Mini Medium Format Shootout

[Chairman Bill]

INterestingly, the Pentax is offered for sale here in the UK at about £9,000 - clicky. I've just had an email from Hasselblad, offering me a H4D-31 plus 80mm lens for £8,995 plus VAT (local sales tax at 17.5%). I can't help but think this new Hassy price has something to do with the new Pentax.


Edit: spelling

[Mark D Segal]

Mark,
Could this be the reason why you ended up moving to MFDB, your 1Ds3 had a misaligned sensor?

This review is really short on hard facts. Okay! The two of you had lots of fun comparing cameras. No harm in that. But I can't help wondering, after your taking the plunge to move into the ultra-expensive MFDB system, if there remains a need to justify that expensive move.

The subconscious can play interesting tricks. Getting you to produce an incompetent result, when using your 'rejected' 1Ds3, is not difficult for the subconscious  .

Ray, are you a shrink or a photographer? 

Yes of course the review is short on "hard facts". We made photographs, we presented the methodology, the data and the results. What more do you want? We should be camera designers and engineers too?
My subconscious doesn't play interesting tricks when it comes to shelling out over 20000 bucks on a camera system. I'm not wealthy and it was a huge decision. It had absolutely nothing to do with the performance of my 1DsMk3, which on the whole I happen to be very satisfied with and I still use it extensively. Ray, I don't need to "justify" any expenditure to myself AFTER I made it; only BEFORE. And if I made a mistake, well I made a mistake. It's completely foolish to wrap oneself in delusions - better to recognie reality and get over it, but in this case there is nothing to get over. I didn't make a mistake on either purchase. Yesterday I printed a 60 inch pano I shot of the Toronto skyline from a boardwalk around a bend on the waterfront with the Phase-One. It was a stitch of three shots, put together seamlessly in CS5. The total file size is just over a Gig. The pano is stunning if I dare say so myself - for a shot like that NO DSLR would have achieved that kind of detail and clarity for such scene. But at my son-in-law's Halloween party I used the 1DsMKIII because there's no way I could have captured the combination of spontaneity and image quality I got for those kind of photos using a medium format camera. Horses for courses. Each has its place in the spectrum. That was my perception before buying the MF, and remains so now.

[Mark D Segal]

Hi,

I'd just call your attention to this article: http://www.josephholmes.com/news-medformatprecision.html

I enjoyed the articles very much, well written and interesting. I wouldn't expect the old Pentax lenses to hold up to the digital sensor but they obviously do.

Just three comments:

• I have observed that depth of field is incredibly short, at the pixel level, when shooting digital. This applies also to medium apertures.
• There is also field curvature, it's not given that the optical image projected by the lens is entirely flat.
• In recent testing I have done it has been my experience that foliage and treetops are far more critical than brick walls or book shelfs. My Sigma 12-24 is quite decent on both book shelf and brick walls but the image has serious problems at edge and corner shooting real world subjects.

Best regards
Erik

Erik,

I really appreciate these observations - I think because they also correlate with a recent experience I had photographing fall foliage. The leaves were the problem. I think they were moving more than I noticed and possibly the shutter speed wasn't high enough. Well, there's always next Fall! Meanwhile brick walls usually remain quite still which is great for testing cameras, but incredibly boring for photography - unless they have "character" of course.

[Mark D Segal]

INterestingly, the Pentax is offered for sale here in the UK at about £9,000 - clicky. I'f just had an email from Hasselblad, offering me a H4D-31 plus 80mm lens for £8,995 plus VAT (local sales tax at 17.5%). I can't help but think this new Hassy price has something to do with the new Pentax.

Interesting. Phase-1 recently also reduced the price for the P40+/645DF combo. Medium format digital is "coming of age" in the sense that it is beginning to walk the technological and commercial path of much other high-tech equipment: "more for less" as a function of time. This will be a particularly difficult scene for both the producers and purchasers, because the market for this stuff remains very thin in terms of camera sales world-wide per year. A company like Leica will continue to need to charge premium prices because they are providing custom fine-tuning and extremely exacting QC on every unit of camera and lens they ship. The others will be faced with serious dilemmas about the balance to be struck between volume, price, quality, R&D and new models to meet the evolving demands and expectations of the market. Competition has both its costs and rewards.

