Hasselblad X1D: Contining Impressions and Learning Curve

[Bjørn Rørslett]

A different approach that is carefully constructed to take all factors into account might or might not support the current hypotheses ('theories'). Just repeating something possible wrong won't make the outcome more correct. We need urgently complete independence in other set(s) of data to advance our understanding of these relationships.

[Ilkka Nissilä]

Why on earth would a camera sensor have better dynamic range in Dx format than in Fx format?

(Sorry! I misread your post. Edited.)

The D500 sensor has more dynamic range at low ISO than the D5 likely because of different sensor architectures; sports photographers and journalists likely prefer a camera optimized heavily for high ISO and many wildlife photographers would use the D500 and shoot in relatively bright light. So Nikon (or whoever they work with, Sony etc.) designed each sensor so that it is optimal for the target applications.

The same sensor in DX crop mode does have less dynamic range than in FX, at equal ISOs, unless there is an error in the measurement or presentation. This is because each pixel in the DX image has to be magnified by 1.5x more than each pixel in an FX image to reach the same print size. This magnification makes the noise more apparent.

[Bjørn Rørslett]

Just my point. The "difference" occurs after the data has been captured and is due to processing not true difference in sensor performance.

Ie, just a very convoluted way of stating the obvious, that a larger format needs less overall magnification of detail to reach a fixed print size. A fact that has been known since the dawn of photography. One allegedly blame the sensor performance instead of pointing to the underlying assumptions responsible for the claimed "difference". Take away the fixed print size and what remains of value in the model output??

[Ilkka Nissilä]

Comparisons between algorithms, instrumentation, technologies or products usually involve each contestant performing a fixed task. I don't think any methods or instrument focused scientific journal would publish a comparison where the tasks performed by two methods are different, or evaluated using criteria which is not the same for each method.

You can use several different tasks and metrics if you want to gain a more in depth understanding of the advantages or disadvantages of each method. For example a big lawn mover might finish an acre faster than a smaller one, but the smaller one is more dexterious and can reach places where the larger one can not. So the speed of covering a large field might not be the most important metric. Still it seems not very scientific to present the following comparison result:

Lawn mover 1  finishes a town in a fortnight
Lawn mover 2  tweaks a home lawn in an afternoon.

It is difficult to grasp these results as to what exactly the relative performance was, since the same task or metric were not used and they were not precisely defined (even though the above speed evaluations might appeal to some for their practical value to a common person, they would not be considered scientific).

As long as the methods are described clearly, and the experiments are correctly carried out and presented, one can find value in the numerical data that is presented. Of course, a truly comprehensive analysis of a camera's performance needs to consider lens availability, focusing, etc. all factors which contribute to the experience of the users. But there is a lot of fixation on the numerical data these days. This is not a good or bad thing in itself, it is just important to remember that this is only a part of the story.

Personally I value greatly the viewfinder quality of the cameras that I use. Yes, it is subjective, but it has an important role in my experience of the photography process and also indirectly influences the outcome (and sometimes directly as well as the information I get through the viewfinder is different and that influences my choice of when to take the shot). And ultimately my personal experience is what drives me to do photography, not some numerical performance criteria. That doesn't mean the numerical criteria isn't of some value. I enjoy the D810's super silky smooth tones at ISO 64 and the quantitative sensor performance evaluations give some rigorous validation of the subjective findings. However, when I'm photographing people, I usually use the D5 as it gives me a very high rate of focus keepers and so I don't have to lose shots to slight misfocus as much as I used to and above all I don't have to spend as much time editing. In this situation I choose the metric which is more important to me. I think the cameras have been fairly described in reviews - there is discussion of the dynamic range, and autofocus capabilities. For me the D8x0 always put me in an uncomfortable place focus-wise; the D810 is an improvement over the D800 in that respect (and many others) but the D5 really kicks the (autofocus) ball right out of the park.

