Phase One IQ180 - 80 Megapixel sample images

[Bernie]

The IQ backs can be exposed for as long as you want to. PhaseOne does not recommend exposures longer than 120 seconds and if you are picky you might want to stick to even less. The problem is image noise, which, of course, is always a temperature related issue. A sensor cooled down to absolute zero would not produce any noise at all. The backs do have passive cooling elements as well, but it's not enough if it comes to longer exposures or multiple exposures in a row. I've heard reports from photographers claiming the image quality worsens during an extensive photo session. Sinar backs once used Peltier elements (thermoelectric coolers) for cooling, but I don't know if they still do. Leaf puts a cooling fan into every back. Of course, one has to be clear that the backs are designed for photographers who produce a few frames a day on which they then spend even more time in post. And let's not forget that the files take up a lot of space. In my experience with the P45+ a 100 frames take up about 10 gigabytes of disk space (with the previews, CaptureOne files etc.).

[popo]

To answer Bob's earlier question on the DxO scores, remember DxO score the whole sensor, not the individual pixels. The overall scores being 91 vs 79, or 12 points ahead. Note that on DxO's score scale, 15 points is equivalent to a stop. Given it's based on the whole sensor, it has about 2.5x the area so should have over 15 points advantage. That it does not means, per sensor area, it is lagging the 5D2. Alternatively, it you simply scaled up the 5D2 sensor to the same size, it would score higher.

[MrKman]

curious if this applies to medium format: when I upgraded to a full-frame camera (5D mk1) my field of view increased from the 40D's cropped sensor. When I look through 50mm I see more--more is captured in the image on my 5D than my 40D. I understand why.

So my question is would a medium format give me an even greater field of view at, say, 50mm than on a full-frame sensor? In other words, is full frame really a crop sensor for medium format?

p.s. I'm interested in learning more about dynamic ranges different camera's have. Anyone got a helpful resource that shows that?

[popo]

I think the term "crop factor" was mainly used because digital SLRs used smaller sensors than the film models it replaced, so that had to be used to explain the differences. So while "full frame" sensors are smaller than medium format, I'm not sure anyone has ever made one by taking a medium format camera and shrinking it! At least, not in the digital era.

So if we take the existing reference as 36x24mm, then medium format cameras would have a "crop factor" of less than 1. Using a 50mm lens, assuming it has a big enough image circle, would provide a wider angle of view.

As for dynamic range, the earlier link to DxO scores also includes a component about dynamic range.

[Radu]

It’s all about the angle of view really. It is a function of focal length and sensor size.
The angle of view of the human eye (about 51 degrees) it’s called the normal angle of view
Any angle smaller than it is a telephoto angle and the ones wider than normal are called, well, wide.
A camera with a full frame format has a normal angle of view with a 50 mm lens. A smaller sensor, like the APS, decreases the angle of view and the same 50 mm lens becomes a telephoto lens. Similarly a large sensor like the medium one increases the angle of view and the 50 mm lens becomes a wide lens. For the medium format, if I recall well, the normal angle of view is with a 75 mm lens.

[Radu]

One more thing, in the case of the medium format we may name a "stretch factor" instead of the crop factor that is used for smaller formats than 24x36 mm. Taking into account the fact that the medium format frame is 54X54 mm, the stretch factor is 0.66 or about 0.7. So your 50 mm lens on a medium format is a 50X0.7= 35 mm full frame equivalent.

[Bernie]

The angle of view of the human eye (about 51 degrees) it’s called the normal angle of view.

Since I know how controversial this topic is I would like to point you to my sources at the bottom of my reply. I have been studying the technical aspects of photography and optics for more than 20 years and am not making anything i am going to write up. It is all thoroughly based on the current scientific view.

Human eye
Contrary to a lens made of glass our eye contains a flexible lens situated behind the iris. This results in a variable refractive power of 19 to 33 dpt for a healthy and young human. Additionally the light is refracted by the retina, the anterior chamber as well as the vitreous body. The whole system has a refractive power of 58 to 70 dpt, which equals a focal length of about 14 to 17 mm. Therefore the maximum field of view looking straight ahead is roughly 95°. In 35 mm film terms (full frame sensor) this would mean our eye has the field of view of a 20 mm lens (when we are young). Since the iris can vary the aperture between 6.5 and 2.5 mm our eye has a variable focal ratio of f/2.4 to f/6.4. This conclusion should already show you that a "normal lens" has nothing to do with the field of view of the human eye and as I am going to show the term has solely optical relevance.

