Sony A7R 2 vs PhaseOne P45+

[peter_c]

In response to this LL post, I'm posting a comparison I did a couple of weeks ago for my own benefit.  My current working kit of several years is a Linhof 679CC with P45+ and various Rodenstock & Schneider digital lenses, mounted on a Rollei electric shutter system.  A perfect system for my food work and I see no reason to upgrade.  The P45+ also lives on a Contax on occasion.  My personal system has progressed: GH2, OM-D 5, A72, A7R2.  What you see is a quick, spontaneous studio shot of stuff I had on hand:  P45+ and Rodenstock 70mm Apo Digital (not HR) vs Sony A7R2 with 55mm FE.  I'm not going to comment, the raw files can are available here..  BTW I did recently sell my Canon 5D2.

[ACH DIGITAL]

Thanks Peter for the test.
My conversion on ACR, same values for everything except, white balance and exposure. I think they are fair. In some part of the images you see a bit of difference in focus distance and depth of field.

[Manoli]

Thanks for that comparison, Peter.

[Theodoros]

Thanks for posting this Peter... Having a Hasselblad CF-39MS myself which I often also use on Contax, I find it very useful since P-45+ shares the same sensor... One question please... would the result be better for the P-45+ if it was used with the Contax and a dedicated lens, or would it be worst? (I'm not familiar with the 70mm Rodie lens).

Personally, I like the P-45+ image better, I have excluded resolution from my judgement (because there's more than enough with both) and concentrated on color accuracy, color saturation and mid tone contrast... I also believe that if one develops both images to have a similar look as if it was the same contrast curve, the P-45+ image would come ahead in DR too... Lets see what others think...

[Jack Hogan]

Personally, I like the P-45+ image better, I have excluded resolution from my judgement (because there's more than enough with both) and concentrated on color accuracy, color saturation and mid tone contrast... I also believe that if one develops both images to have a similar look as if it was the same contrast curve, the P-45+ image would come ahead in DR too... Lets see what others think...

Hi Theodoros,

Tongue in cheek, right?

A fit to 2007 DxoMark.com data shows P-45+ read noise at base ISO of about 13.5e- and clipping about 41500e- (the fits are not perfect, blame it on early days at DXO).  That means that the camera requires about 11.6 bits to fully encode information rolling off the sensor into the raw data.  12-bits would have been plenty, 14-bits way overkill, 16-bits for marketing purposes only - so we can cancel any potential P45+ advantages in bit depth.  Compare that to the a7RII read noise of 7.5e- and clipping at 52842e- and you realize that the a7RII comes out on top in terms of DR, SNR and tone gradations throughout the range at the sensor level.

Looking at color next, the Sensitivity Metamerism Index is 72 for the P-45+ and 85 for the a7RII, close to worst in class and best in class respectively, no inherent P45+ sensor color advantage here.  In fact just the opposite again.   Playing with contrast curves/profiles would only emphasize the IQ difference.

ANY potential advantages the P-45+ might have in contrast, local contrast, micro contrast would therefore become apparent solely at the picture level because of its larger sensor size and presumed quality of the lens used: we would expect to see better sharpness from it for instance.  But we see just the opposite, meaning that the newer Zony 55FE is a better (or better focused) lens than the Rodenstock 70mm in this composition.  Have I forgotten anything relevant?  If not, I am afraid that any and all advantages that one sees on the 8 year old P-45+ image vs the state-of-the-art a7RII's can be ascribed to ye olde placebo effect.



Jack

PS. Summary at http://www.dxomark.com/Cameras/Compare/Side-by-side/Phase-One-P45-Plus-versus-Sony-A7R-II___456_1035.

[Bart_van_der_Wolf]

If not, I am afraid that any and all advantages that one sees on the 8 year old P-45+ image vs the state-of-the-art a7RII's can be ascribed to ye olde placebo effect.

Maybe Jack, but I'm afraid it more like Cognitive Dissonance 

But you are right, the specs are in favor of the A7R II, so it would be hard for the P45+ to pull ahead based on sensor size alone. Of course, in a studio setting with controlled lighting and with custom profiles the differences will become smaller. Both are potentially capable of producing good to exellent results.

I'm not sure, but the tethering abilities of the A7R II seem to have received little attention so far, but well supported tethering can also make a difference in actual (studio) use.

Cheers,
Bart

[Jack Hogan]

But you are right, the specs are in favor of the A7R II, so it would be hard for the P45+ to pull ahead based on sensor size alone. Of course, in a studio setting with controlled lighting and with custom profiles the differences will become smaller. Both are potentially capable of producing good to excellent results.

