A small shadow detail comparison between my P45+ and my Sony Alpha 99

[ErikKaffehr]

Note: Reposted from another thread!
P45+ image updated and 2X crops added.

Raw images are here: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/ShadowDetail/
Hi,

I had a hunch that shadow detail in my Sony Alpha 99SLT images are cleaner than in my Phase One P45+ images. Today I shot two quite comparable high contrast images, and try to push shadow detail. Exposure was quite similar on both, according to RawDigger, with the P45+ having a bit more exposure.

Sony Alpha image has been upscaled to P45+ image size. So the Sony image is expected to have less detail.
Actual pixel detail is enclosed below, P45+ left and Sony Alpha right.

Both images processed in LR5.3 RC, using my default processing on both cameras (same settings for both AFAIK). It include convolution  sharpening, some masking and some noise reduction.

I also added a screen dump at 2:1 view (twice actual pixels) in PS.

I have tried to process the image in Capture One, too. It removes noise but for me that image seems washed out. But I am no good at C1 processing.

I will post the raw images in a couple of days.

Best regards
Erik Kaffehr

[Ken R]

Hi Erik, I played with both files extensively in both LR5 and C1Pro 7.

Which lens did you use on each?

Both files are very close in all aspects, detail, color, shadow recovery and highlight recovery. But my instincts are telling me that the lens used on the A99 image is getting the most out of the sensor (specially in the center!) and the one on the P45+ is not. Yes, the P45+ has more pixels but the per pixel sharpness is lacking a bit. There is just an overall softness, not just a difference in DOF, that is taking away from the detail a bit.

Specifically regarding Color the A99 image is a bit warmer overall specially in the gray tones but it can be adjusted to taste. Lightroom is doing a really good job on both files. But C1P7 sharpens the P45+ image with less artifacts and looks a bit cleaner and punchier. You can really see it on the edges of the rocks in high contrast areas. Shadows are a hair more clean and the file can be sharpened more with cleaner results in the shadows areas.

Also, on C1P7 with the P45+ file color is more sensitive to color settings. Small WB changes seem to make larger differences in the look of the file, specially in the shadow areas, compared to LR5. LR5 seems to produce pleasing color easily albeit a touch more bland. Of course if you get deeper into each programs controls then you can go crazy with either but the sharpening issue with LR and the P45+ file is always evident.

[Doug Peterson]

I have tried to process the image in Capture One, too. It removes noise but for me that image seems washed out. But I am no good at C1 processing.

Yep the default noise reduction in C1 will result in a mild loss of detail in such deep shadows.

Simply decrease luminance noise reduction, and increase single pixel noise reduction to taste.

[ErikKaffehr]

Hi Ken,

I would like to thank you for your thoughtful comments on the issue, but also for your postings in general.

Regarding the lenses, I used a Sony SAL 24-70/2.8ZA at 35 mm setting and at f/8 on the Sony Alpha. On the Hasselblad I used a Distagon 50/4 at f/8-11.

This shoot was not intended as a test, but I normally carry both equipments, so I just shot a couple of images on the Alpha 99 in addition to the "blad". I also found the images on the soft side, but this time I was really looking at shadow noise. The lack of sharpness may have been a focusing issue, I normally use 3X monocular for focusing but I couldn't find it in the bag, and I also had my eyes tearing.

I would be thankful if you posted a pair of crops from C1 as my skills with C1 are non existent. It is a good program but we sort of make no friends. I also guess that I am accustomed to LR/ACR rendition and object to something different.

Regarding the Zeiss lenses I started with a Sonnar 150/4, and I think that is my best lens, but the Distagon 50/4 I used here is also pretty good. First I missed the "byte" I expected from MF, but I use the same sharpening as on the Alpha 99 and I am quite happy.

In general I feel that the Hasselblad/P45+ has the "MP advantage". I have run some tests on the Sonnar 150/4 and I would say it performs like my Minolta/Sony lenses when used with an adapter on Alpha 99.

I enclose a screendump of my sharpening settings, a bit on the extreme side, but it works for me. It will generate some ringing artefacts, tough.

