The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Ricoh GX100 : Architectural photography

A

Arch

Guest
Hello everybody, and thanks for a serious forum and all the interesting threads. I thought I could share the enthusiasm of my initial experiences with Ricoh GX100 for serious architectural documentary photography. With serious, I mean straight vertical perspective, zero barrel distortion and sharpness in details, speaking merely in technical terms. Since 1992, I’ve been using a 4x5” plate film Cambo Wide with a Schneider Super Angulon 1:5.6/65 mm. A fantastic tool, to put it mildly. But you can’t put it in your pocket.
For a more everyday use, I wanted to find a decent pocket camera with a similarly wide angle high quality lens, and there was no other alternative than the Ricoh GX100. It’s an impressive little camera, and works beautifully even for serious architectural photography, with a little help from some basic software. Small sensor noise is not an issue with this type of shooting, which is low ISO, anyway. And small sensor huge depth-of-field is pretty handy, indeed.
My example shows two images. The starting point is a very casual, underexposed, hand held shot on a dark and depressive day, with a seriously tilted perspective, in order to fit the whole building in the frame. This dng file was first processed in Lightroom with basic tonal adjustments. From Lightroom, a quick external (TIFF) edit in LensFix CI (for mac, =PTLens for PC), to get rid of the barrel distortion. This inexpensive software has specific presets for the GX100, and it does the job in a flash, automatically. After (and only after) LensFix, another external edit in Photoshop (perspective crop), and final adjustments in Lightroom (which is the unnecessary extra here).
With this test, I also used the wide angle converter. It’s a quality product, but it shows minor softness at the other(!) end of the frame. The 24mm equiv. wide end of the zoom is wide enough for most architectural work, but the 19mm option is a treat. The pocketable 24mm, by the way, makes the GX100 an easy choice over the GRDII with its 28mm (/21mm), or any other 28mm quality pocket camera, for that matter. Even though the barrel distortion with the GRDII is minimal, it too has to be corrected with post-processing for serious results. The ease of using LensFix really puts aside the whole issue of lens distortion, and makes a prime lens lover reconsider the attraction of the zoom.
This simple and quick test makes me re-evaluate camera technology and ways of taking pictures. Let’s just hope that in future cameras lens distortion correction happens in-camera, without you noticing a thing. And let’s also hope that Lightroom adopts the invaluable perspective cropping from PS.
 
W

wbrandsma

Guest
Welcome on GetDPI and thank you for sharing your thoughts and images.
I have personally never worried about lens distortion and actually used it for compositional reasons.
For your second image the distortion correction may correct the appearances of the buildings, but I think it makes the cars in front look a bit funny (like 4:3 videofootage on a 16:9 TV). Did you deliberately underexpose this image? For instance to avoid highlight clipping?
I have used the GX100 for architectural photography as well sometimes, but probably never as serious as you do.





 
A

Arch

Guest
Thanks for your comments, WB.
I'd like to clarify the difference between lens distortion and perspective distortion. Lens distortion is something that all optical engineers want to avoid, and they use their utmost ingenuity and sleepless nights for it. The finest lenses have zero barrel & pincushion distortion, and (documentary) architectural photography takes this as a minimum requirement.
Perspective distortion is not a technical fault, and it can be used for expressive purposes, as you have done in many of your fine shots. Yet, most architectural photography is based on vertical verticals. This is not only a long tradition, but has to do with the way we experience spaces and volumes on a two-dimensional surface. It also has to do with being humble in front of the subject, letting the architecture speak instead of the photo itself. This makes me ask if the building in your third photo is a cubical box or something else. I would guess it's cubical, but the photo sort of misleads you to think it's a cut pyramid.
I'm tempted to refer to some of the finest of architectural photography, the work by Bernd and Hilla Becher. Their ultimately "dry" objectivity is pure documentary photography (not even sunlight to take your attention from the subject), yet their work is shown in the finest art galleries and museums in the world. Their art is in choosing the subjects and revealing beauty through a typology. Which breaks your heart, of course.
I forgot to mention that my initial quick test with GX100 included a horizontal scale transformation after the perspective crop. This is necessary to adjust the changed proportions that the crop gives. For a more serious approach, you need a reference photo taken at the same spot, but levelled horizontally. This can be used as a check layer in PS. For my eye, the cars look funnier in my original dng file. Things in the wide angle corners always look funny, anyway. Their was no intention to underexpose the shot, it was a true point-and-shoot situation, full exposure, 1/20th sec. Using the "bad" dng was part of my test.
The seriousness with the combination GX100/LensFix/PS perspective crop reaches all the way up to a full magazine page, and it will be hard to believe that the images are not made with plate film, shift and tilt, and prime lenses. For door size prints in exhibitions you still need plate film. But let's wait for five years or so.
 
