Is the difference of DR on MFDB vs 35mm dslr discernible on print?

[bjanes]

     An excellent example of what not to do. For the benefit of anyone contemplating doing a serious comparison between the P65+ and the D3X, one should (1) pay particular attention to equalizing the FoV in one of the dimensions, through choice of lens focal length, and (2) one should equalize DoF by changing the F stop by the same multiplier that has been applied to the focal length.

The photographer on Flickr has completely ignored point (2) and got point (1) slightly wrong. The P65+ shot uses a 150mm lens at F11.  For a portrait aspect ratio, the equivalent FL for the D3X should be 90 mm, and the equivalent F stop either F6.3 or F7.1. The Flickr guy has used F13 with the D3X.

I think everyone except a complete beginner would appreciate that any good lens will be noticeably sharper at F6.3 than at F13 on a high-pixel-density camera such as the D3X.

In the following 100% crops, I downsampled the P65 image to the same file size as the D3X. If your monitor has a resolution of 120dpi, the 100% crops represent a print size of 33"x50". At larger sizes than this, the P65 would have a resolution advantage. In the crops below, the P65 actually does show a resolution advantage, but only as a result of an inappropriate choice of F stop.

[attachment=16795:P65_v_D3X.jpg]

Ray's comments confirm what Michael stated when he published his megapixel shootout: Whatever you do, there will always be those who will nitpik--this or that should have been done. There are so many variables it is difficult to control them all. I downloaded the P65 and D700 shots. Both look very good and noise is not a problem with either. The photographer used the base ISOs, 50 and 200, respectively. Should he have used the same ISO for both? I would say no, since in a studio situation, one could shoot at the optimum ISO. Resolution is less critical with a portrait, which is basically a low frequency image, than it would be with a landscape. Also, the scene appears to fit into the DR of both cameras, so we really are not looking at a dynamic range comparison, which is the goal of this thread. For making an A4 print, I doubt that the D65 would show a definite advantage. The P65 would allow printing at larger sizes with no loss of quality.

These shots do not really demonstrate the advantages of the P65: resolution and purported high DR. A high contrast landscape would show the P65 to best advantage, and a low light street scene would show the D700 to better advantage. Something not mentioned thus far in this thread (AFAIK) is that vibration reduction technology is available for dSRLs and not for MFDBs. As Michael likes to quote the British: "Horses for courses".

[Dick Roadnight]

The P65 would allow printing at larger sizes with no loss of quality.

These shots do not really demonstrate the advantages of the P65: resolution and purported high DR... "Horses for courses".

A half length portrait of a lady with waist-length hair, printed 24" * (18" or 34"), (like you see in hairdresser's windows) showing every hair, would be a good test to see the difference between serious cameras!

[bjanes]

A half length portrait of a lady with waist-length hair, printed 24" * (18" or 34"), (like you see in hairdresser's windows) showing every hair, would be a good test to see the difference between serious cameras!

Yes, it would show resolution, which was my point. Retouching of skin pores, facial hair, and blemishes might be necessary! No one doubts that 65MP can have better resolution than 12 MP. If you need large prints, you should use a high MP camera or perhaps stitch with a lower resolution camera. Stitching works well for landscapes, but for a portrait, the subject might move. You would be back to the 19th century, when the models had to remain motionless for several seconds or longer. No wonder they had a stiff countenance  . To increase DR, you could stack the images and perform HDR, but the same limitations would apply.

[tho_mas]

In the crops below, the P65 actually does show a resolution advantage, but only as a result of an inappropriate choice of F stop.

you don't have to take the P65 capture to show that MFDBs have an advantage in resolution/sharpness. Take the CF22 shot which is still much more detailed than the D3x. All the MFDBs in the comparision are much sharper than the DSLRs.
IMO the kind of comparision posted at Flickr is the only one that counts: the photographer shoots a motif he/she regularly shoots and compares what he/she likes better. For this kind of motif - given the examples - me personally I'd prefer the CF22, the H3D50 or the S2... simply for the look of the images. The DSLRs show some clumpy noise in the grey background and the skintones are ... unsatisfying. Matter of taste...

[Panopeeper]

IMO the kind of comparision posted at Flickr is the only one that counts: the photographer shoots a motif he/she regularly shoots and compares what he/she likes better. For this kind of motif - given the examples - me personally I'd prefer the CF22, the H3D50 or the S2... simply for the look of the images. The DSLRs show some clumpy noise in the grey background and the skintones are ... unsatisfying. Matter of taste...

I would like to know, where exactly you see the "clumpy noise"? (I have not looked at the Finepix S5, nor at the Nikon D90.)

The skintones are not produced by the camera but by the raw rendering; I wonder, why this has to be posted again and again.

[tho_mas]

I would like to know, where exactly you see the "clumpy noise"?