On color rendition, I generated a DNG-profile for my camera where I reduced saturation in yellow/reds quite a bit, and I am quite happy with it.

The main issues I may have with C1 regarding color is probably that it applies a "film curve" as default which is just ugly, the other is that it seems to interpret auto WB in an odd way. I feel LR makes a better job on WB. My greatest problem with C1 is that I like LR5. LR5 has sort of tone mapping built in that can be used with gradients and correction brush. That is my most used tool, really!

Best regards
Erik

Hi Erik, I played with both files extensively in both LR5 and C1Pro 7.

Which lens did you use on each?

Both files are very close in all aspects, detail, color, shadow recovery and highlight recovery. But my instincts are telling me that the lens used on the A99 image is getting the most out of the sensor (specially in the center!) and the one on the P45+ is not. Yes, the P45+ has more pixels but the per pixel sharpness is lacking a bit. There is just an overall softness, not just a difference in DOF, that is taking away from the detail a bit.

Specifically regarding Color the A99 image is a bit warmer overall specially in the gray tones but it can be adjusted to taste. Lightroom is doing a really good job on both files. But C1P7 sharpens the P45+ image with less artifacts and looks a bit cleaner and punchier. You can really see it on the edges of the rocks in high contrast areas. Shadows are a hair more clean and the file can be sharpened more with cleaner results in the shadows areas.

Also, on C1P7 with the P45+ file color is more sensitive to color settings. Small WB changes seem to make larger differences in the look of the file, specially in the shadow areas, compared to LR5. LR5 seems to produce pleasing color easily albeit a touch more bland. Of course if you get deeper into each programs controls then you can go crazy with either but the sharpening issue with LR and the P45+ file is always evident.

[ErikKaffehr]

Hi Doug,

Thanks for suggestions! I will try. Still, I feel that even if C1 is a fine tool for sure, we make no friends, C1 and me.

Best regards
Erik

Yep the default noise reduction in C1 will result in a mild loss of detail in such deep shadows.

Simply decrease luminance noise reduction, and increase single pixel noise reduction to taste.

[BernardLanguillier]

Regarding the Zeiss lenses I started with a Sonnar 150/4, and I think that is my best lens, but the Distagon 50/4 I used here is also pretty good. First I missed the "byte" I expected from MF, but I use the same sharpening as on the Alpha 99 and I am quite happy.

In general I feel that the Hasselblad/P45+ has the "MP advantage". I have run some tests on the Sonnar 150/4 and I would say it performs like my Minolta/Sony lenses when used with an adapter on Alpha 99.

Is that a lens issue or a focusing issue?

Very small errors in focusing can easily reduce the resolving power from 40 to 20 megapixels.

Cheers,
Bernard

[ErikKaffehr]

Hi,

Yes could be focusing, especially in this case. But this case was intended to be a noise test and not a sharpness test. In "lab conditions" I can achieve better sharpness (in lp/mm) on the Sonnar 150/4 than on my zooms on the Sony's, real world conditions may be different. I use a PM5 viewfinder and a Zeiss Victory 3X monocular, which gives me 9X magnification. These time I head tears in my eyes and did not have my monocular with me.

In general I feel that the Hasselblad has the 39MP vs. 24 MP advantage, but the lenses have a lot of field curvature, except the Sonnar 150/4, that lens is almost perfect at f/4. The Distagoon 50/4 used here is also quite neat. Macro Planar has bad curvature of field at infinity, probably fine close up.

Distagon 40/4 FLE is interesting. In test, the corners are awful, in reality a large part of the image is awful, but sometimes field curves in your direction and you get excellent sharpness on foreground grass. The Distagon 40/4 has also a lot of lateral chromatic aberration that doesn't correct fully in Lightroom. When the lens is shifted it is bad. An oddity, the off axis region sharpens really well in Lightroom, I just use a radial filter and crank up sharpening on the outer region, and voilá the Distagon 40/4 FLE is sharp!

Zeiss made a much better Distagon 40/4 FLE IF, that is the lens group Hartblei uses for their 40 mm TS.

Here are some recent samples: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples3/

I can add that I have a lot of aliasing and Moiré which indicates that the lenses have some decent MTF at pixel pitch.