W

wbrandsma

Guest
I fully agree with you Arch and thank you for correcting me. I also meant perspective distortion, but unfortunatelly wrote lens distortion.
You are right with the building in the third image. It is cubical box. I delibaretly shot this image with this skrewed perspective. The building itself is misleading as well. When you first see this building the concrete structure looks like a useless floating shell. But instead is a fundamental structure for this building (by the way designed by Rafael Vinoly Architects).
I tried to correct the perspective distortion for the same image. That way you can get a real sense of the building :thumbup:.
 
A

Arch

Guest
What a difference! There's a whole new peace within the image, it catches you for a thoughtful moment. The black&white conversion suits the subject perfectly, since color plays such a minor role in this very piece of architecture. Also, the darkened sky focuses the attention to the elemental aspects of the white diagonal grid, and the straightened perspective lets your mind travel through the structural forces. Thanks for going through the trouble. I'm sure that Rafael Vinoly, too, would be pleased.
 
H

hiro

Guest
When depicting a 3D world on a flat surface, there will always be a tradeoff between preserving straight lines, shapes, and areas. Except when shooting something straight-on, you will never get more than 1½ out of the 3. Lens makers usually try to preserve straight lines by going for rectilinear perspective, but this introduces distortions of both shape and area, as can be seen in the first example shots above where the square window panes of the LHS building become distended parallelograms whose original shape—square or oblong?—is no longer discernible. At the other extreme we have fisheye perspective where area is preserved but straight lines become curved, and shape is again deformed. Half-way between the two, shapes are more or less preserved but at the expense of the other two.

The situation is precisely analagous with map projections: with the mercator projection we have a rectilinear world where you can plot a straight course but countries at the margins such as Greenland become hideously deformed. If OTOH you choose an equal area projection, all the straight lines of latitude and longitude become curved, and countries at the edges become elongated. Somewhere between the two you can more or less keep the country shapes correct but with a modest loss of both area and straight lines.

In the end it is horses for courses, there is no one correct map projection nor perspective, it depends on your requirements. For myself, I welcome the modest barrel distortion of the GX100 since it reduces the distortions of area and shape at the edges of the picture that a true rectilinear lens would produce. However I am not usually photographing straight lines, i.e. architecture.
 
S

Sean_Reid

Guest
When depicting a 3D world on a flat surface, there will always be a tradeoff between preserving straight lines, shapes, and areas. Except when shooting something straight-on, you will never get more than 1½ out of the 3. Lens makers usually try to preserve straight lines by going for rectilinear perspective, but this introduces distortions of both shape and area, as can be seen in the first example shots above where the square window panes of the LHS building become distended parallelograms whose original shape—square or oblong?—is no longer discernible. At the other extreme we have fisheye perspective where area is preserved but straight lines become curved, and shape is again deformed. Half-way between the two, shapes are more or less preserved but at the expense of the other two.

The situation is precisely analagous with map projections: with the mercator projection we have a rectilinear world where you can plot a straight course but countries at the margins such as Greenland become hideously deformed. If OTOH you choose an equal area projection, all the straight lines of latitude and longitude become curved, and countries at the edges become elongated. Somewhere between the two you can more or less keep the country shapes correct but with a modest loss of both area and straight lines.