[attachment=16796:bgs.jpg]

of course printed at camera resolution and at 300dpi you won't see it. But you clearly see it when you do some blow up... especially in the sky or similar parts of the motif.

The skintones are not produced by the camera but by the raw rendering; I wonder, why this has to be posted again and again.

yes, of course. But... why do all the MFDB files (in the comparision) look okay and all the DSLR do not? Why do the MFD files (all together) look so different to the DSLR files (all together)?
Don't get me wrong - I find all that captures quite good and if I'd do portraits at moderate print sizes I can't see any reason not to choose one of the DSLRs (other than a waist level finder and the super shallow DOF of MF... if this is the style you might prefer for portraits). But there clearly is a difference - whether you value the difference high or low is up to you.

[BJL]

The P65+ shot uses a 150mm lens at F11. ... The Flickr guy has used F13 with the D3X.

The rest might be nit-picking, but this is a valid point in sharpness/resolution comparisons if those are the apertures used. Using the smaller sensor and smaller photosites of the D3X at this higher f-stop introduces significant potential for diffraction effects that could be easily avoided. Opening the D3X up to about f/8 would still give about as much DOF as the P65+ shot and significantly less diffraction.

Do not mark a camera down on the basis of a poorly executed sample photo!

[Panopeeper]

But you clearly see it when you do some blow up... especially in the sky or similar parts of the motif

I see it clearly in the S5 Pro (not S2) shot, a three years old DSLR, far from the state of the art, and in the D90 shot, an entry level DSLR. I can't fathom why the photographer included these with professional cameras.

why do all the MFDB files (in the comparision) look okay and all the DSLR do not? Why do the MFD files (all together) look so different to the DSLR files (all together)?

Let's separate the issues. The color is only rendering. ACR supports camera color profiles; there is no need to use the defaults. Btw, I like the 5D2's skin rendering; I think the D3X is too reddish.

The "look" is another issue. I am not saying, that the other cameras are real competitors of these MFDBs in every situation; I am trying to understand, what exactly causes the different perception (the lack of AA on the MFDBs should not make a difference when viewing so huge images for example in monitor size).

Btw, the 5D2 and the D3X shots were made with 70mm, f/13, the CF22 with 135mm, f/11. Though I don't think the possible diffraction because of the smaller pixels and the smaller aperture with the DSLRs is visible on monitor size.

[tho_mas]

I can't fathom why the photographer included these with professional cameras.

Just to know how they look like in the comparision? He didn't make the comparision with regard to the "topic" in this thread.

ACR supports camera color profiles

:-)

I am trying to understand, what exactly causes the different perception (the lack of AA on the MFDBs should not make a difference when viewing so huge images for example in monitor size)

I think the lack of an AA filter makes a huge difference - at 100%, downsized or upsized. A friend of mine is a (very) high volume shooter of portraits (magazines, editorials): Whenever it's possible he uses the M8 (actually he uses 2 at the same time to get the relatively slow buffer managed) only because he likes the crisp look (without applying agessive sharpening). Me I thought about an A900 but decided for the P21+ in addition to my P45. Of course not only because of the lack of the AA-filter... but because of the whole look and because - for me (!) - ISO400 and ISO800 are high ISO. I never used any higher ISO in 25 years... I don't say the P21+ is "better" (regarding what?) than the A900... I don't know. I just prefer the crisp and somehow "bold" look of the files. And I appreciate that I can use it on my Contax (all those knobs on current DSLRs, especially the Nikons, drive me crazy).
So... I'm the fanclub of "there is something different in MFD files". At the same time I'm quite impressed of A900 or D3X images and I fully understand that people say it's "enough". Yes, it's much more than enough unless you print very large.
Finally... I don't understand the topic here. "Jim2" didn't mention how he is printing... size, motfis and so on. I find it totally pointless to ask for the "best" solution for "landscape", "architecture" or "people". All these genres appear in very different modes. DR? IMHO totally overestimated :-)

[bjanes]

The rest might be nit-picking, but this is a valid point in sharpness/resolution comparisons if those are the apertures used. Using the smaller sensor and smaller photosites of the D3X at this higher f-stop introduces significant potential for diffraction effects that could be easily avoided. Opening the D3X up to about f/8 would still give about as much DOF as the P65+ shot and significantly less diffraction.

Do not mark a camera down on the basis of a poorly executed sample photo!

Not only that, but it is ridiculous to talk about sharpness/resolution without specifying the lens!! 70 mm is not usual for a prime Nikkor, suggesting that a zoom may have been used. Of course, sharpening applied to the image is another important parameter, and is especially important with a camera such as the D3x with a blur filter. Sharpening also affects noise. Unfortunately, no information regarding sharpening of these images is available. These tests provide useful information, but are far from definitive.