Although ergonomics, viewfinder and AF are important criteria for me, I still always check the numerical sensor data as well, as they give me important information on what I can expect image quality wise. I have not been disappointed by the much-criticized D5 base ISO dynamic range because I knew to expect it and I've been very pleased with the results from ISO 1000 to 25600. Without the measurements I would probably try to try to shoot at low ISO and wonder what's going on when the image looks a bit rough. In my figure skating photography with the D810 (and D800 before that) I was only happy up to ISO 1000 and even ISO 2000 seemed a bit noisy especially if I had to crop. With the D5 I can shoot at ISO 6400 with impunity and while it obviously isn't quite as good as D810 at ISO 1000, it gives me additional options as to what kind of lens I can use for this application, such as the 300/4. I've shot wedding dance at ISO 8000-11000 and been very happy with the image quality of the prints, something that would have been totally outside of the domain of the D800 or D810 in my view. So by checking the quantitative sensor data I was able to start on the right foot and use each camera for where it excels rather than go on a rampage about the D5's base ISO image quality. Which is still fine, not the best by today's standards, but at least the shots are in focus. I was testing the D5 and 300 PF at Helsinki City Run on Saturday to get some practice and to get a feel for whether it would be better to use the D810 at this week's two outdoor fashion shoots for its low ISO DR.  If I shoot with the 300/4 PF the D810 nails the focus most of the time but not so much with the 200/2 - in backlight the camera is likely to pick on the brightly lit, high contrast detail in the hair which is several cm behind the eyes and so shots are likely to be backfocused in this scenario (I actually trick the system by subtracting 1 or 2 points from the normal focus fine tune value to increase my odds).  The D5 has enough AF sensitivity and specificity to focus on the facial details without being lured by the brighter, contrasty hair most of the time. And I think the DR will be ok as long as I don't make a gross misjudgment of my exposure.

(Sorry, I always seem to slip onto the topic of AF in any sufficiently long post.)

[David H. Hartman]

Take away the fixed print size and what remains of value in the model output??

A smaller print?

[bclaff]

...

In the referred charts, modelled output based on unspecified or implied conditions deemed relevant to the average photographer

The output is not modeled it is measured from raw test images.
Nothing is unspecified it is all documented on the site.
Assumptions are made (and stated) regarding visual acuity and reference is made to well established (eg. ISO) levels of acceptable SNR.
I think that's relevant to the average photographer.
FWIW, I think this test is better designed to provide relevant information than the "equivalent" DxOMark Landscape (print) score; and feedback from photographers indicates that many agree.

is presented as were it to show the DX sensor is inherently inferior to the FX. Or by extension, any smaller vs. larger format sensor. ...

There is no intent to show anything. The measurements speak for themselves.
You could interpret it that way but like anything it needs to be put into context.

That said, the results are unsurprising.
Since this is a measure normalized to a fixed print size and viewing distance; the smaller sensor area must be enlarged more and suffers more loss in quality.
Surely no one expects a small crop from a full image to retain the same image quality when both are enlarged to the same size; this is no different.

[bclaff]

As far as I know the numbers come from DxOMark and the analysis is by Bill Claff himself. Both have been described in more or less detail and discussed extensively. What else does "science" want ? Does it want someone to independently repeat DxOMark's measurements ? Who is going to pay for it ?

To clarify ... everything in the first section of my site, for example Photographic Dynamic Range, are from my own independent measurements.

The second section comprises measurements that are derived from DxOMark data.
This is provided largely because they have tested more cameras.
Also as a resource for those who had become accustomed to getting data from sensorgen (which still exists but is not longer maintained).
And finally, as a cross-check.
So in the second section you would find DxOMark Derived Photographic Dynamic Range
Same PDR methodology using DxOMark data rather than my own.

Although I don't repeat DxOMark's methodology I do think a valuable cross-check already exists in PhotonsToPhotos.

[JohnMM]

To clarify ... everything in the first section of my site, for example Photographic Dynamic Range, are from my own independent measurements.

The second section comprises measurements that are derived from DxOMark data.
This is provided largely because they have tested more cameras.
Also as a resource for those who had become accustomed to getting data from sensorgen (which still exists but is not longer maintained).
And finally, as a cross-check.
So in the second section you would find DxOMark Derived Photographic Dynamic Range
Same PDR methodology using DxOMark data rather than my own.

Although I don't repeat DxOMark's methodology I do think a valuable cross-check already exists in PhotonsToPhotos.

Thank you.

[Bjørn Rørslett]

We agree on the effect of secondary magnification and that'll suffice for me. Again, just bringing the focus (sic) back to what was known centuries ago.