In the end it is horses for courses, there is no one correct map projection nor perspective, it depends on your requirements. For myself, I welcome the modest barrel distortion of the GX100 since it reduces the distortions of area and shape at the edges of the picture that a true rectilinear lens would produce. However I am not usually photographing straight lines, i.e. architecture.
I strongly agree with what you say here and have been writing about just this topic in several articles over the past couple of years. A lens that gives a perfectly rectilinear rendering will, as you point out, "stretch" the appearance of objects close to the camera. That effect, while it is mathematically correct, in a sense, can be as or more disturbing than the curvature of straight lines.

For a lot of work, I tend to prefer wide-angle lenses that have just a bit of barrel distortion, especially when I don't need to be concerned with strong verticals at the right and left edges of my pictures. I do shoot architecture professionally, though, and the lenses I use for that usually have to show very little barrel distortion. Thanks for an interesting thread all.

Cheers,
 
A

Arch

Guest
This thread is becoming interestingly theoretical. Hiro pointed out a nice analogy between photography and cartographical projections, which takes us further. The image captured by the camera is, after all, circular like that of our eyes, but cropped in order to fit the generally approved, limited format, which is always a compromise.
Architectural (and again I want to use the word documentary) photography, for its part, has an analogy, and parallel history, with architectural perspective drawing. These drawings are made with straight rulers, and since the 1980's or so, more and more with straight lines generated by cad software. They have always been one step further from projection drawings: elevations and sections. Interestingly, cad perspective drawings - or renderings - have become misleadingly photorealistic, and the tools for generating these within cad software are now called "cameras", which imitate photography in various aspects, such as lens angle, aperture, dof etc. Probably, in some applications, you can add barrel distortion, if you wish.
Yet, I have doubts that anyone would wish to do so. The subjects in cad rendering or architectural photography are most often based on rectilinearity. Frank Gehry and his league are exceptions of the rule. Vertical lines tend to appear next to the edges of the frame, and unintentional line curvature really hurts the architect's eye. And architects are usually clients for architectural photography...
The main content of any architectural photo is the building or the space. It doesn't really matter if cars or objects like that in the foreground corners become anamorphously twisted. But there is another kind of architectural photography, where distortions can even help to create a more "natural" vision of the world. Mitch Alland pointed out that Stuart Richardson's first subway shot in the "Conclusions on GRD2 vs GX100" thread (page 3) is "a good example of using a camera's faults for aesthetic purposes". Note, however, that in Stuart's second example on page 4, barrel distortion seems to be corrected, for obvious reasons. Really fine shots, by the way. Sean Reid, as you can read in this thread, seems to make a sophisticated choice between lens types according to the subject's needs.
But back to our tool, the lovely GX100. Later this winter, I'll try to make some serious tests to push my Ricoh's performance. I'll fix, crop and print magazine full-page size images, and compare them with those taken with my 4x5" Cambo Wide/Super Angulon. Not pixel-peep really, just look. I expect to surprise myself.
 
S

Sean_Reid

Guest
Hi Arch,

Another interesting post from you... I find that I sometimes have more flexibility to use lenses with some barrel distortion (less "EOS" <G> - edge object stretching) when I'm shooting certain interiors where strong vertical don't fall in the outer zones of the picture. In fact, one of my clients prefer that pictures made for him be made with lenses that stretch the edge objects a bit less.

But for most exteriors and interiors, you're, of course, absolutely right - zero barrel distortion is what we hope for.

If the opportunity presents itself, see if you can include a GR2 with GW-1 adapter into your informal comparison.

Cheers,

Sean
 
W

wbrandsma

Guest
Thank you Arch for your interesting post. I am very interested in your testresults later this winter with the GX100.
I think architectural photography is a very interesting subject and your post made me curious. I am still learning so threads like these are very inspiring.
 
Top