Normal focal length
In photography, the term “standard or normal lens” is understood to mean a lens with a focal length about as long as the diagonal of the image field (field size). Its purpose is to simply describe the characteristics of the lens's focal length on the used format, which boils down to the corresponding field of view. A lens that produces a wide field of view (wide-angle lens) has a focal length significantly shorter than the diagonal of the film format. If it is about the same as the long side of the format, the lens is considered to be a moderate wide-angle lens. Super wide-angle lenses are those with focal lengths between the length of the short side of the format and half the diagonal. Those with even shorter focal lengths are often referred to as extreme wide-angle lenses, though the delineation between “super” and “extreme” is fluid, of course, and to some extent a matter of taste. Long, normal and short focal lengths simply describe the field of view of a lens on a specific format relative to the format's diagonal. Basically it is just a way for manufacturers to put their lenses into understandable categories. Of course, this implies that wide-angle lenses supposedly produce images that are perceived to be wide, but this is not necessarily the case. Human perception is quite a complex thing and even differs between genders. I would like to use following image to illustrate it a little more.


Le sentier de grande randonnée 10 (GR 10), Leica M9, 24 mm Summilux

Although it was taken with a wide-angle lens (24 mm on a full frame sensor) it is what I like to call a natural appearing landscape. I found the scene worthy a photograph at quite the same point of view and it represents pretty much what I saw. A normal lens would have cropped the image way too much to represent my view (the 24 mm lens was the only one I was carrying anyways). This is a topic I have talked to many fellow photographers about and many agree that wide-angle lenses can produce quite ordinary looking images. As photographers our job is to find a subject, a point of view, the best light and crop the image with a lens to come as close to what we are seeing with our "inner eye". Human perception can make an image that was taken with a long focal length look just as normal as one taken with an extremely wide lens. In studio work, for example, we often tend to use longer lenses to get more appealing images.

So why did the engineers decide to draw the line between wide and long right at focal lengths equal to the diagonal of the format? It has mainly to do with what is known as the normal viewing distance for a print, which is about the diagonal of the print. Sounds familiar? The reason is a human limitation: the closer you are to a subject the more your eye literally focuses on detail. So if you were standing too close to a print you would not be able to see it as a whole. If you were too far away the image would get lost in its surroundings. Also don't forget that photography is always about enlarging the original image by reducing the resolution and after further fiddling with it displaying the result (print, screen etc.).


So let's return to the technical aspects. If you are not into mathematics just skip to Misconceptions down below. The drawing above shows the light cone created by the lens and projected onto the film/sensor. Usually this projection has the shape of a circle and is known as the image circle. To cover the entire film the image circle's diameter has to be at least equal to the diagonal of the format (d). Traditionally this is defined by what is known as the angle of field (ω), which is exactly half the diagonal angle of view of the film format.

tan(ω) = d/2f
<=> ω = arctan(d/2f)

The diagonal field of view would have to be 2 * ω or 2 * arctan(d/2f).

Now what angle of view would a lens have to create to reflect the one from looking at the print from the normal viewing distance (diagonal of the print)? Through similarity of triangles we conclude that in this case f = d and therefore:

Normal field of view:
2 * arctan(d/2f) = 2 * arctan(d/2d) = 2 * arctan(1/2) = .93 rad = 53.13°

So this is where the approximately 50° for a normal lens, as Radu quoted, come from. Again, a normal or standard lens has nothing to do with representing the human field of view – it is merely a technical term for lenses where f = d. It is determined by what thinkers found to be the normal viewing distance centuries ago.

Misconceptions
1) A normal lens produces about the same perspective as the human eye.
Lenses do not determine the perspective. Only the position of the lens's nodal point relative to the photographed subject as well as a potential swing/tilt of the film plane affect the perspective. The closer you move the camera to the subject, the extremer your perspective. Vice versa, the farther away from your subject, the more you'll find the perspective to be compressed. Photographed from the same point of view a normal lens has exactly the same perspective as an ultra wide angle lens.

2) A normal lens has the same magnification as the human eye.
Magnification is solely determined by the size you print your image or display it on your computer screen and your viewing distance. Consider this example: You take a picture of a coin with a magnification of 1:1 (1 mm on the coin is projected as 1 mm onto the film). The image is then printed 100 metres wide. If you were standing right next to the print you would see a much larger magnification than 1:1, but from three miles away it would be quite the opposite. If you are arguing that the magnification of a normal lens is about the same at hyperfocal distance (far away objects) I'd urge you to take a picture with your standard lens and look at it from a normal viewing distance. As mentioned above our eye is more of a "wide angle zoom" – especially at hyperfocal distance.

3) Standard lenses are the easiest ones to design.
To some degree this is actually true. Moderately longer lenses often involve less complex lens designs and extreme wide angles are the greatest challenges for optical engineers.