[BernardLanguillier]

Erik,

I really appreciate these observations - I think because they also correlate with a recent experience I had photographing fall foliage. The leaves were the problem. I think they were moving more than I noticed and possibly the shutter speed wasn't high enough. Well, there's always next Fall! Meanwhile brick walls usually remain quite still which is great for testing cameras, but incredibly boring for photography - unless they have "character" of course.

Hi Mark and Nick,

Regarding this asymmetrical lack of sharpness. What makes you think that the issue you are seeing is the result of a back alignement issue instead of being a camera positioning issue?

In the absence of live view, it is my experience that critical camera alignement is very challenging with 6 micro class photosites. Even with liveview it isn't easy.

Cheers,
Bernard

[ndevlin]

INterestingly, the Pentax is offered for sale here in the UK at about £9,000 - clicky. I've just had an email from Hasselblad, offering me a H4D-31 plus 80mm lens for £8,995 plus VAT (local sales tax at 17.5%). I can't help but think this new Hassy price has something to do with the new Pentax.

Can someone please explain to me how the retail photographic industry continues to exist in the UK?? They rape their customers like Vikings on a weekend cruise, in a day-and-age when Fedex puts B&H at everyone's fingertips.  This is shameless and pathetic, and you guys just shouldn't take it anymore.

I'll personally smuggle you a 645D for return airfare to London, for a lot less 

- N.

[ndevlin]

Thanks Guys!! 
...
If I travel to shoot specifically for landscapes and find that I get superior results when shooting my Nikon and stitching, I cannot help but question why should I invest in MF digital for this purpose? I have yet to receive the "WOW" from my MF digital landscape kit, but before I sell it off, I will be asking my vendor to take it all back and to please tweak all of its components for the "wow." I feel there is something missing in my expensive landscape kit and the experts should have a chance to test it, tweak it, and charge me for the necessary time to bring it all together (if possible).
...
Pentax has made me hopeful, and it is easier for me to trust their future with my dollars mainly because of the "WOW" factor I repeatably receive through their equipment and their timing into the MF digital market. Pentax IMHO, seems to have waited on the sidelines, observing others R&D before they went to their own drawing board, and this I like! It is a smart business model and one that I think will eventually pay-off.

Kind regards,
Darr

Darr,

I know what you mean about the "wow" factor. And also about the power of the pano (from 35mm).  The moral of our story so far is that MF is very demanding, and it's hard to make it much better than 35mm. When it works, it can yield spectacular results, but whether the margin is worth it is very much an individual question. 

Michael wrote an article here many years ago showing convincingly that a 1Ds (Mark I) matched 6x7 film quality, for all purposes save significant enlargement.  This remains true, and has been made (slightly) truer by the larger FF sensors (though I remain convinced that sensors with larger photo-sites give nicer results, somehow). This puts MFDSLRs into the rare-air of 4x5. Few needed it before, and few were willing or able to acquire the skills to perfect the art.  The same may be true of digital MF....with the added twist that we are much more at the mercy of the ghosts in the machine.

As for the 67 lenses, thanks for your generous offer.  I would take you up on it, except I just today received three 67 lenses from friends of mine for testing (90-180, 135 macro and 165 f2.. I'll include the results in a future article on the 645D. I plan to shoot the 165 in-studio on the weekend.
Cheers,

- N.

[darr]

Darr,

I know what you mean about the "wow" factor. And also about the power of the pano (from 35mm).  The moral of our story so far is that MF is very demanding, and it's hard to make it much better than 35mm. When it works, it can yield spectacular results, but whether the margin is worth it is very much an individual question. 

Michael wrote an article here many years ago showing convincingly that a 1Ds (Mark I) matched 6x7 film quality, for all purposes save significant enlargement.  This remains true, and has been made (slightly) truer by the larger FF sensors (though I remain convinced that sensors with larger photo-sites give nicer results, somehow). This puts MFDSLRs into the rare-air of 4x5. Few needed it before, and few were willing or able to acquire the skills to perfect the art.  The same may be true of digital MF....with the added twist that we are much more at the mercy of the ghosts in the machine.

As for the 67 lenses, thanks for your generous offer.  I would take you up on it, except I just today received three 67 lenses from friends of mine for testing (90-180, 135 macro and 165 f2.. I'll include the results in a future article on the 645D. I plan to shoot the 165 in-studio on the weekend.
Cheers,

- N.

