DxO P45+ test: Let the Games Begin

[Bob]

OK, I really really know that there is more to a system than a sensor, but here is the latest log for the fire from DxO Labs

http://news.cnet.com/8301-13580_3-10154386-39.html?part=rss&subj=news&tag=2547-1_3-0-5

 

[LJL]

Bob,
Thanks for posting this link. I had just seen it myself and thought the same thing ("let the games begin").

Maybe I am a bit slower on the uptake or something, but the data directly from the sensors that DxOMark uses surely seems to suggest the high cost MF sensors are not delivering more....at least in the parameters that are posted. Personally, I still have a hard time resolving some of this. It finally starts to hit home when one really looks at the 35mm DSLR units in their own comparison. Then it becomes clear that there are some measurable differences, with the D3X coming out on top. I still am having a hard time translating those differences to what one sees in many images, so this MF data is just going to add more twist to the knots.

My personal take on this is that the data suggests MF is only producing a much larger image, which alone can be very important. The other point that seems to come back is that the glass really must be doing a lot in MF, and we know it trumps 35mm DSLR stuff already. From all of this, one could conclude that IF the 35mm DSLR folks could start fielding better optics, they would be matching or maybe even beating MF for all but overall image size. Trust me, I find it hard for me to even type that, hence my somewhat veiled concern over just what the DxOMark stuff really means by itself, and then just how much "system" stuff goes into the mix after that.

I know this sounds all over the map (need another espresso or something today), but on the surface, it does not bode well for MF, except for maybe the very highest end new stuff (P65+ and others), which have not been tested. From other shots and discussions here, it suggests the newer, bigger sensors may have added a bit, but one has to start asking just how much extra is there and at what costs? For a handful of folks it surely will be important, but these data are suggesting that for most things, one will not see vastly superior image files from MF over DSLR....with the exception of size and maybe overall resolution with better glass.

Am I the only one starting to have that sinking feeling?

LJ

 

[Bob]

I have peeped at a lot of files, and although the DxO tests might be "scientific", they don't show what I see in the files. Perhaps it is AA related, or perhaps it is related to lenses, or maybe to the relationship of pixel size to lens MTF, but despite their scores, A P45+ file looks a hell of a lot better (to me at least) than a Canon 1DsMIII. I haven't had the chance to look at the D3X yet, but the D3 was a terrible disappointment to me.
So I shoot with what gives me the most pleasing results. That is one reason that an M8 is still in my bag along with the Phamiya/P45+
-bob

 

[Jack]

Shoot a two-dimensional high-contrast grayscale test target or a 7-degree tilted black and white Koran rectangle with a top end DSLR and lens outfit against a top MF back and lens and guess what? They won't look (or calculate out) all that different; the edges on those targets that define the differences will look and equate pretty darn similarly. Now shoot real three-dimensional subjects in real studio or available light situations and guess what? The difference between systems becomes immediately recognizable, and in most cases quite significant.

Cheers,

 

[Guy Mancuso]

I will refrain from my usual comment on this stuff because it would not be pretty. LOL

Not selling my MF system anytime soon.

 

[LJL]

Bob and Jack,
I am not arguing with you guys on this at all. That is why I keep scratching my head over just what the DxO stuff means. They are only dealing with the sensor....no processing, no optics, etc. That remains my point about what the 35mm DSLRs are NOT providing (better glass), and that we know makes a huge difference in the final image.

The question remains.....IF, and that is the big word, IF Nikon, Canon, Zeiss or whomever could produce truly stellar lenses for the 35mm format size to close that gap, would the images we look at really look better than what we tend to see now with MF?

LJ

 

[Paratom]

Bob,
Thanks for posting this link. I had just seen it myself and thought the same thing ("let the games begin").

Maybe I am a bit slower on the uptake or something, but the data directly from the sensors that DxOMark uses surely seems to suggest the high cost MF sensors are not delivering more....at least in the parameters that are posted. Personally, I still have a hard time resolving some of this. It finally starts to hit home when one really looks at the 35mm DSLR units in their own comparison. Then it becomes clear that there are some measurable differences, with the D3X coming out on top. I still am having a hard time translating those differences to what one sees in many images, so this MF data is just going to add more twist to the knots.

