The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

DxO P45+ test: Let the Games Begin

yaya

Active member
"http://www.dxomark.com/index.php/eng/Technologies/Medium-format-ranking"

Professional portrait and landscape photographers often use medium-format cameras because of their superb performance under controlled lighting conditions. However, as these cameras are definitely not designed for so-called “action photography” scenarios, they generally do not perform well with respect to DxO Labs’ Low-Light ISO metric. Because of this inherent low-light limitation, medium-format cameras do not receive top marks on the overall DxOMark Sensor scale, even though they may show outstanding performance with respect to Color Depth or Dynamic Range.
The DxOMark Sensor scale is designed to weight equally three photographic scenarios that, taken together, cover nearly the entire photospace: Portrait, Landscape, and Action Photography. Each scenario represents use cases that stress a specific parameter of the camera—Color Depth, Dynamic Range, and Low-Light ISO, respectively. When looking at cameras with narrow or specialized uses, considering the specific ranking for the right metric or metrics is critically important.
Medium-format cameras are designed to perform best in particular use cases—specifically, they are mostly used in a studio environment where light level is not a problem, and in landscape photography, where they are most often used with tripods to facilitate long exposure times. In light of their specific uses, medium-format cameras are optimized for low ISO performance, and so do not feature a wide “analog” ISO latitude, meaning that they show some limitations at high ISO speed. Consequently, medium-format cameras end up with lower Low-Light ISO rankings compared to DSLRs, and this affects their overall DxOMark Sensor score.


I think there are some inherent conflicts between the purpose of these tests and the way they are being performed and analysed.

Also, between Portrait, action and landscape, they seem to have dropped the still life/ product area, which takes about 70% of all commercial imagery produced today...

I also think that what most people relate to as "better image quality" in MF is coming from greater colour sensitivity and tonal range, which are both actually demonstrated in these tests.

This is not to diss DxO by no means, but perhaps their tests should be re-designed to accommodate new types of cameras and more realistic measures?
 

Stuart Richardson

Active member
Yeah Guy, I am with you. I like to follow testing and even argue about it at times, but the real proof is in what the photos look like -- that is hugely dependent on lens and format, far more than just the sensor. This is something I think we have forgotten with digital. It's like everyone is only thinking of the film, and they forgot that the size of the format and the lens in front of it is a larger determinant on the resultant image. TechPan in 35mm and FP4 in 6x6 might have a similar level of grain and resolution, but the two photos will look very different.

Medium format has a special look not only because of the high resolution and tonality, but also because of the fact that an 80mm is a standard lens, while it is a telephoto on 35mm. Same goes with large format. 4x5 looks different in many ways not just because of the fact that it has intense resolution, but because you generally have complete freedom of movement and a 150-210mm lens as a standard.

These differences in format are not trivial unless you shoot every photo at f/22...
 

mark1958

Member
I think the article is interesting and was written to provoke discussion-- which is fine. I mean M.R. even states that --"are we going to find that there's more to how a sensor performs that can be can be told by instrument measurements alone? I believe so."
There is clearly an artistic issue that cannot be measured as there is a subjective component to most things in life. Similar to Medicine-- geez even in hardcore science there is always debate about which parameters are most important to measure to come up with a conclusion and depending on how one does the measurements--do the study differently and can come up with different conclusions. THere is not always one right and wrong way either. That is why there is little full agreement on many things in the world. Anyway, I think his article is provocative and will stimulate lots of discussion --...
 

glenerrolrd

Workshop Member
Why is this so surprising? It seems that the DxO tests only represent a single dimension and doesn t factor in size of the sensor in any meaningful way. Isn t it closer to comparing films .You could use the exact same film in both 35mm and 2 1/4.....and immediately see the difference . They really doesn t seems to be any way to compare different sensor sizes..but within a group it looks pretty accurate. Reminds me of the old lens tests ...didn t the original zeiss lens on the Hasselblad look weak compared to the Leica glass?
 

Paratom

Well-known member
Until I find a t/s option for 35mm that will compare to my Schnider and REodenstock glass, it's a mute point for me. I shoot architecture for a living, and I've tried the offerings from Canon. Believe you me, there is no comparison.
I think it would be a little unfait to compare a Schneider or Rodenstock - since those are not lenses for MF-Reflex-Cameras.

I agree that a 35mm-DSLR-lens has to be better than a MF lens because of the smaller sensor/higher magnification.

I wrote this because I am not totally sure if the "problem" of DSLRs are the lenses or if the problem might be the AA-filters and 12-bit color depth.

Looking at DMR images I am again and again impressed and without direct comparison just from my impression I dont find the DMR images to stay behind MF images - of course this is limited to a certain print size due to the resolution limits.
But then the DMR doesnt offer any real benefits over a mf-system (like higher ISO or fast AF)
 

Geoff

Well-known member
"http://www.dxomark.com/index.php/eng/Technologies/Medium-format-ranking"

Professional portrait and landscape photographers often use medium-format cameras because of their superb performance under controlled lighting conditions. However, as these cameras are definitely not designed for so-called “action photography” scenarios, they generally do not perform well with respect to DxO Labs’ Low-Light ISO metric. Because of this inherent low-light limitation, medium-format cameras do not receive top marks on the overall DxOMark Sensor scale, even though they may show outstanding performance with respect to Color Depth or Dynamic Range.
The DxOMark Sensor scale is designed to weight equally three photographic scenarios that, taken together, cover nearly the entire photospace: Portrait, Landscape, and Action Photography. Each scenario represents use cases that stress a specific parameter of the camera—Color Depth, Dynamic Range, and Low-Light ISO, respectively. When looking at cameras with narrow or specialized uses, considering the specific ranking for the right metric or metrics is critically important.
Medium-format cameras are designed to perform best in particular use cases—specifically, they are mostly used in a studio environment where light level is not a problem, and in landscape photography, where they are most often used with tripods to facilitate long exposure times. In light of their specific uses, medium-format cameras are optimized for low ISO performance, and so do not feature a wide “analog” ISO latitude, meaning that they show some limitations at high ISO speed. Consequently, medium-format cameras end up with lower Low-Light ISO rankings compared to DSLRs, and this affects their overall DxOMark Sensor score.
I remember now - some thirty years ago the magazines concluded that a normal car, used in a normal way, would give better results (value per $) than a Ferrari.

