Sony A7R 2 vs PhaseOne P45+

[shadowblade]

Hi,

1. I only want to make the worst image look like the better one... (mind you that I don't like P1 backs either - I think they have the worst color out of all MFDBs - but still much better than all DSLRs with much more processing latitude as with all backs)...

You see, I find that the MFDB gives me the worse image, with more artifacts...

[Jack Hogan]

Hi Jack, no matter what DXO findings are, I'm still among those that trust their eyes more. On the pictures Peter posted, I don't like the magenta cast of the Sony when compared to P45+, I don't like the color or the (luck of) metal structure on the (metal) cup that he included and I surely think that there is a lot of "holding" happening in HL area contrast which leads the paper ball to have a totally unreal presentation... IMO, if one aims to match the contrast of the K39000 sensor in the HLs (so that the paper will have an equally "live" presentation) he'll end up with less DR with the A7RII sensor... The same happens with my D800E when compared with my CF-39MS to my experience...

EDIT: Note that to my experience, there are two (totally different) DRs one may aim for... One (not important) is the range that a sensor may be able to record, the other (and most important) is what is left after processing as to produce a (subjective) print that has the appropriate tonal balance to make the photo-graph acceptable...

EDIT-2: Note also the color saturation difference at the red and blue ball at the lower right of the frame...

Hi Theodoros, I don't know whether this is an appropriate comment for your background but it's what I understand:

• visual information from the scene captured with good technique in the raw data represents the best IQ possible for the hardware setup; rendering and PP that come after that are arbitrary, software driven adjustments that can only interpret and present in a subjectively pleasing manner the information that is there - not make up for information that's not there.
• Better visual information captured in the raw data offers more options to the camera/artists.  More options during rendering/PP typically result in a more, not less, pleasing final photograph.  If a profile/algorithm makes an image look better from one capture, there will be an algorithm/profile that will make a different capture with better information look better still (or at least as good).
• Some indicators of the quality of visual information captured can be measured quite accurately.
• LR is not good software for this type of comparison: it is designed for ease of use, not for best IQ.  For instance it changes profiles and parameters under the hood (i.e. with all sliders at zero) on the fly based on camera settings, let alone from camera to camera, so what you see is not what you get - it's what Adobe thought was pleasing at the time it came up with them, saturation twists and all.

I've downloaded the files.  In the test at hand the a7RII+55FE at ISO 100, f/5.6, 1/8s objectively captures better visual information in the raw data than the P-45+70APO at ISO 100, f/11, 1/2s.

Cheers,
Jack

PS This shared understanding coupled with actually measured, objective IQ indicators is the reason why, I think, we no longer get apocalyptic posts from Edmund about the coming a7II green discrimination

[BernardLanguillier]

I've downloaded the files.  In the test at hand the a7RII+55FE at ISO 100, f/5.6, 1/8s objectively captures better visual information in the raw data than the P-45+70APO at ISO 100, f/11, 1/2s.

I would think that f11 explains this right? This should take the P45+ well within diffraction area I would think.

Cheers,
Bernar

[Jack Hogan]

I would think that f11 explains this right? This should take the P45+ well within diffraction area I would think.

Yes, it probably negates the P45+'s sensor size advantage.

Jack

[Theodoros]

Hi Theodoros, I don't know whether this is an appropriate comment for your background but it's what I understand:

• visual information from the scene captured with good technique in the raw data represents the best IQ possible for the hardware setup; rendering and PP that come after that are arbitrary, software driven adjustments that can only interpret and present in a subjectively pleasing manner the information that is there - not make up for information that's not there.
• Better visual information captured in the raw data offers more options to the camera/artists.  More options during rendering/PP typically result in a more, not less, pleasing final photograph.  If a profile/algorithm makes an image look better from one capture, there will be an algorithm/profile that will make a different capture with better information look better still (or at least as good).
• Some indicators of the quality of visual information captured can be measured quite accurately.
• LR is not good software for this type of comparison: it is designed for ease of use, not for best IQ.  For instance it changes profiles and parameters under the hood (i.e. with all sliders at zero) on the fly based on camera settings, let alone from camera to camera, so what you see is not what you get - it's what Adobe thought was pleasing at the time it came up with them, saturation twists and all.