My personal take on this is that the data suggests MF is only producing a much larger image, which alone can be very important. The other point that seems to come back is that the glass really must be doing a lot in MF, and we know it trumps 35mm DSLR stuff already. From all of this, one could conclude that IF the 35mm DSLR folks could start fielding better optics, they would be matching or maybe even beating MF for all but overall image size. Trust me, I find it hard for me to even type that, hence my somewhat veiled concern over just what the DxOMark stuff really means by itself, and then just how much "system" stuff goes into the mix after that.

I know this sounds all over the map (need another espresso or something today), but on the surface, it does not bode well for MF, except for maybe the very highest end new stuff (P65+ and others), which have not been tested. From other shots and discussions here, it suggests the newer, bigger sensors may have added a bit, but one has to start asking just how much extra is there and at what costs? For a handful of folks it surely will be important, but these data are suggesting that for most things, one will not see vastly superior image files from MF over DSLR....with the exception of size and maybe overall resolution with better glass.

Am I the only one starting to have that sinking feeling?

LJ

Bob,
I just spent 1 week vacation and brought a M8, D3x and Hy6 with Sinar back.
Even though I dont really want to post comparison images, because there were small "faults" like changing light, slight missfocus, differences in converters etc. my overall opinion however is the following:

1) Compared to the D3x the MF back seemed to show smoother tonal transitions, the images looks "deeper", clearer, more "3d" to me. If you pixel peep at 100% there might also be a small detail advantage, but not that much IMO. Its more the overall look of the MF image, and I believe that no AA filter and 16bit are still an advantage, however my personal feeling was that the differences are not as clear as shown in some other comparison threads. Maybe since I used high quality primes on the d3x like the 24PCE.
I also think that there is a difference in the transition between sharp and unsharp areas of the image caused by the sensor size-which leads to a smoother transition for the bigger sensor.
I am not sure about the dynamic range yet, there were some images were I though the DR from the d3x could even be a bit better than that from the SInar back-but I am really not sure at all here.

2) A slight missfocus and the detail advantage of MF seems to be gone. Besides the fast and good AF the D3x has a very good metering, a nice AWB, great high ISO etc etc. There were some situations where I took images with the d3x which would have been very difficult if not impossible to take with the MF-camera. We went into a hole of the glacier, with dim light - not much room and time and light - the d3x worked great.

3) My conclusion would be that:
MF-digital: Image quality: 99% (if you get everything right), flexibility: 40%
D3x: Image quality: 90%, flexibility 99%
For landscape and slow pace I prefer the MF, not only the IQ but also the handling, the viewfinder, the feel of the camera and lenses.
On the other side it is really hard to not nail focus, exposure etc with the d3x, in each and every situation and if you dont nail it, you can see on the display that you didnt nail it (in comparison to the display of the SInar back- they could have made it B&W and it wouldnt be much worse)

4) Regarding price of the D3x: I see that it is expensive but if you allready own good Nikon glass and see that you can get more out of your glass I think it is

5) If you dont have to print big I am stil surprized again and again by the IQ I get from the M8 - besides somewhat interesting colors sometimes.

 

[LJL]

Guy,
Not encouraging anybody to do that or not, nor throwing rocks at MF. Folks have nearly obsessed over the minute differences in the 35mm DSLR data, and will argue over that forever. This new MF data suggests that there is a lot of something else going into the final piece....lack of AA, better optics, etc., and that alone shows a big difference. I am simply wondering if the image quality gap between MF and the top end DSLRs could be closed more with really better glass and maybe tearing off the AA filter. I know folks have tested this before, but not with the newest sensors, and the question about is there any 35mm glass that can really deliver?

Again, not trashing MF. Just taking a look at the sensor data, and having to ask the question about what accounts for the differences we see and prefer in MF. If it really boils down to just the glass......?

LJ

 

[evgeny]

Yes, please say the MF does beat 35mm DSRL.

