 [well ahead of the pack].

The framing/magnification slightly favours the two H cameras in this test, but probably not so much as to alter the conclusions.
I guess the 3 main variables here are: pixel count, sensor size, and AA filter or not. (Assuming that all the lenses are excellent at the aperture used).

So in technology terms, the H4D60 wins over the D3x on all 3 variables: 2.5 more pixels, 2.5 times the sensor size, and no AA filter. No surprise then that they fall on either extreme of the image ranking.

The most interesting thing is that the M9 is competitive with the H3D-22. They are pretty much tied on 2 variables (pixel count, no AA filter) but are 2x different in sensor size. So perhaps sensor size is not that important, in good light and with lenses that outresolve at least the bigger pixels.

Can't comment on colour, as the lighting/WB was changing during the test.

Ray

[ondebanks]

I think the most suprising is to see how little the differense is between Hasselblad 60mp and Hasselblad 22.

Maybe Hasselblad are doing a better job with Kodak?

I see a vast difference in resolution, in favour of the 60. But the shadow tones are definitely better on the 22 (Kodak) than the 60 (Dalsa), where they block up and become muddy. So the Kodak has better signal to noise.

Ray

[EricWHiss]

No offense but I think its difficult to conclude much from this set of test images.  Tests like this can be misleading if the focus has not been very carefully set because even when stopped down the apex of the focus can be seen when zoomed in to 100%.    Furthermore, one of the bigger differences between 35mm format and MF is the more rapid transition from in focus to out of focus.  You can't see that when shooting a flat surface, and if you just compare the one facet of images, sharpness, you can't know that the point you are comparing is at the apex of focus in all the shots.

[David Watson]

No offense but I think its difficult to conclude much from this set of test images.  Tests like this can be misleading if the focus has not been very carefully set because even when stopped down the apex of the focus can be seen when zoomed in to 100%.    Furthermore, one of the bigger differences between 35mm format and MF is the more rapid transition from in focus to out of focus.  You can't see that when shooting a flat surface, and if you just compare the one facet of images, sharpness, you can't know that the point you are comparing is at the apex of focus in all the shots.

Don't disagree entirely but as Keith says it is interesting to see how good the 22 and M9 are compared to both the D3X and 60. 

[hasselbladfan]

I have to admit, the M9 is really good. Wish the H4D60 was a bit cheaper.

[bjanes]

No offense but I think its difficult to conclude much from this set of test images.  Tests like this can be misleading if the focus has not been very carefully set because even when stopped down the apex of the focus can be seen when zoomed in to 100%.    Furthermore, one of the bigger differences between 35mm format and MF is the more rapid transition from in focus to out of focus.  You can't see that when shooting a flat surface, and if you just compare the one facet of images, sharpness, you can't know that the point you are comparing is at the apex of focus in all the shots.

I agree with the above, but one can check shadow noise by looking at the darkest neutral patch of the ColorChecker. Since some of the images are not white balanced, I split out the RGB channels in Photoshop for the D3x, M9, and HasH4D60 images and determined the means and standard deviations with the Photoshop histogram tool.

D3x
red channel: mean = 49.34 SD = 0.89
green channel: mean = 48.55, SD = 1.92
blue channel: mean = 47.96, SD = 1.27

M9
red channel: mean = 48.61, SD = 2.24
green channel: mean = 43.91, SD = 2.74
blue channel: 40.34, Sd = 3.16

Hasselblad H4D60
red channel: mean = 44.22, SD = 0.78
green channel: mean = 44.38, SD = 0.87
blue channel: mean = 44.32, SD = 0.73

The noise is considerably greater with the M9 than with the D3x. Owing to its large total sensor area, the Hasselblad has the least noise. These values are likely within what would be considered photographic DR.

Regards,

Bill

[ErikKaffehr]

Hi,

I'd say that it's hard to compare the images as they are in different light. I would place them

1) Hasselblad D60
2) Hasselblad D22
3) Nikon
4) Leica

But that may depend on light, processing and many other factors. Initally I preferred Nikon to Hassy, it has a more natural rendition in my view but tha Hassy has more detail.

I enclose a screen dump of what I see, the samples are arranged:

HassyD60     Nikon
Leica           Hassy D22

Best regards
Erik

[BernardLanguillier]

Thks for these test results. For once DSLRs were represented by their best batter instead of picking the older 5DII. :-)

The results are pretty close indeed. The hd60 is of course ahead of the pack in terms of resolution in the sharp areas, but I am not sure the focus could be perfectly adjusted to the target area.

It would also be interesting to see how sharpening could help the different images.

Anyway, I would be seriously worried about upcoming 36mp DSLRs if I were working at Hassy, especially if they do indeed have an AA filterless version.

For us photographers this is excellent news.

Cheers,
Bernard

[ErikKaffehr]

Hi,

Regarding AA-filtering, both the 22 MP Hasselblad and the Leica show some aliasing, in my view. The Leica also has tendency to color Moiré.

Regarding 36 MP full frame, I don't think it exactly corresponds to 16 MP on APS-C, which seems to be what you are using right now. I don't expect a major boost in quality going from 24MP FX to 36MP FX.

Best regards
Erik

Thks for these test results. For once DSLRs were represented by their best batter instead of picking the older 5DII. :-)

The results are pretty close indeed. The hd60 is of course ahead of the pack in terms of resolution in the sharp areas, but I am not sure the focus could be perfectly adjusted to the target area.

It would also be interesting to see how sharpening could help the different images.

Anyway, I would be seriously worried about upcoming 36mp DSLRs if I were working at Hassy, especially if they do indeed have an AA filterless version.

For us photographers this is excellent news.

Cheers,
Bernard

[BernardLanguillier]

Regarding AA-filtering, both the 22 MP Hasselblad and the Leica show some aliasing, in my view. The Leica also has tendency to color Moiré.

Regarding 36 MP full frame, I don't think it exactly corresponds to 16 MP on APS-C, which seems to be what you are using right now. I don't expect a major boost in quality going from 24MP FX to 36MP FX.

Sure, moire would be a real issue. As far as the gap goes, the increase of detail btwn an AA filtered 24mp camera and an AA filterless 36mp camera should be pretty significant I would think.

Cheers,
Bernard

[ErikKaffehr]

Hi Bernard,

Your D7000 corresponds exactly to a full frame 36MP cropped to APS-C.

Best regards
Erik

Sure, moire would be a real issue. As far as the gap goes, the increase of detail btwn an AA filtered 24mp camera and an AA filterless 36mp camera should be pretty significant I would think.

Cheers,
Bernard

[NikoJorj]

Regarding AA-filtering, both the 22 MP Hasselblad and the Leica show some aliasing, in my view. The Leica also has tendency to color Moiré.

Yes, what has struck me most in that test is the color Moiré of the H22. I see some, but not much, in the H60 and M9 shots in comparison.
That does make a difference, especially with the D3x.

[ErikKaffehr]

Hi,

Big pixels + sharp lenses -> Moiré

But, Moiré is not the only issue, we also get a lot of fake detail. Small pixels and correct OLP-filtering is probably better. Unless we like artifacts, of course!

Best regards
Erik

Yes, what has struck me most in that test is the color Moiré of the H22. I see some, but not much, in the H60 and M9 shots in comparison.
That does make a difference, especially with the D3x.