I've downloaded the files.  In the test at hand the a7RII+55FE at ISO 100, f/5.6, 1/8s objectively captures better visual information in the raw data than the P-45+70APO at ISO 100, f/11, 1/2s.

Cheers,
Jack

PS This shared understanding coupled with actually measured, objective IQ indicators is the reason why, I think, we no longer get apocalyptic posts from Edmund about the coming a7II green discrimination

Hi Jack, I agree with you that capture information looks at first site to be more with the A7RII, but the same (with the HLs) looks to happen with my D800E when compared to my CF-39MS (85mm/2.8 micro PC on the Nikon set at zero movements, 120APO micro on the Contax)... My point is different though, if I develop both files using phocus (which does a much better job with the D800E than LR or CR), I find that to keep that extra HL information with the D800E, one has to end up with a dull (at the HLs) final result... OTOH, If one aims for the same punch in the HLs at both images, then (using phocus) he ends up with (slightly) more DR with the CF-39MS.

That said, the behavior of the two sensors at the LLs is also different, with the CF-39MS holding color information with much less noise (back at ISO 50, D800E at ISO 100) in the deeper shadows. OTOH, if one under exposes the CF-39MS by 2/3rds of a stop (which almost compensates for the ISO difference) he can match the information captured in the HLs, while at the same time there is almost no loss in the LLs... So, most of the times I use the 39-MS in single shot mode in other than stills or without controlled lighting, I usually underexpose it by 2/3rds of a stop...

Other than the above, because with the CF-39MS one can also use the 4x MS mode when comparing (which of course beats both results easily) he can have a reference at the same scene with the same exposure and lens and thus have a better comparison trying to bring both images as close to the MS image as possible...

Off course one has to admit that the Nikon D800E is not the same as the A7RII which I haven't tried, but the images that Peter posted, have much of the differences I've noticed when doing similar tests comparing my DSLR with my MFDB and it led to me posting my opinion on this thread... That said, I do find the Sony very attractive and seriously thinking to invest in one to combine it with a Cambo Actus... what I wait for, is an electronic mount for the Contax lenses on the Actus which will allow me to control aperture, so that I will avoid investing in more lenses (the mount is highly rumored).

[IanB]

Stepping away from the technical talk for a moment (sorry folks...):

I prefer the P45+ image to the A7rII by quite a long way. It looks almost effortlessly natural in a way which I love, whereas the A7rII image looks technically good, but pretty ordinary from an aesthetic standpoint.

That begs a question, though - how much of this is to do with the sensor/software package, and how much is it actually due to the relative qualities of the lenses?

[BernardLanguillier]

Stepping away from the technical talk for a moment (sorry folks...):

I prefer the P45+ image to the A7rII by quite a long way. It looks almost effortlessly natural in a way which I love, whereas the A7rII image looks technically good, but pretty ordinary from an aesthetic standpoint.

That begs a question, though - how much of this is to do with the sensor/software package, and how much is it actually due to the relative qualities of the lenses?

That should be easy to check since it must be possible to mount that lens on the a7rII, right?

Cheers,
Bernard

[Manoli]

That begs a question, though - how much of this is to do with the sensor/software package, and how much is it actually due to the relative qualities of the lenses?

.. and how much of it is to do with work in post ?

That should be easy to check since it must be possible to mount that lens on the a7rII, right?

Bernard, you know perfectly well that says b** all about f** all !

Best,
M

[BernardLanguillier]

Bernard, you know perfectly well that says b** all about f** all !

I must be slow today, but sorry I have really no idea what you mean here.

Cheers,
Bernard

[Theodoros]

.. and how much of it is to do with work in post ?

Hi Manoli,

You are quite right in mentioning this, IMO the best method to compare two different sensors, is to use them with the associated equipment that they will be matched with for the photographer's/potential customer needs and develop them for the best possible (hypothetical) salable outcome rather than trying to match the looks as much as possible... Then compare the final result and see which suits the tester most...

I have this (not usual) opinion about testing, simply because I believe that "better" is what satisfies the user most according to his needs/likes... Of course one may claim that if one aims for the best possible image, then the development should lead to similar looks, but if one tries the method without looking when he develops the second image of what the processing of the first image looks like, he may be surprised with how different results he may come up with when using different sensors and different associated equipment that leads the mind to adapt itself with the equipment's characteristics.

