I wouldn’t comment on product announcement unless the product is exceptionally interesting, or as the teaser said, a game changer. However, those words have been (ab)used before, in particular in the mirrorless* world, and have eventually left something to be desired.
For everyone who missed it: Hasselblad have announced a 50 MP medium format** (44×33 mm) mirrorless camera with a 2.36 MP electronic viewfinder (EVF). There also will be a new lens line with leaf shutters, for a large range of shutter speeds (60 minutes to 1/2000 seconds) and with full flash synchronization to boot. Rumors had it for Sony and Fuji for a while and I would not be surprised to see a similar offering by Pentax; possibly at Photokina in September (my flight is booked).
*They really should come up with a new name for those cameras, on the same footing that we are not driving horseless coaches.
**Well, not quite to medium-format film standards ranging from 60×45 to 60×90 and even 60×170 panoramic.
The Hasselblad X1D is a handsome, relatively light (725 g) camera with the option (hopefully) for mounting it on a view camera and to use other medium format, leaf-shutter lenses with an adapter. Effectively, we are looking at a H6D-50 in a Mamiya 7 gestalt, the body being a bit smaller while the lenses are a bit bigger. Although the weight of the camera body plays a role for the landscape shooter, it’s the entire system that counts; body, lenses, l-bracket, batteries, chargers, flash-unit, and tripod. The camera is light enough though for a 1-series Gitzo tripod, or equivalent, which I would not use for a Pentax 645z or a Hasselblad H6D.
And to get this out of the way: The camera will retail at around 7900 Euros (8900 USD) and the 45 mm lens will be 1900 Euros; not particularly cheap but reasonable if you consider the price for the naked sensor. And btw, who would always complain about a Porsche not being particularly cheap. Remains the question for how long the image quality of this camera will be state of the art. Moreover, I wouldn’t go to Antarctica or ever shoot another wedding without a second body (I had a shutter failure on a Nikon F3 exactly when they were exchanging rings). And carrying two incompatible systems isn’t an option either.
For me, the reason for getting excited about this announcement by Hasselblad is the setup shown in the image, which yields a second live for my beloved Linhof Technika. Equipped with a Fotodiox adapter for the Nikon F mount, it lets me view the image through the viewfinder (with diopter adjustment), allows full movement (shift/rise/tilt/swing) and is surprisingly easy the shoot (at least for everyone who has experience with view-cameras). Just set the digital camera to aperture priority, close down the aperture at the large-format lens and use the shutter delay option on the camera. The adapter also allows stitching of up to six images by shifting the camera position at the back and thus avoiding stitching artefacts in foreground/background compositions.
Btw, this setup also allowed me to check if the large-format lenses can keep up with the sensor resolution of the Nikon D810. And yes, they can, albeit with a relatively small shooting envelope because of their maximum apertures of f/5.6 and f/8.
However, and this is an important point, the large flange-focal distance of the F mount, together with the thickness of the camera standards and bellows, makes it impossible to focus lenses shorter than 150 mm at infinity.
I was therefore seriously tempted after the price drop for the Hasselblad CVF-50c digital back. But a back is a back is back, while a larger sensor in a thin body has the potential to combine the best of both worlds. The CVF-50c can be used on the Hasselblad V-system cameras, but let’s face it: The handling of these cameras (in particular for rectangular formats) is slightly awkward and not really up to the digital age.
A comparison of the relative sensor size and pixel pitch should give us a hint of the image quality. One may argue that 51 MP is not much more than 36 MP. But the combination of larger sensels and more sensor real estate add up to an appreciable advantage over 36 MP. The sensor is about 1.7 times larger than full frame. Moreover, the native image aspect ratio is 4:3 rather than the wider 3:2 ratio of both APS-C and the full frame 35 mm. I have always found 3:2 to be either too wide, or not wide enough. This is, of course, a matter of personal preference, but I often find myself cropping 35 mm and thus throwing away 10 MP or so. This gives an even greater advantage to the X1D over the Nikon D810.
However, there is no such thing as a medium-format look; on the same footing that there is no large-format look. Using a lens with the same angle of view for the film/sensor size, and the same camera to subject distance, there is absolutely no difference in the perspective. Of course there is a difference in the depth of field; an f/2.8, 90 mm on (real) medium format 6×7 cm (horizontal angle of view 39 deg.) yields a depth of field 17 cm at a subject distance of two meters, comparable to an f/1.8, 50 mm on the 36 mm, full-frame sensor.