Thanks Nick!!
I cannot wait for the results...my fingers are crossed!  

[Mark D Segal]

Can someone please explain to me how the retail photographic industry continues to exist in the UK?? They rape their customers like Vikings on a weekend cruise, in a day-and-age when Fedex puts B&H at everyone's fingertips.  This is shameless and pathetic, and you guys just shouldn't take it anymore.

I'll personally smuggle you a 645D for return airfare to London, for a lot less 

- N.

You can ask that question more broadly - my general observation about prices based on any window-shopping I've ever done during my many visits to London was simply to call a pound a dollar and the numbers remain the same. Beats me how our friends in the UK manage. The cost of just about everything there looks really high to us when converted into devalued dollars.

[Mark D Segal]

Hi Mark and Nick,

Regarding this asymmetrical lack of sharpness. What makes you think that the issue you are seeing is the result of a back alignement issue instead of being a camera positioning issue?

In the absence of live view, it is my experience that critical camera alignement is very challenging with 6 micro class photosites. Even with liveview it isn't easy.

Cheers,
Bernard

Hi Bernard, yes - our initial thoughts too were along these lines, but we weren't really comfortable with that as a conclusion, because EVEN for MF, when your lens is set at f/8 or f/11, and you're 5~6 meters from a flat subject, and you've taken considerable pains to align the camera as parallel to the wall as you could get it - with some pretty sophisticated gear - you'd want to think that the DoF would be MORE than sufficient to cover-off any minor mishaps of camera alignment. But once you realize that errors in front of the lens are measured in cms and those behind in micro-millimeters in terms of their impact on sharpness, it does seem more reasonable to suspect lens/sensor rather than camera/wall. I'm not coming down hard on one explanation or the other - it's simply that the lens/sensor aspect does seem more plausible in the circumstances.

[Ray]

Ray, are you a shrink or a photographer? 

Mark,
I don't believe in strict categories of compartmentalization  .

Yesterday I printed a 60 inch pano I shot of the Toronto skyline from a boardwalk around a bend on the waterfront with the Phase-One. It was a stitch of three shots, put together seamlessly in CS5. The total file size is just over a Gig. The pano is stunning if I dare say so myself - for a shot like that NO DSLR would have achieved that kind of detail and clarity for such scene.

I'm curious to know why no DSLR could have achieved that kind of detail and clarity. The P40 has approximately double the pixel count of the 1Ds3, and less than double the pixel count of the A900 and D3X..

A stitch of 3 x P40 images should be roughly equivalent to a stitch of 6 x 1Ds3 images, depending on camera orientation.

Now I accept that the lack of an AA filter on the P40+ (I'm presuming it has no AA filter) might result in some marginal increase in apparent resolution with nose pressed against the print. Perhaps the ventilation grid on an airconditioner sticking out of the wall of an apartment on the Toronto skyline might have more clearly defined edges.

If that's the case, that would be no big deal for me. What's the size of your printer, Mark?

I have a 6ftx2ft print clipped to a board in the living room of my new house. I'm not sure whether to paste it on the new plasterboard, or mount it so it can be easily removed if/when desired.

It consists of 5 stitched images from the Canon 5D of the Himalayas, with camera vertical. The detail is so good from close up, I get the impression if there were a climber on one of the mountain peaks waving an Australian flag, I'd be able to see him. (But maybe not    ).

Because the image is so impressive, I've been wondering if I should make it larger. My Epson 7600 can manage only 24" wide prints, so anything larger would have to be segmented with an obvious division.

I see two ways of managing this, and I'm not sure which would be more effective. One is not to attempt to disguise the segmentation, as opposed to creating an imperfect seamless join.

The other is to create a realistic representation of a window frame (photographing the existing windows in my house), and use that as a more natural division of the segments in order to create the effect that one is looking out of a window and seeing a view of the Himalayas. What do you think?

Because of the resolution limitations of the 12.8mp of the 5D, I would be reluctant to make a print larger than 3ft x9ft (from 5 x 5D stitched images). For that, I would want a camera with at least the resolution of a 5D2 or D3X, or even an MFDB.