I'm not sure, but the tethering abilities of the A7R II seem to have received little attention so far, but well supported tethering can also make a difference in actual (studio) use.

All agreed, although I would toss the reference to custom profiles as immaterial presuming both camera/lens combinations get their own: a 'perfect' custom profile can only present at best what's there, it cannot make up for what isn't there in the first place.

Jack

[Theodoros]

All agreed, although I would toss the reference to custom profiles as immaterial presuming both camera/lens combinations get their own: a 'perfect' custom profile can only present at best what's there, it cannot make up for what isn't there in the first place.

Jack

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...

[shadowblade]

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...

Conversely, if you adjust the P45 image to match the A7rII instead of the other way round, you're left with even worse shadow detail than you started off with. The magenta cast is just a matter of white balance, or a difference between the two lenses, not a difference in sensor colour accuracy (which is always relative to a white point).

When I shoot an image, I want the camera to produce as technically good an image as possible - not to add any of its own character, not to achieve a certain 'look', but merely to capture the photons as accurately as possible. Any image, no matter the camera, is going to end up tortured under a multitude of curves and layers in Photoshop anyway - the better the original is technically, the better the final result.

I really wish the A7rII had a base ISO of 25, though, for minimal low ISO noise (and a base ISO almost two stops higher)... shooting four frames and averaging them isn't quite the same.

[eronald]

Camera shake/ Shutter vibration in the case of the Sony?

Edmund

Maybe Jack, but I'm afraid it more like Cognitive Dissonance 

But you are right, the specs are in favor of the A7R II, so it would be hard for the P45+ to pull ahead based on sensor size alone. Of course, in a studio setting with controlled lighting and with custom profiles the differences will become smaller. Both are potentially capable of producing good to exellent results.

I'm not sure, but the tethering abilities of the A7R II seem to have received little attention so far, but well supported tethering can also make a difference in actual (studio) use.

Cheers,
Bart

[Bart_van_der_Wolf]

Camera shake/ Shutter vibration in the case of the Sony?

Hi Edmund,

I haven't seen any comparison between the two. All cameras/lenses with moving parts will produce some movement/vibration. The longer focal length required for a physically larger sensor will magnify subject movement on the sensor, and the physically smaller sensor needs more magnification for the same size output. Sensor stabilization against camera shake may help with handholding, but it is also needed more on the smaller sensor.

Cheers,
Bart

[Bart_van_der_Wolf]

All agreed, although I would toss the reference to custom profiles as immaterial presuming both camera/lens combinations get their own: a 'perfect' custom profile can only present at best what's there, it cannot make up for what isn't there in the first place.

The differences between CFAs/sensors may be smaller than the difference due to the profiles.
So it's hard to say without actually testing it.

Cheers,
Bart

[AlterEgo]

I really wish the A7rII had a base ISO of 25, though, for minimal low ISO noise (and a base ISO almost two stops higher)... shooting four frames and averaging them isn't quite the same.

base ISO25 is not a guarantee of low "readout" noise in deep shadows - it just typically a sign that sensel well capacity is big (big sensor big sensels) or pushed to the limits (when manufacturer tries use every single electron even all sensels have physically different wells) or that sensor schematics allows in process readout from sensels effectively making more than 1 exposure behind the scenes... but DR (which is defined by both sensel well saturation capacity and readout noise) can still be technically worse (even S/N above deep shadows better)... and averaging will increase S/N (and/or resolution based on setup and software - see photoacute.com for example) like it or not...

[Theodoros]

Conversely, if you adjust the P45 image to match the A7rII instead of the other way round, you're left with even worse shadow detail than you started off with. The magenta cast is just a matter of white balance, or a difference between the two lenses, not a difference in sensor colour accuracy (which is always relative to a white point).

When I shoot an image, I want the camera to produce as technically good an image as possible - not to add any of its own character, not to achieve a certain 'look', but merely to capture the photons as accurately as possible. Any image, no matter the camera, is going to end up tortured under a multitude of curves and layers in Photoshop anyway - the better the original is technically, the better the final result.

I really wish the A7rII had a base ISO of 25, though, for minimal low ISO noise (and a base ISO almost two stops higher)... shooting four frames and averaging them isn't quite the same.

1. No sensor is accurate to the software's white point... it's all an interpolation tuning matter (of the processor) in combination with sensor ability...
2. That's what YOU do... Obviously you don't use MF...
3. If the A7rII had 25 ISO as sensor base, then it would be 2 stops worst with higher ISO... Obviously that wasn't Sony's aim...

