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!

The A7r versus MFDB

Ben Rubinstein

Active member
Thanks Bill for the second test! I also believe the compression really isn't helping. I've noticed the highlight graduations too in my own tests between the D800e and the Aptus II-8. It's very subtle but there. I wonder if it's just bigger pixels on a bigger chip? Or is it CCD?
 

Taylor Sherman

New member
To me, looking at the images above, the most important difference is the colors. The Phase colors are more natural.

Now, this could have everything to do with the color profile(s) used while developing. I've stopped using the "Adobe Standard" color profile w/ the A7 because it's just not realistic (I know Bill was using C1 so that's not directly applicable).
 

RVB

Member
To me, looking at the images above, the most important difference is the colors. The Phase colors are more natural.

Now, this could have everything to do with the color profile(s) used while developing. I've stopped using the "Adobe Standard" color profile w/ the A7 because it's just not realistic (I know Bill was using C1 so that's not directly applicable).
I would agree with this,the Phase color looks better...

Rob
 

philip_pj

New member
Thanks for the final processing version, Bill, a clear difference. I see it as more photorealism - a higher level of 'being there', that comes with higher resolution. The Phase sensor is a (very) low ISO monster.

These guys below got similar results using a D800e against the IQ180:

D800(E) vs IQ180 | The Circle of Confusion

The next big thing arrives with the 44x33mm 50Mp Sony MF sensor, presumably with the same level of tech development as the a7r, so very nice ISO 6400 - but with 68% more sensor real estate - pixel dimensions will equal those of a 30Mp 24x36 camera.

So please Sony, are you making an MF RX2 with a leaf-shuttered 45/2 Sonnar lens on that one, just scaled up a little? Everyone says you make cameras not lenses, so don't disappoint them - or us. MFD in your hand, an everywhere carry, choice of compressed files or not. Sony could thus keep their MFD customers happy and still own the high end compact camera market.

And stick the 36Mp sensor in the present RX1 for good measure, so they can offer RX1/RX1r (24Mp) - RX1/RX1r (36mp) - RX2r (50Mp).

Phase must be happy with any trade-off at the low end of ISO in this new sensor for them - gaining versatility for any loss in low ISO performance, unless Sony has made a further breakthrough, of course.

IQ250 Back Announced

http://news.softpedia.com/news/Sony...ll-Be-Completely-Different-Rumor-434024.shtml
 

Georg Baumann

Subscriber Member
I'm not so sure, because the 16bits in MFDB is pretty much a marketing thing: Digital Medium Format / 35mm equiv difference - FM Forums

That being said, the A7R's raw files are compressed, and that could be the difference. They are essentially what the "cRAW" option used to be on the A900.
Pheew. Thanks Douglas, I guess I have to revisit this thread after breakfast, it is a little hard to digest on an empty stomach. :D
 

philip_pj

New member
Listening to that LL interview re the new 50Mp sensor, the Phase team remark that the noise even at high ISO (with some paraphrasing):

"the behaviour, the nature and how we treat the noise, it's interesting noise, like a silky noise that (waves arms) seems to merge with the image rather than ugly noise...old one I would not use at more than 200 to 400, now I am easily shooting at 6400...that's 32 times up."

On CCD to CMOS: we can achieve exactly the same thing, when you see it you will really get blown away...we would of course very much prefer to have a full size (MFD) CMOS chip...I shoot at ISOs so I can capture things I have never been able to capture, before I had to use a tripod...it's really amazing...it will help wedding photographers...heat is not a problem, not at all, compared with CCD...we've started to explore into the video world...this is like 8k...we are not there yet...14 stops DR (a7r=14.1 stops)"

So the feedback from these guys is very enthusiastic.
What Sony could do with optimisation for an 'RX1' type of camera would appear to be pretty special wrt microlenses and assembly. The noise character is a good description of the 24mp and 36mp sensors, as well. It looks like being a 68% higher res version of the a7r part.
 

Ulfric Douglas

New member
Yup that's the second thing I noticed ;
To me, looking at the images above, the most important difference is the colors.
In fact to my eyes the top prints (from both tests) look "purple", and that's something that very often pops out at me from a wide range of Sony photos ... and it shouldn't because everyone's processing them differently. :confused: and whitebalance can't be blamed in files processed from RAW.
But still ... I see it often.
 

Stefan Steib

Active member
It took me about a week to find my setup for the Sony A7r.
I tried Aperture first, then switched to CaptureOne and finally landed at Lightroom. That worked by far best.
Especially as I am now using a lot of old Canon FD Glass which is amazing but has some longitudinal chromas and lightroom´s new chroma control is totally superior to both Aperture and C1.

The sharpening needs also twiddling and the curves for the lower ISO´s of the Sony also need tweeking, but after that ....?

Greetings from Germany
Stefan
 

lowep

Member
Given the cavernous price delta between these systems, I would expect the MFD rig to maybe make me dinner once in a while.

How many of us take perfect shots every time? Yes, this is more likely to happen when there is no pressure, no imposition of other demands apart from following the light, and no need to try and squeeze orange juice out of a lemon if the light happens to be awful, the distractions too many, the client too stupid, or the concentration not what it should be.

For this reason I would be quite interested to know how the A7 file compares to the MFDB file when you have to squeeze and twist curves, exposure, etc etc in post to salvage an image that in the original form leaves a lot to be desired?