I can attest to the fact that in practical field situations, the 'theory' plays no role for the final image.  We do not make the "same" images with different gear, which is a nonsensical approach, we use the best tools for the purpose or even whatever tool is at disposal.

[David H. Hartman]

In the days of film I never owned a half frame camera. I would still have used Tri-X as my primary film. I would still have made 8x10" prints just like with a full frame 35mm camera. I would not have liked the increase in grain form the extra enlargement required. I would not been able to crop images and accept the grain. The loss of image quality would have been too great for me.

Today my primary format is 36x24mm (FX, FF). I would make most use of a D5 if I could afford one. I don't believe in equivalence. If you need a medium format camera you can't make a DX or FX camera do medium format. I would find 24x16mm (DX, APS-C) as available in the D500 very useful. I can but dream of a medium format digital camera.

I hope Michael continues to enjoy his Hasselblad X1D and finds the lenses that facilitate his photography.

Dave

[Jack Dahlgren]

I agree with David. Good to hear Michael is using his gear for making images.

[Ilkka Nissilä]

I can attest to the fact that in practical field situations, the 'theory' plays no role for the final image.  We do not make the "same" images with different gear, which is a nonsensical approach, we use the best tools for the purpose or even whatever tool is at disposal.

Right but you have decades of experience and use that to decide what gear to use.

Someone with no experience can use theory to get an idea of what can be expected using a particular setup in a particular situation. This is helpful to get started. An author doesn't usually refer to grammar when writing a novel or a poem but nevertheless grammar was taught at school and this education likely contributed positively to their writing skills.

[Ilkka Nissilä]

I don't believe in equivalence. If you need a medium format camera you can't make a DX or FX camera do medium format.

It's not a question of a religious belief but the theory comes from basic physics and geometry and has been experimentally validated.

Under diffraction limited conditions you can get the same image using a medium format or small format camera, to a good approximation. I will be happy to demonstrate this some time (with FX and micro four thirds, the formats which are available to me) when I have time to set up the experiment.

[Michael Erlewine]

SELLING MY X1D

After shooting some 1100 or so shots with the Hasselblad X1D system, I have decided that it is not what I need for my work. I’m sure, many will point out that I don’t get it, but I am only trying to “get it” for my own work. It’s embarrassing to admit this, after all my praise, but at my age, who cares? I have put my X1D system, including the 45mm and 90mm lenses, plus the lovely RSS L-Bracket and 5 batteries for sale on Ebay under my nickname ALLMUSIC.

Although there are a number of druthers and small reasons for giving up the camera, the two main reasons are:

(1) The lack of lenses I need now for the camera. I have waited months for the 30mm and, of course, probably will would have waited for the announced 120mm Macro, as well. This is summer, and now is the time I need those lenses. As a close-up photographer, neither the 45mm or 90mm can get me close enough, especially since there are no extensions available.

(2) And secondly, having very carefully done hundreds of test shots for overall sharpness, I am, despite what others say, disappointed with the degree and kind of sharpness I can attain with the X1D system and their lenses. Please don’t ask me to prove this. It’s just my opinion.

No one is more sorry than I am, and perhaps selling this system is a stupid thing to do. I lose a bunch of money. Or, is it that I’m hooked on the Nikon system and how the D810 works? I am not arguing that the D810 IQ is better than the X1D, but only that what I am able to get from the X1D is not worth what I have to put up with to get it. And, of course, there are all the great lenses I have that will never work on the X1D, but that is a minor thing.

And finally, for me and the work I do, the X1D is just not ready with what I need. I should have waited for perhaps the second edition. And, I can always get another copy, should an X2D comes out. Meanwhile, I will wait for the rumored 46 Mpx D820 and have to be happy with that. I have TRIED to love this system, but I can’t get there from here.

I will say that the X1D haptics are extraordinary and that, if I wanted to afford what for me would basically be a $20k system, I would keep it. The system is easy to use and I love the touch-screen of the LiveView.

So, there you have it. I report this because I owe it to those who have read any of my other comments on this system. C'est la vie

[Bjørn Rørslett]

A very fair and succinct summing-up of your decision and the underlying rationale. Kudos.

This also confirms my conviction that photographers basically are more interested in what they can achieve rather than any numeric specification of the same, however enticing the latter might be on paper.