However, if the situation permits taking multiple rows of images for stitching purposes, I can't imagine needing more resolution than a 5D2 can provide, for my purposes.

Cheers!

[Mark D Segal]

Mark,
I don't believe in strict categories of compartmentalization  .
 

I'm curious to know why no DSLR could have achieved that kind of detail and clarity. The P40 has approximately double the pixel count of the 1Ds3, and less than double the pixel count of the A900 and D3X..

A stitch of 3 x P40 images should be roughly equivalent to a stitch of 6 x 1Ds3 images, depending on camera orientation.

Now I accept that the lack of an AA filter on the P40+ (I'm presuming it has no AA filter) might result in some marginal increase in apparent resolution with nose pressed against the print. Perhaps the ventilation grid on an airconditioner sticking out of the wall of an apartment on the Toronto skyline might have more clearly defined edges.

If that's the case, that would be no big deal for me. What's the size of your printer, Mark?

I have a 6ftx2ft print clipped to a board in the living room of my new house. I'm not sure whether to paste it on the new plasterboard, or mount it so it can be easily removed if/when desired.

It consists of 5 stitched images from the Canon 5D of the Himalayas, with camera vertical. The detail is so good from close up, I get the impression if there were a climber on one of the mountain peaks waving an Australian flag, I'd be able to see him. (But maybe not    ).

Because the image is so impressive, I've been wondering if I should make it larger. My Epson 7600 can manage only 24" wide prints, so anything larger would have to be segmented with an obvious division.

I see two ways of managing this, and I'm not sure which would be more effective. One is not to attempt to disguise the segmentation, as opposed to creating an imperfect seamless join.

The other is to create a realistic representation of a window frame (photographing the existing windows in my house), and use that as a more natural division of the segments in order to create the effect that one is looking out of a window and seeing a view of the Himalayas. What do you think?

Because of the resolution limitations of the 12.8mp of the 5D, I would be reluctant to make a print larger than 3ft x9ft (from 5 x 5D stitched images). For that, I would want a camera with at least the resolution of a 5D2 or D3X, or even an MFDB.

However, if the situation permits taking multiple rows of images for stitching purposes, I can't imagine needing more resolution than a 5D2 can provide, for my purposes.

Cheers!

Ray, these discussions reach a point where making inferences about image quality by comparing superficial data sets simply isn't useful. You need to see the results to appreciate it, and likely me - yours. However, in principle, when you are working with a sensor which provides twice the data, 16 rather than 14 bit depth, no AA filter and other design features - Phase-One's "secret sauce" unique to that product, you are going to get a set of raw images which tolerates more cropping and allows less stitching with larger print sizes at higher output PPI than you can get from any DSLR, and all this added together makes a difference to what you see. You know what - I believe Phase One is organizing one or more of their PODAS Workshops in or not too far from where you are between this year and next. You should check-out their offerings and sign-up for one of them. It will give you a pretty thorough hands-on understanding of what their brand of MF can do and why. You are under no obligation to buy anything they put in your hands to work with. Some attendees do and others don't. It's a great learning experience.

[Schewe]

Ray, these discussions reach a point where making inferences about image quality by comparing superficial data sets simply isn't useful.

Mark, I guess you haven't tracked Ray's MO. He's never used a MFDB but is quick to discount any advantage over DSLRs. He's trying real hard to resist the medium format force...I suggest you resist getting into the arena with him.

As far as the test and results...congrats. Although in the grand scheme of things you would have been better off leaving the 1Ds MIII results out of the loop because of the lens used. Heck even a relatively cheap 50mm F1.4 would have out performed the lens used...

Also, I'm inclined to agree that ultimately, the corner/edge sharpness may come down to the sensor alignment. Depth of Field is one thing (and I agree that F8 should have been enough) but Depth of Focus is super critical. I have an older Canon Rebel that could simply not bring anything into focus on the far left side of the frame (right side of the sensor)...regardless of the subject distance. Rather than try to get it fixed, I just got another upgraded Rebel (that was fine).

The bottom line in terms of IQ is the total system...which includes lens, sensor and ISO, and raw processing. Any of the elements can impact IQ.

[Mark D Segal]

Mark, I guess you haven't tracked Ray's MO. He's never used a MFDB but is quick to discount any advantage over DSLRs. He's trying real hard to resist the medium format force...I suggest you resist getting into the arena with him.