[Manoli]

Theodore,

Well, yes and no ...

If I've understood him correctly, and to paraphrase Jack Hogan's posts - what he's said is that the differences in sensors and lenses are demonstrably evident and proof lies in scientific measurements – some of which he's referred to above.  In short it can be, and is, measured. It's not based on subjective opinion which, to a degree rules profiles, and to a larger extent, post_p.

One needs to distinguish one from the other.
But, in the end result, the 'print', the two are inexorably entwined.

[peter_c]

As I suspected, a can of worms 

A few points:  The comparison included other lenses for both cameras, with a range of apertures throughout.  I elected to simplify the post with just 1 capture from each, 3 stops from wide open.  Both cameras shot tethered in C1, with focus on main eye of bank note.  Difference in focus likely my error.  The question of whether the P45+ would have superior result on Contax with native Zeiss lens- no.  In the past, I have compared that combination vs. my Rodenstock and Schneider lenses, and found the latter set of lenses to resolve better fine detail.  I hope anyone commenting on the advantages of one or another will take the time to download the RAW files first.

As a tethered studio camera,  Sony / Capture One have a few issues.  As pointed out in other threads: the HDMI output for live-view does not work when tethered via USB; the live-view will not rotate for portrait orientation in C1; the naming in C1 with camera counter does not work.  Also, you cannot use a less than HD monitor from the HDMI out, when recording 4K internal video.

Still, my conclusion is the Sony is pretty impressive.

[Torbjörn Tapani]

More depth of field in the phase shot. Just a quick n dirty conversion (had to use dng converter) in lightroom. There is not much in it. Impressive performance. But on the eye of the bank note I would have to give the nod to the Sony for sharpness. Diffraction and f-stops aside, thats mighty impressive from that 55.

[BernardLanguillier]

More depth of field in the phase shot. Just a quick n dirty conversion (had to use dng converter) in lightroom. There is not much in it. Impressive performance. But on the eye of the bank note I would have to give the nod to the Sony for sharpness. Diffraction and f-stops aside, thats mighty impressive from that 55.

You would have had very similar results with a a7r, D800E or D810, probably even closer with an Otus.

So this is basically 3 years old news.

Cheers,
Bernard

[RobertJ]

According to my RawTherapee with Deconvolution sharpening workflow (also looked at the files in C1), the Sony is miles ahead in terms of sharpness and the way it renders detail.  The P45 looks very "digital" in the way the detail is rendered.  It looks like a file from a digital camera, while the Sony looks like I'm standing there looking at the damn table.

At first, it appeared the P45 had more dynamic range after recovering the highlights of that crumpled up white paper/napkin, but I think they're pretty equal.  Is the Sony dynamic range overhyped?  IMO, it is WAY overhyped (thanks DPREVIEW).  Don't care what those clowns at DXOmark have to say.

On the bright side, Nikon and Sony cameras will continue to get better and better.  Right now, I would take a Nikon D810 instead.

[synn]

Hi Peter,

It's funny, I was doing an unrelated test last night and then realized that my test image was in some ways similar to yours. I was testing the Mamiya 150 f/3.5 (Non D) on the 645DF+ with my Credo 40 back. Here is the test image, straight out of C1P with some basic processing.

Here is the image. F/4.5, ISO 100, 1.6s. Right click-> View image to see at 100%.



It just might be me, but I feel that this is sharper than both samples that you have provided, even though this is in no way the sharpest Mamiya lens out there. Maybe diffraction is hurting your images or maybe there is some sort of shake inducing element. I am not sure.

But all that aside, lack of sharpness is not a big deal. It can be cured in post. But what caught my eye was how the gentle sub-tones come out in the P45 image and are lost in the sony.

For example, look at the bottle, right above the label. There are light green streaks running down in the Phase image that are completely lifeless in the Sony. Same goes for the shades of brown and green in the wooden board on the side. The Phase image shows way more differentiation in the colors, to my eyes.

I am quite sensitive to these sub tonal thingies, which is why I moved to MF in the first place. Other people may not see this/ care much for it, so the difference may not be so clear cut for them.