Any look has to do with lens design and optical corrections. Therefore, the rendering of the new lenses and whether the camera is able to focus accurately will need to be seen. And it is bizarre: In the large-format days we struggled to get more depth of field rather than less, using Scheimpflug and apertures at around f/22. Resolution wise, film was the limit, not the lens aberrations and diffraction. But with the small digital sensors, everybody gets excited about bokeh, and this in turn has resulted in monstrous, f/1.4 lenses mounted on tiny camera bodies.
But image quality is not only about resolution. Even more important are the dynamic range, color fidelity, and tonal separation. Here the claim is exciting: 14 apertures dynamic range and a genuine 16-bit workflow.
Consider the Nikon D810 with its 35.9 x 24.0 mm sensor, 36.1 MP, and a pixel area of 23 μm². For the 51.3 MP sensor of the X1D we arrive at a pixel area of 28 μm², which is about 17% larger. But the Nikon D5 has a pixel area that is again 48% larger and allows an ISO range up to 102400 (with respect to 25600 for the Hasselblad). As the ISO range indicates, roughly, how much you can amplify the low signals without creating nothing but noisy mush, it is an indication for the signal-to-noise ratio that can be obtained for a particular pixel size and for a given sensor architecture. So why then, does Nikon*** not offer a 16-bit workflow for the D5?
Processing speed may be the key, but they could offer 16-bit processing as a setting when high frame rates are not required. As the quantisation of an analog signal introduces an error by rounding off to a discrete value in the analog-to-digital (AD) converter, it would seem obvious that the highest possible bit depth yield the smoothest tonal gradients****. However, noise amounts to random jumps of tonality between neighboring pixels. If these jumps are larger than the discrete intervals, banding will be rendered invisible. This smoothing of transitions is known as dithering. Discretizing the signal from the sensor in much finer steps than the level of the noise floor is therefore meaningless. And if you require the same depth of field for equal exposure, you must double the ISO gain for the medium-format camera, and consequently one stop of input (at the highlights) is pushed past the saturation point. So maybe, the claim of 16 bits is nothing but a unique selling proposition. The sample images available on the web are useless; what would be required are 16-bit tiff files of head portraits (skin tones), desert landscapes (tonal separation), or the earth shadow (gradients).
Mirrorless cameras are also very power hungry, which can quickly turn into a situation. Two years ago I walked the Annapurna circuit, days after the snowstorm disaster on the Thorong La. An avalanche had destroyed the power line in the Marsyangdi river valley and so we were without electricity for more than 10 days. With two spare batteries I managed comfortably, limiting live view to the minimum and refraining from chimping. Try this with a mirrorless.
I do like the design, apparently, what looks like metal is metal, but it needs to be seen if they didn’t go too far: where is the focus select button? Focus and recompose will just not be good enough for this sensor resolution and using the touch screen will be impractical for hand-held shooting.
There are a few other facts from the catalogue that make me curious: 3.0 inch 920K pixel display (Nikon D5: 3.2 inch, 2359K), exposure metering spot, center and center-spot (only), 2.36 MP electronic viewfinder (Leica SL: 4.4 MP) and no mention about the autofocus system (Sony Alpha 7rII: 399-point phase detection plus 25-point contrast detection). Unlike the Leica SL, the X1D has no general purpose, electronic first curtain shutter (EFC). Leaf shutters are not vibration free, however, and there is a lack of any speed higher than 1/2000 second, while properly implemented EFC offers a range of up to 1/16000 second.
When connecting the X1D to a view camera, a method is thus required to physically trigger the lens shutter and the sensor unit simultaneously. For the digital backs this can be done via a flash-sync connection from the copal-type lens to the sync port on the back. I hope they kept this functionality, because otherwise, it will be a game killer, not a game changer.
Update 28.08.2016: There will be definitely no sync port on that camera. But the rumor mill has it that Fuji will be on the market with an announcement at Photokina.
***And neither do Pentax (645Z) and PhaseOne (IQ250), which employ the same 51 MP Sony sensors.
****You wouldn’t be able to tell the difference on a computer screen anyway (unless you do serious cooking in Photoshop where banding and tonal overlaps may result), because they are currently limited to 10 bits. Even then, a graphics board and software that support 10-bit output would be necessary.