Glad to see little has changed: if you take a niche product and average its results over a larger set of criteria, surprise: it won't do so well in the test.

This isn't to take anything away from the wonderful quality of the dslr products now available, but rather to wonder why DxO didn't "de-couple" their test a bit more?
 
Note that even two medium format systems based on the sensor are producing meaningfully different results: the H3DII-39 and Phase One P45+ both use the 6.8 micron 1.1X Kodak 39 megapixel sensor and the Phase beats the Hassy by half a stop of dynamic range and 17% better "low light ISO".

Doug Peterson, Head of Technical Services
Capture Integration, Phase One & Canon Dealer | Personal Portfolio
Doug, you are toooo predictable. ;-)

Anyone know if they analyze the RAW files? If so how? As they are missing out on each companies processing.

If they are analyzing a TIF, what settings do they use?

...and so on.

Best,



David
 

PeterA

Well-known member
I cant connect a Nikon to an Alpa..is that in their 'test'?

Chill David - no need to be defensive, Hasselblad owners know the difference between this that and the other thang -:)
 

tashley

Subscriber Member
I wonder if MF glass is really better than a Leica 19mm or Leica50mm or Zeiss Macro100 or Nikon 24PCE or Nikon 200/2.0 or Canon 35/1.4 or Canon 135/2.0 or . . . . ?
I agree with you. My experience so far, with four lenses for my Phamiya, two of which are 'D' lenses, is that they are not in the same league as Leica glass. The 28D has poor corner sharpness that Leica wouldn't let out of the door (OK that would be my imaginary Leica with the QC department!) and the 80D is brutally and clinically sharp to the corner but it is brutal and clinical. A look I like sometimes... the Leica genius is accuracy and poetry in the same frame. Aaahhhhh....
 

Guy Mancuso

Administrator, Instructor
Tim, Leica has brutal and clinical lenses also. Also the Leica 19mm has poor corner sharpness on FF. Leica is far from perfect with some of there glass, I like the kool aid too but your talking 35mm lenses and comparing them against ANY MF lens is a different ball game.
 

tashley

Subscriber Member
Tim, Leica has brutal and clinical lenses also. Also the Leica 19mm has poor corner sharpness on FF. Leica is far from perfect with some of there glass, I like the kool aid too but your talking 35mm lenses and comparing them against ANY MF lens is a different ball game.
Yup, I've had some Leica dogs and they've all been 35 lux or cron but aside from that all my Leica glass is amazing corner to corner and none of my Phase glass is that good, with the exception of the 80D but as I said there's no poetry to its accuracy. I stand by what I said: I can't trust most of my Phamiya glass to give me an edge to edge usable frame (I include the 28D here) and I can trust all of my Leica glass to do so. So I get all my Leica pixels and fewer of my Phamiyas, which makes my framing accuracy slightly more accurate on the Leica!

Having said that, I have a Bicam arriving soon with some nice Schneider glass and I do expect that to be different.
 

robsteve

Subscriber
Tim, Leica has brutal and clinical lenses also. Also the Leica 19mm has poor corner sharpness on FF. Leica is far from perfect with some of there glass, I like the kool aid too but your talking 35mm lenses and comparing them against ANY MF lens is a different ball game.
Guy:

I would argue that the corner sharpness with the 19mm on full frame has more to do with the sensor than the lens. Has anybody shot it on a sensor with angled micro lenses in the corners? Is wasn't bad on film.

I bet you could mount one to a technical camera and shoot it on a mfdb and get better results than it did on a Canon full frame sensor.

Robert
 

dougpeterson

Workshop Member
Bob - I thought you had a horseman with an M plate??

We have at least one. If we look around the shelves I bet we have more.

Technically called a "Flexadapter Plate for Mamiya mount digital backs". The moniker being important to note since it's the same plate as used in the Phase One Stitching FlexAdapter for traditional view cameras.

Or am I missing something of irony??
 

Bob

Administrator
Staff member
Sorry, Doug I was not clear.
Yes, I have the flexadapter plate.
I am looking to see fi I can find a horseman lens plate with a Leica M lens mount on it.
thanks
-bob
 

Guy Mancuso

Administrator, Instructor
Robert you should not need micro lenses on the sensor to have corner sharpness.The canon does not have it nor does the Phase 25 Plus. We all know extreme wide angles can and most do have corner sharpness issue. The leica 15mm 3.5 and even the 2.8 are not exactly stellar in the corners. I owned all of them on the DMR with the crop not bad but on a FF canon it was there also. Very hard to make any extreme wide angle lens to begin with. Hassy corrects it in there software and i am waiting for Phase to do the same thing with there 28mm in the meantime i have a cheat going for it and it works pretty good. I shot a lot of very important interiors with it and they look great. Most of the issues with these lenses has more to do with the curvature of the front element and that seems to be more the case than anything else.
 

robsteve

Subscriber
Robert you should not need micro lenses on the sensor to have corner sharpness.
You need it with the really wide angles because the light hits the sensor at such an oblique angle it is not hitting the photo site square on, but from the side.

A good example are lenses that looked like crap on the Epson RD1, look better on the M8.

Robert
 
Top