Personal beliefs
Nevertheless many renowned photographers claim that normal lenses are the ones that come the closest to reproducing natural looking images in terms of magnification and perspective. One could argue for years over this extremely controversial topic. Personally I'd say there is some truth to it since it is all about perception anyways. What I tried to clarify simply is the technical aspect, although some might have a different opinion on that as well.

Sources
Applied Photographic Optics, Sidney F. Ray, http://www.amazon.com/dp/0240515404
A History of the Photographic Lens, Rudolf Kingslake, http://www.amazon.com/dp/0124086403
Lens Design Fundamentals, Rudolf Kingslake, http://www.amazon.com/dp/012374301X
Distagon, Hologon, Biogon, H. H. Nasse, http://blogs.zeiss.com/photo/en[...]Distagon.pdf
Experimentalphysik 2: Elektrizität und Optik, Wolfgang Demtröder, http://www.amazon.com/dp/3540682104
Grundlagen der Optik in der Fotografie, Jost J. Marchesi, http://www.amazon.co.uk/dp/3933131626

[Bob Andersson]

Hi Bernie,

Thanks for taking the time to compose your post above - an interesting read. A question if I may, and my apologies if the answer is in one of your sources...

While I know that the FoV of the eye is considerably wider than that offered by a 50mm focal length lens in front of a full-frame sensor (20mm from your post) I had always assumed that the idea of 50mm corresponding to what we generally consider to be a normal (in the non-technical sense of the word) FoV was due to an implicit weighting by the brain when we view a scene because of the higher density of rods and cones near the eye's optical axis. Any thoughts on that proposition?

Bob.

[Radu]

Thanks Bernie for making the issue clearer. I recall my instructor telling me some 35 years ago about the normal lens and it’s relation to human vision. Thanks for clearing this misconception.

[Bernie]

Bob, you are correct about the brain paying more attention to areas closer to the image centre. It does not limit our FoV though. I edited my previous post to elaborate further. Always think of it this way: Normally a lens behaves abnormal. It is the way we perceive the image it creates that makes it normal. Technically it is the simple statement focal length = format diagonal. Why? Because that's when it reflects the angle of view of the print from the normal viewing distance, which was determined more or less randomly.

[MikeI]

By Bernie:
Normal field of view:
2 * arctan(d/2f) = 2 * arctan(d/2d) = 2 * arctan(1/2) = .93 rad = 53.13°

Bernie opened up the "normal FOV topic" by indicating how controversial it is. Not enough controversy?!

The ascending angle of the Great Pyramid's Great Gallery is arctan(1/2)=26.5650..°
This angle comes up during the construction of the Golden Proportion.

Cheers,

http://www.hyperflight.com/golden_numbe ... ct_Great_P

[jouster]

I recently saw exhibited some of Clifford Ross's large format work, and I can tell you that no single image digital camera has anything like its resolution. Sure, he uses very large format film, but the claim earlier in the thread was that digital has surpassed all film in terms of resolution. This is simply not true.

[Bernie]

Since it was I who made the statement I guess I'll have to clearify it a little. When I wrote "digital has long surpassed the quality of film" I should have been a little more precise. What did I mean by quality or what does quality in photography stand for?

Print quality
I would like to quote George DeWolfe here: A fine print comprises aesthetic (artistic) qualities that can be controlled by technical means. Printing is not just a technical exercise or an aesthetic one; it is a combination ruled by both in harmony. We control six major aesthetic qualities when we produce a print: cropping, contrast, brightness, color, defects and sharpness. Although we can control those qualities in post production it takes a good file or negative to begin with.

Quality of the capture medium
This is merely technical and has no aesthetic aspects to it. The capture medium should record the image produced by the optical system (lens) as accurately as possible. This involves detail, color and dynamic range/contrast.

When I made my statement I was writing about the capture medium (the PhaseOne back), which I considered to be obvious. This has nothing to do with print quality, which also involves perception. For further reading I would recommend Markus Zuber's comparison of the IQ180 with 8x10 inch film, which I consider to be the best "shootout" so far and concludes that the IQ180 out-resolves 8x10 film. Again, this does not mean the prints are better.

[Gordon Laing]

Thanks for all your detailed posts Bernie, all very interesting stuff!

PS - MrCliff it depends which film and which digital sensor you're talking about, but if it's 35mm film, most frames only contain about 12-15 Megapixels worth of real-life detail, which means most DSLRs or ILCs have been roughly matching or out-resolving 35mm film for some time. And remember this is under ideal conditions. If you have average film, less than perfect processing and an unremarkable scanner, 35mm could dip below 10 Megapixels worth.

[cybermystic]

Now if only they would Peltier cool that beast I would have the ultimate deep-sky imager
Greg