As far as the test and results...congrats. Although in the grand scheme of things you would have been better off leaving the 1Ds MIII results out of the loop because of the lens used. Heck even a relatively cheap 50mm F1.4 would have out performed the lens used...

Also, I'm inclined to agree that ultimately, the corner/edge sharpness may come down to the sensor alignment. Depth of Field is one thing (and I agree that F8 should have been enough) but Depth of Focus is super critical. I have an older Canon Rebel that could simply not bring anything into focus on the far left side of the frame (right side of the sensor)...regardless of the subject distance. Rather than try to get it fixed, I just got another upgraded Rebel (that was fine).

The bottom line in terms of IQ is the total system...which includes lens, sensor and ISO, and raw processing. Any of the elements can impact IQ.

Hi Jeff, thanks, glad you enjoyed the article.

Ray and I have discussed MF previously so I do know where he is coming from and that is why I suggested he should enroll in PODAS - a great learning experience.

When we were preparing the article, we did discuss between ourselves leaving this 1Ds3 result out of the loop as you recommend, but for better or worse, we decided to tell it like we got it. And I tried to balance it with my comments about this camera and lens at the end of my section. That 24_105mm zoom can be a fine lens - despite the existence of sample variability which I'm sure you know well can affect particularly zoom lenses quite a bit, due to problems with alignment of the moving element sections. The sample I have is a very good one. My result in that test was disappointing beyond expectation, and purely out of curiosity it is something I intend to drill down to further as time permits.

[ErikKaffehr]

Hi!

Thanks for your suggestion on PODAS. I was considering to rent some equipment for doing my own tests, but it comes at a very significant expense. Taking part in PODAS would be a nice way to learn about MF digital, getting some nice pictures and meeting some nice ladies and gentlemen.

I will consider PODAS workshop next year. Of course there are some boundary conditions, like calendar, checkbook and so on.

Best regards
Erik

Ray, these discussions reach a point where making inferences about image quality by comparing superficial data sets simply isn't useful. You need to see the results to appreciate it, and likely me - yours. However, in principle, when you are working with a sensor which provides twice the data, 16 rather than 14 bit depth, no AA filter and other design features - Phase-One's "secret sauce" unique to that product, you are going to get a set of raw images which tolerates more cropping and allows less stitching with larger print sizes at higher output PPI than you can get from any DSLR, and all this added together makes a difference to what you see. You know what - I believe Phase One is organizing one or more of their PODAS Workshops in or not too far from where you are between this year and next. You should check-out their offerings and sign-up for one of them. It will give you a pretty thorough hands-on understanding of what their brand of MF can do and why. You are under no obligation to buy anything they put in your hands to work with. Some attendees do and others don't. It's a great learning experience.

[ErikKaffehr]

Hi,

Regarding the print size limitation there are probably professional labs around who can make larger prints at reasonable cost. The lab I use here in Sweden has something called Lambda Prefix. The idea is that you make all the work and send a file ready to print to the lab. They print it using Durst Lambda and post it next day. It's without manual intervention so it's quite economical, like 70 USD for a 70x100 cm print.

The Lambda prints are very close to my Epson 3800, so I can use that for proofing. The Durst Lambda prints on photopaper using lasers, so the result is an old fashioned photochemical print. Gamut is significantly smaller than the Epson but the prints are great.

There are some comparable labs in Australia (where you live?), for sure.

Best regards
Erik

Mark,
I don't believe in strict categories of compartmentalization  .
 

I'm curious to know why no DSLR could have achieved that kind of detail and clarity. The P40 has approximately double the pixel count of the 1Ds3, and less than double the pixel count of the A900 and D3X..

A stitch of 3 x P40 images should be roughly equivalent to a stitch of 6 x 1Ds3 images, depending on camera orientation.

Now I accept that the lack of an AA filter on the P40+ (I'm presuming it has no AA filter) might result in some marginal increase in apparent resolution with nose pressed against the print. Perhaps the ventilation grid on an airconditioner sticking out of the wall of an apartment on the Toronto skyline might have more clearly defined edges.

If that's the case, that would be no big deal for me. What's the size of your printer, Mark?