Best regards
Erik

Is that a lens issue or a focusing issue?

Very small errors in focusing can easily reduce the resolving power from 40 to 20 megapixels.

Cheers,
Bernard

[ErikKaffehr]

Hi,

I added two crops from some shoots this year. Both are at actual pixels (sort of).

The tree trunk image is actual pixels on unscaled images. In this case I had better framing on the DSLR, using my 70-400/4.-5.6G zoom on the Alpa. I could simply look at the image at same magnification and all advantage from MF comes from lens and no OLP filtering. I would say it is obvious the P45+ image is sharper, and it is also sharp. The Sonnar 150/4 was used in this case on the Hasselblad. I can add, that I in the end preferred the Hasselbald crop.

The ruin image was a good crop on both Hasselblad and Sony, so here I upsized the Sony image to the same width as the Hasselblad. In the actual pixels view we can see that the Hasselblad wins and the Sony has some upressing artefacts. In this case I would also say that the Hasselblad image has some hazy look to it. I would say it is a lens issue. Sony: Alpha 99SLT + 24-70/2.8 at 35/8 LVMF (11X) . Hasselblad: Distagon 50/4 at unknown aperture (guess f/11), PM3 + 3X mag (9X mag in total).

Raw images here: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/MFDB_VS_DSLR2/

Best regards
Erik

Hi Erik, I played with both files extensively in both LR5 and C1Pro 7.

Which lens did you use on each?

Both files are very close in all aspects, detail, color, shadow recovery and highlight recovery. But my instincts are telling me that the lens used on the A99 image is getting the most out of the sensor (specially in the center!) and the one on the P45+ is not. Yes, the P45+ has more pixels but the per pixel sharpness is lacking a bit. There is just an overall softness, not just a difference in DOF, that is taking away from the detail a bit.

Specifically regarding Color the A99 image is a bit warmer overall specially in the gray tones but it can be adjusted to taste. Lightroom is doing a really good job on both files. But C1P7 sharpens the P45+ image with less artifacts and looks a bit cleaner and punchier. You can really see it on the edges of the rocks in high contrast areas. Shadows are a hair more clean and the file can be sharpened more with cleaner results in the shadows areas.

Also, on C1P7 with the P45+ file color is more sensitive to color settings. Small WB changes seem to make larger differences in the look of the file, specially in the shadow areas, compared to LR5. LR5 seems to produce pleasing color easily albeit a touch more bland. Of course if you get deeper into each programs controls then you can go crazy with either but the sharpening issue with LR and the P45+ file is always evident.

[ondebanks]

but the lenses have a lot of field curvature, except the Sonnar 150/4, that lens is almost perfect at f/4.

I think that may be a characteristic of Sonnars in general. At least, that's been my experience too with the CZJ MC Sonnar 300/4. However, the CZJ MC Sonnar 180/2.8 was not as good off-centre wide open, but few people ever notice this as they primarily shoot it for the famous portrait bokeh.

Macro Planar has bad curvature of field at infinity, probably fine close up.

Interesting to know. I wonder if the Mamiya/Phase 120/4 behaves similarly at infinity. It's a lens I've often considered buying, but I would hope that it would be at least as good at infinity as my 110/2.8 N when set to f/4.

Distagon 40/4 FLE is interesting. In test, the corners are awful, in reality a large part of the image is awful, but sometimes field curves in your direction and you get excellent sharpness on foreground grass.

That is exactly like the Mamiya 35/3.5 C and N (and probably the AF too with the same old optical formula).

Ray

[ErikKaffehr]

Hi Ray,

The Planar is a macro lens, so it is optimised for close distances. The MTF data from Zeiss tells the story. Ugly at infinity quite nice close up. MTF curves below.

Most modern lenses have floating elements, that is a group that moves with focusing to keep the field flat. Check out MTF curves, they tell the story!

Best regards
Erik

Interesting to know. I wonder if the Mamiya/Phase 120/4 behaves similarly at infinity. It's a lens I've often considered buying, but I would hope that it would be at least as good at infinity as my 110/2.8 N when set to f/4.

Ray