[tho_mas]

Not only that, but it is ridiculous to talk about sharpness/resolution without specifying the lens!!

That's the typical small-minded shit… sorry :-)

Compare all the DSLR files as a whole to all the MFD files as a whole at the Flickr page…
There are forums where you can compare tons of DSLR files and tons of MFD files (downsized for web use) side by side… as a whole...
Compare 10 DSLR prints to 10 MFD prints (at native resolution) as a whole...
they always look different. Even A900/ZA 1.8/135 @ f4 files look different to MFD files.
Maybe you find an explanation for the difference in clever articles or not… just look at the difference by yourself.

[Ray]

Ray,

In my humble opinion we may be happy that we have a good comparison. The person who published these pictures made a real effort to help.

Regarding your view on equalizing parameters I don't agree. Reason is that whatever DOF we need is dependent on setting. Landscape shooters normally focus on far away objects and may very well choose optimum aperture. In portrait we may strive for short depth of field.

Best regards
Erik

Erik,
Fair enough! Shall we say, good marks for trying. However, it simply isn't helpful to create wrong impressions by ignoring fundamental requirements when doing a comparison. Those requirements are; same subject; same lighting; same FoV; same DoF; same point of focus, and of course a shutter speed sufficient to freeze any movement of subject or camera.

The fact that landscape shooters often desire a more extensive DoF is totally irrelevant in this context. The subject in these comparisons is not a landscape. I can't think why you mentioned it.

I haven't studied the other comparisons at Flickr because I see the D3X as being the closest competitor to the MFDB format.

[Jeremy Payne]

... But you clearly see it when you do some blow up... especially in the sky or similar parts of the motif ... But... why do all the MFDB files (in the comparision) look okay and all the DSLR do not?

I think you are seeing what you want to see ... I don't see this clumpy noise and don't agree that the MFDBs look okay and the SLRs do not.

[douglasf13]

That's the typical small-minded shit… sorry :-)

Compare all the DSLR files as a whole to all the MFD files as a whole at the Flickr page…
There are forums where you can compare tons of DSLR files and tons of MFD files (downsized for web use) side by side… as a whole...
Compare 10 DSLR prints to 10 MFD prints (at native resolution) as a whole...
they always look different. Even A900/ZA 1.8/135 @ f4 files look different to MFD files.
Maybe you find an explanation for the difference in clever articles or not… just look at the difference by yourself.

  I don't agree.  If we are trying to isolate what gives MFDB it's "look," lenses are as much or more of the equation than the sensor.  It is very, very easy to buy medium format lens adapters for 35mm DSLRs, and I think comparing a MFDB with an D3x/A900 would be interesting if the same medium format lens was used on all of the cameras.  I used to have a MFDB, and now I have the A900.  I've begun using my Hasselblad lenses on the A900, but I don't have a direct reference for comparison anymore.  Maybe I need to get that new CFV back so I can do some testing

[Ray]

Ray's comments confirm what Michael stated when he published his megapixel shootout: Whatever you do, there will always be those who will nitpik--this or that should have been done. There are so many variables it is difficult to control them all. I downloaded the P65 and D700 shots. Both look very good and noise is not a problem with either. The photographer used the base ISOs, 50 and 200, respectively. Should he have used the same ISO for both? I would say no, since in a studio situation, one could shoot at the optimum ISO. Resolution is less critical with a portrait, which is basically a low frequency image, than it would be with a landscape. Also, the scene appears to fit into the DR of both cameras, so we really are not looking at a dynamic range comparison, which is the goal of this thread. For making an A4 print, I doubt that the D65 would show a definite advantage. The P65 would allow printing at larger sizes with no loss of quality.

These shots do not really demonstrate the advantages of the P65: resolution and purported high DR. A high contrast landscape would show the P65 to best advantage, and a low light street scene would show the D700 to better advantage. Something not mentioned thus far in this thread (AFAIK) is that vibration reduction technology is available for dSRLs and not for MFDBs. As Michael likes to quote the British: "Horses for courses".

Excuse me! Bill. Failing to get the DoF similar in two shots taken for comparison purposes is 'nit-picking'? What would you say is 'not nit-picking'? Failing to take off the lens cap?

Shallow DoF is one of the characteristic advantages claimed for the larger format MFDB. At any given F stop the MFDB will produce a shallower DoF. The difference in sensor size between full frame 35mm and APS-C is of the same order of magnitude as the difference between the P65 and the D3X sensors. What would you say if I were to compare my Canon 50D with a 5D2 using F11 with the 5D2 and F13 with the 50D, shooting the same subject?