I have a 6ftx2ft print clipped to a board in the living room of my new house. I'm not sure whether to paste it on the new plasterboard, or mount it so it can be easily removed if/when desired.

It consists of 5 stitched images from the Canon 5D of the Himalayas, with camera vertical. The detail is so good from close up, I get the impression if there were a climber on one of the mountain peaks waving an Australian flag, I'd be able to see him. (But maybe not    ).

Because the image is so impressive, I've been wondering if I should make it larger. My Epson 7600 can manage only 24" wide prints, so anything larger would have to be segmented with an obvious division.

I see two ways of managing this, and I'm not sure which would be more effective. One is not to attempt to disguise the segmentation, as opposed to creating an imperfect seamless join.

The other is to create a realistic representation of a window frame (photographing the existing windows in my house), and use that as a more natural division of the segments in order to create the effect that one is looking out of a window and seeing a view of the Himalayas. What do you think?

Because of the resolution limitations of the 12.8mp of the 5D, I would be reluctant to make a print larger than 3ft x9ft (from 5 x 5D stitched images). For that, I would want a camera with at least the resolution of a 5D2 or D3X, or even an MFDB.

However, if the situation permits taking multiple rows of images for stitching purposes, I can't imagine needing more resolution than a 5D2 can provide, for my purposes.

Cheers!

[NigelC]

Can someone please explain to me how the retail photographic industry continues to exist in the UK?? They rape their customers like Vikings on a weekend cruise, in a day-and-age when Fedex puts B&H at everyone's fingertips.  This is shameless and pathetic, and you guys just shouldn't take it anymore.

I'll personally smuggle you a 645D for return airfare to London, for a lot less 

- N.

I don't think it's quite as simple as that. Generally yes prices are higher for consumer durables in the UK - just compare the price of a MacBook Pro! Specifically on photographic equipment, firstly forget about high street chains like Jessops and Jacobs - they trade on consumer ignorance or apathy. The online market is quite competitive and I know their margins are not that great - if anyone is creaming it the distributors who control how low resellers can go. As you are no doubt aware, manufacturers fix prices in different markets according to what they think they can get away with - the margins they have to play with are pretty substantial. Traditionally they have been able to get away with high prices in the UK. Also, in the UK, MFD is sold in more restricted channels where discounting is less common as they know they are mainly selling to professionals who are looking at the reliability of product back-up as much as the initial purchase price. Also, the first price set by Pentax UK is a bit finger in the wind - the UK professional market is quite conservative and it will take some doing to shift people away from Hasselblad/Phase/Leaf.
A few other points:


1.I think the euro price set by Pentax is similar, taking into account current £/euro rate.
2.You should compare prices net of VAT with US prices net of sales tax - if you are doing well enough as a pro to contemplate MFD, you are likely to be above the VAT registration threshold and will therefore claim VAT back.
3.For a retail customer not sure this is as clear cut anyway. Just looked at B&H price for a Canon 5D2 - $2499 or £1567 @ 1.6 exchange rate. However, this presumably excludes sales tax. If I could get them to ship to me in UK I would presumably have to pay shipping, 17.5% VAT plus import duty. I could of course  go over to US and try to smuggle it back but I would still have to pay air fare. However, I can buy 5D2 in UK online for £1561, VAT and shipping included if I pay by bank transfer rather than credit card.
 

[Joseph Yeung]

I would just like to note that with one lens tested on each camera it's not possible to tell whether lens decntring or sensor alignment is at fault.  Perhaps if several lenses were tested on one camera and they all leaned the same way on the brick wall, you would be able to conclude that the sensor inside the camera is misaligned.

[ndevlin]

Here is a sobering article, which we should all read:

https://www.lensrentals.com/news/2010.03.06/this-lens-is-soft-and-other-facts

Then we should all take a deep breath and think of every inspiring image we have seen in the last year, and remember that all of them were taken with some iteration of this 'flawed' technology :-)

- N.

[darr]

Here is a sobering article, which we should all read:

https://www.lensrentals.com/news/2010.03.06/this-lens-is-soft-and-other-facts

Then we should all take a deep breath and think of every inspiring image we have seen in the last year, and remember that all of them were taken with some iteration of this 'flawed' technology :-)

- N.

Thanks!  It is nice to see this written by a guy whose business is camera & lens rentals--nothing like calling a spade a spade!