 

[Bob]

Maybe they could close the gap in glass quality, but in order to do that on a smaller sensor they have to be better by the ratio of sensor size.
The dslr makers are also selling to folks who have different goals than I do apparently.
Not shooting weddings, Not shooting sports, Not shooting (much) macro, Not using AF (much), Not needing anything higher than iso 50 actually I would really like ISO 12 to save me from carrying ND filters), I guess for me, much of the big advantages that are touted for the dslrs are mostly don't-care or little-care. So I look at the images and unless I can get lost in them, they are not worth much to me.
Oh I am positive that there are a whole lot of folks who need, or think they need, all that stuff, and maybe a lot of them just push the camera jpgs to their clients to save time in their work-flow. I just an sure that I am not a member of that set. On the other hand, once there exists something that tickles my retina, I will be all over it in a split-second.
-bob

 

[dougpeterson]

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".

These sorts of comparisons are always useful, but:
1) There are dozens of variables other than sensor specs that make the print on the wall sing. Lenses, color fidelity, post-production flexibility, raw convertor, conversion settings, proper sharpening, etc etc etc
2) In a tough spot a small difference in DR or Resolution can make a large difference. The difference between light puffy clouds at sunset and pure paper white can be a fraction of a stop of DR.
3) You don't make prints of resolution charts and DR step-charts. Though they probably could qualify as modern art nowadays.

Doug Peterson, Head of Technical Services
Capture Integration, Phase One & Canon Dealer | Personal Portfolio

 

[LJL]

Yes, please say the MF does beat 35mm DSRL.

I think there is little doubt that this is the case. All one has to do is look at the images to be convinced. My "devil's advocate" questioning in this thread is more for discussion. No panic needed. I think the DxOMark data is not truly representative of what the actual MF systems deliver, and should be taken with a lot more than a few grains of salt. I just do NOT see the data matching what folks produce and see from their MF systems every day.

LJ

 

[LJL]

Maybe they could close the gap in glass quality, but in order to do that on a smaller sensor they have to be better by the ratio of sensor size.
The dslr makers are also selling to folks who have different goals than I do apparently.
Not shooting weddings, Not shooting sports, Not shooting (much) macro, Not using AF (much), Not needing anything higher than iso 50 actually I would really like ISO 12 to save me from carrying ND filters), I guess for me, much of the big advantages that are touted for the dslrs are mostly don't-care or little-care. So I look at the images and unless I can get lost in them, they are not worth much to me.
Oh I am positive that there are a whole lot of folks who need, or think they need, all that stuff, and maybe a lot of them just push the camera jpgs to their clients to save time in their work-flow. I just an sure that I am not a member of that set. On the other hand, once there exists something that tickles my retina, I will be all over it in a split-second.
-bob

All good points, Bob, and that is what I was trying to get at. As I mentioned in another reply, I just do NOT see the DxOMark data on MF sensors truly being able to characterize what is delivered. Not saying the data is not correct or accurate, but more that it does not take into account all the other very important components that go into creating a great image capture. The glass and sensor size interplay in a way that DxOMark is not measuring, and that also translates into things we find so wonderful in the MF images.

Sorry if folks got nervous or defensive over my comments and questions, but it is good to wonder what is missing in 35mm DSLR that may or may not ever be done to "close the gap".

LJ

 

[Paratom]

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 . . . . ?

 

[LJL]

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 . . . . ?

Not sure it is, but there is something about how it works with the sensor sizes in MF, plus the lack of AA filter on MF sensors that is not quite the same on the 35mm DSLR side. The lenses you mention are all reputed to be among some of the best optics of the format. We already see some differences in images when used on cameras without the AA filter, such as the M8 or DMR, but for whatever reasons, the resulting images are still not the same as those taken with MF. Can somebody else help explain that part? We see it, but how does one measure it?

LJ

 

[Bob]

Until the sensors get larger in dslrs, then mf glass needs to be only a bit better than half as good as dslr glass to win. It is the resolved spatial frequency at mid-range contrast that tells the tale.
-bob

 

[jklotz]

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 . . . . ?

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.

 

[KeithL]

OK folks, you're just having a nightmare, you'll wake up in an hour or two and all will be well...

 

[Jack]

Until the sensors get larger in dslrs, then mf glass needs to be only a bit better than half as good as dslr glass to win. It is the resolved spatial frequency at mid-range contrast that tells the tale.
-bob

Precisely.

 

[Guy Mancuso]

Exactly what we are finding out.