Each format has its advantages and disadvantages. Understanding what those are allows one to choose the best tool for the job. Choosing a specific DoF, whether it be shallow for portraits or extensive for landscapes, is surely a very basic requirement. There may be situations when one of the camera formats in the comparison is so small that it's not possible to select an F stop to match the shallow DoF of the larger format. In such a situation one then understands that this fact points to an inherent weakness of the smaller format.

However, in the Flickr comparison between the D3X and P65+, it makes no sense at all to move in the opposite direction and make a DoF mismatch even worse by stopping down in the D3X shot.

[bjanes]

That's the typical small-minded shit… sorry :-)

Compare all the DSLR files as a whole to all the MFD files as a whole at the Flickr page…
There are forums where you can compare tons of DSLR files and tons of MFD files (downsized for web use) side by side… as a whole...
Compare 10 DSLR prints to 10 MFD prints (at native resolution) as a whole...
they always look different. Even A900/ZA 1.8/135 @ f4 files look different to MFD files.
Maybe you find an explanation for the difference in clever articles or not… just look at the difference by yourself.

I see no need for profanity and my suggestion is hardly small minded. Your post is incoherent. The first step of any experiment is to control the variables and the quality of the lens is a most important variable. As you can see from the MFD backs on the Flickr page, the quality is high and the differences are small. Also, they likely were used with high quality primes. Comparing the dSLRs as a whole is not useful, since there is so much variation between them. Before drawing any conclusions, it would be proper to look at a range of image types or at least an image that has a range of contrasts and image details (as in Micheal's previous shootout). I suspect that naive types such as yourself have already made your judgment before even looking at the pictures. In that case there is no need to control variables and perform a scientific test.

[bjanes]

Each format has its advantages and disadvantages. Understanding what those are allows one to choose the best tool for the job. Choosing a specific DoF, whether it be shallow for portraits or extensive for landscapes, is surely a very basic requirement. There may be situations when one of the camera formats in the comparison is so small that it's not possible to select an F stop to match the shallow DoF of the larger format. In such a situation one then understands that this fact points to an inherent weakness of the smaller format.

However, in the Flickr comparison between the D3X and P65+, it makes no sense at all to move in the opposite direction and make a DoF mismatch even worse by stopping down in the D3X shot.

I'm not sure that equalizing the depth of field among the cameras would be a good idea, but you reasoning concerning the DoF mismatch is good. That is definitely a flaw in the tests.

[bjanes]

I think the lack of an AA filter makes a huge difference - at 100%, downsized or upsized. A friend of mine is a (very) high volume shooter of portraits (magazines, editorials): Whenever it's possible he uses the M8 (actually he uses 2 at the same time to get the relatively slow buffer managed) only because he likes the crisp look (without applying agessive sharpening).

Again, you are not controlling the variables. No MFDB has a blur filter and I know of no dSLR (out of the box) except for the ill fated Kodak full frame which had severe aliasing problems. There are companies who would be happy to remove the blur filter from your friend's dSLR if he doesn't like to have one and does not mind Moire. Perhaps then he would find that his Leica has qualities other than the lack of a blur filter that endear its images to him. Perhaps the legendary Leitz lenses?

[Khun_K]

ALERT: All the pros have left the building and what's left is Nerdsville. People who want to endlessly debate nothingness. (Oh crap, a split infinitive. It just doesn't sound right any other way.)

Profanity is fun. Profanity is the result from frustration of tolerating social ineptitude, artistic ineptitude and what-an-effin'-camera-is-for-in-the-first-place ineptitude.

Be sure to clean up and turn off the lights if you're the last one out.

I wonder how much sense the test of such kind did make?  Quality of a camera, or for that metter, quality of a system, is like a relationship, the photographer needs to develop it, with his tools.  There is no perfect system, with time, you learn how to use the camera/system, knowing what is best and use its best, knowing what is weaker, and work around it, and 5-10 years later, you tell yourself you have a system really satisfy you.  There are debates made over one to another, brand, type, and etc, but all systems have long list of established masters using them day to day with great images, far above the average photographers spend time doing this and that test and spend hours bashing one another.  

Regards, K

[bjanes]

ALERT: All the pros have left the building and what's left is Nerdsville. People who want to endlessly debate nothingness. (Oh crap, a split infinitive. It just doesn't sound right any other way.)

I don't want to knock pro photographers--some are artists and take an intellectual interest in their craft. I admire them. Others merely want to turn our a workmanlike product with a minimum expenditure of time and effort and take another picture of a baby posed on a rug or of a hackneyed wedding pose. Some like to delude themselves that us mere amateurs behold them with awe. However, many of us have our own professions demanding equal or more skill and learning and are interested in photography because it combines art and science and we find it fascinating rather than a means to put meat on the table. We can allow ourselves to indulge our intellectual curiosities. I like this quote: It's about time we started to take photography seriously and treat it as a hobby. - Elliott Erwitt.

I think we can continue this thread in the absence of John-S.