[shadowblade]

1. No sensor is accurate to the software's white point... it's all an interpolation tuning matter (of the processor) in combination with sensor ability...

Exactly.

Try to make the A7rII file look like the P45+ file and you lose image quality. Try to make the P45+ file look like the A7rII file and you also lose image quality.

2. That's what YOU do... Obviously you don't use MF...

I shot MF and LF film for a long time, and have also shot landscapes using the IQ180. These days, though, between stitching software, better lens selection and a less cumbersome system, improved full-frame sensors make MF barely worth in in most situations.

3. If the A7rII had 25 ISO as sensor base, then it would be 2 stops worst with higher ISO... Obviously that wasn't Sony's aim...

No it wouldn't. Maximum usable ISO has nothing to do with the minimum. Maximum ISO is all to do with signal-to-noise ratio. Minimum ISO is to do with well depth. You can increase one without changing the other.

[Theodoros]

Exactly.

Try to make the A7rII file look like the P45+ file and you lose image quality. Try to make the P45+ file look like the A7rII file and you also lose image quality.

I shot MF and LF film for a long time, and have also shot landscapes using the IQ180. These days, though, between stitching software, better lens selection and a less cumbersome system, improved full-frame sensors make MF barely worth in in most situations.

No it wouldn't. Maximum usable ISO has nothing to do with the minimum. Maximum ISO is all to do with signal-to-noise ratio. Minimum ISO is to do with well depth. You can increase one without changing the other.

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)...
2. There is not much difference in IQ between the IQ 180 and the P45+ (other than res - the least important of IQ factors)... but never the less having used an IQ-180 is different than living with one....
3. Yes it would... It would require significantly more amplifying form the sensor's output...

[Torbjörn Tapani]

Iso doesn't add noise. It is lack of signal that produce more noise in an image. High iso would be the same if you just had a larger well (iso 25, through some magic).

[Theodoros]

Iso doesn't add noise. It is lack of signal that produce more noise in an image. High iso would be the same if you just had a larger well (iso 25, through some magic).

But you don't.... so the avalliable quantity of photons in what you have, has to be amplified...

EDIT: In reality, the problem is not with noise, but rather with S/N ratio... So if we assume that base ISO would have the same S/N ratio, lower base ISO produces more noise in proportion the more the signal is amplified...

[shadowblade]

But you don't.... so the avalliable quantity of photons in what you have, has to be amplified...

EDIT: In reality, the problem is not with noise, but rather with S/N ratio... So if we assume that base ISO would have the same S/N ratio, lower base ISO produces more noise in proportion the more the signal is amplified...

No.

In order to get a lower ISO, you need a greater well depth. Otherwise, the wells will just fill up too quickly and you won't get your lower ISO.

This greater well depth gives you greater SNR, but only at base ISO. A sensor with a base ISO of 25 will have 2 stops greater DR than a sensor with base ISO 100, electronic readout and other architecture being the same. But only at ISO 25 - if you boosted it to ISO 100, it would have the same DR at ISO 100 as the other sensor. The same applies if you shoot it at ISO 25600 - it would have the the same DR as the other sensor with the higher base ISO.

It would be very easy to design a sensor with a base ISO of 25600 - you wouldn't need any well depth. It also wouldn't perform any better at high ISO than any other similarly-designed sensor with a lower base ISO.

[shadowblade]

But you don't.... so the avalliable quantity of photons in what you have, has to be amplified...

EDIT: In reality, the problem is not with noise, but rather with S/N ratio... So if we assume that base ISO would have the same S/N ratio, lower base ISO produces more noise in proportion the more the signal is amplified...

You don't have the same number of photons.

If you shoot at ISO 25, you're collecting four times as many photons as when shooting at ISO 100. But you need a sensor with a deep well capacity, capable of a base ISO of 25, to take advantage of this (a sensor with poor quantum efficiency will also require all these extra photons, but only because they require many more incoming photons to end up with the same number of photoelectrons - a more efficient sensor can do the same at a higher ISO).

MF CCD sensors don't have deep well capacities - they just have poor quantum efficiency, which is why they need to collect so many more photons to achieve the same SNR. That's why they're terrible at ISO 100 and above.

Also, shooting at high ISO isn't multiplying from base - it's collecting a smaller number of photons and multiplying from that. At ISO 25600, two sensors of the same size and quantum efficiency will collect the same number of photons, regardless if one has abase ISO of 200 and the other has a base ISO of 25. Therefore, the SNR will be the same.