Now if one of these two camera types can do that significantly better than the other then it would really make me dinner once in a while
 

Shashin

Well-known member
I have a question regarding your process and I guess for all you guys who test this stuff. I often see statements like "Both files were processed in C1 at defaults; no additional sharpening was performed" and "straight from the camera" or whatever, is there a reason for that? I ask because to my mind, the camera technology has improved greatly but it's nothing without proper processing and by not working to optimise a shot in post then all you're really showing is C1's or Lightrooms default processing of a RAW file. Having done a lot of experimenting with my own images, I have found that the biggest change in my own work has been switching to C1, it blows me away but not necessarily when I open the file.

I could be wrong but to my mind, modern software gives you the tools but doesn't necessarily do the job for you, just opening the file and printing it seems like such a waste.
Mat, because you want to see what each camera does. If each is processed, you have no idea what was the camera/lens and what was the processing. The idea is not to make an ideal or even final image, but to make a baseline image that shows the difference in how each camera performs.

Then comes the analysis of the images, which will be based on the experience of the observer (you) to know whether the difference is significant and could be improved with processing or just whether one is actually "better" than the other. One problem I have with all these test is I have no idea what the original conditions or colors are. This is what makes testing like at DPreview preferable as the condition is far more standard and consistent. But Bill does not have that target and I am grateful for his time and effort.

Processing is rather subjective. And to get two images from two different system to an ideal, it would require different processing. And would would still end up with two different results. And this starts to get more problematic as we don't know the skill of the person doing the processing. Those images would tell me nothing about the cameras. This is the problem with really slick images from a camera company--it tells you more about the skill of the photographer and nothing about the camera.
 
M

mjr

Guest
Hi Will

I sort of understand what you are saying but honestly, I'm happy to accept that i don't get it!

I don't believe that "to make a baseline image that shows the difference in how each camera performs"does in fact show how each camera performs, it's just showing how good or bad C1's default processing of RAW data is, it's showing no more than that. I read and see lots of images from different camera makes that are made or broken by which processor is used, producing a baseline image to see how each camera performs is impossible, especially in this case where it's a landscape. Saying that, it's only the rest of us that can get little but anecdotal information from a test like this, for Bill it's extremely important for all the reasons you mention, it's his scene, processed in his way and printed on his machine, he knows what he saw and what he wants from the shot. As he wanted to test what if any differences where visible in a good sized print, the only way to do it in my view is to process the shot how he wanted to, any other reference point is irrelevant because surely we all want the best from our files?

I'm not against testing, in fact I agree that the target type of testing is more relevant in certain situations although I never do it myself. The target stuff removes interpretation from the equation, it's only showing resolution, sharpness etc. I understand those tests but I shoot what's relevant to me and process in my own way and decide based on that, mainly because the variables are so wide and varied that a single test means nothing. For example...

I have a D800, D800E, Sony RX1 and a couple of older Canon bodies and lenses, in perfect light, landscape scene, the RX1 produces a nicer file to my eye than anything other than the Zeiss 21mm or 135mm on the D800E, any other combination is flat in comparison, thats as long as I want the 35mm focal length. If I'm shooting a portrait, the 135f2 beats everything in natural light, amazing but in the studio under strobes, the cheap 85 1.8 is incredibly sharp, in daylight it's flat and I hate it. etc. etc. I understand my kit enough to know that a certain situation benefits a certain piece of kit that may not work in another situation. I have a ipf6400 and I'm now printing a fairly large amount of images, at 24x36 I can get a nice image from all of my cameras, including my old Canon 20d if I take a shot that plays to its strengths.

Anyway, I'm rambling. The test is great for Bill and interesting to comment on for the rest of us who aren't lucky enough to handle both cameras, I do lots of testing but see no relevance for others. The goal I guess for all of us is to do what we can with what we have and also to enjoy looking and discussing stuff that other people produce, it's all good!

Enjoy your Sunday.

Mat
 

Quentin_Bargate

Well-known member
Considering my first digital SLR was a Fuji S1, with a 3.2mp sensor and only Jpeg files (no RAW), I'd say were are starting to see some progress :ROTFL:


OK, so here's another approach! A rather dull misty day but very little wind to move the trees around. I set up the cameras with their optimum ISOs and apertures. I exposed so the lighthouse was on the verge of clipping.

Note that my qualitative comments are from poring over the prints, not the screen.

A7r with FE 55, at F 5.6, ISO 100.
IQ180 with SK 80 LS at F8, ISO 35.

In both cases I used C1 to wring the most out of the files (bringing up shadow detail primarily.) I corrected for purple fringing and aberrations though the SK lens was automatically adjusted in C1. I sharpened for inkjet printing. I think the files were about as good resolution-wise as I could make them.

The 36 inch prints, like yesterday, were hard to tell apart and I haven't shown them again. The DR is excellent in both cases, holding all the detail in the lighthouse and in the dark areas under the trees. The Phase file is superior with a loupe but at normal viewing distance there's little discernible difference.

I then printed a crop of roughly half the frame, which would be the equivalent of a 60 inch print.



This time there is no contest. The Phase file is crisp and clean and the sign on the left side is easily legible, whereas it cannot be read in the Sony file. The sky has a subtle gradation not present in the Sony file. The bare trees and the communication mast are a bit fuzzy in the Sony print but crystal clear in the Phase print.

My conclusion is the same as yesterday - up to 36 inches, with the FE 55 lens, the Sony holds it's own. For the bigger print, let's just say I'm not giving up my MFDB!
 
Top