Thursday, January 22, 2009

Converging verticals

click photo to enlarge
One of the problems in photographing tall buildings is converging verticals. If you want to include most or all of a tall structure you have to point the camera upwards causing the verticals to converge on a "vanishing point". In the days of film, and still today, photographers used "tilt-shift" lenses that can be adjusted to correct this effect. However, they are prohibitively expensive for most of us, and are largely confined to professional architectural photographers. But, the advent of digital imaging allows anyone to correct verticals in software with varying degrees of success.

Today's photograph shows my attempt to secure an image of the west elevation of the the 282 feet high medieval Gothic church of St Wulfram at Grantham in Lincolnshire. A gateway into the churchyard, nearby buildings and trees prevent the photographer moving back far enough to photograph the building without tilting the camera upwards. So, I took a shot from as far back as I could, knowing that I wanted to "process" it back to vertical. Back home at the computer it was easy enough to correct the verticals, but that resulted in a very vertically compressed image in need of elongating with bicubic interpolation. The width of the church is 79 feet, so to get the proportions right I had to stretch the building until it was 3.6 times as tall as it was wide. But, because I'd tilted the camera the relative sizes of parts of the structure, particularly the spire, were wrong in the original shot, and stayed wrong in the "corrected" version: even though it was "right" it looked wrong. Consequently I compromised with this version that understates the height of the church but looks more correct. In fact, in my photograph St Wulfram's is 2.6 times as tall as it is wide, and though it makes for a reasonable picture, it is a definite failure in architectural photographic terms.

This example of the problem of "converging verticals" is quite extreme: most photographs that need correction are able to be amended without the problems encountered here. Why was I doing it? Well, many consider this spire, the third tallest on a medieval parish church after Louth and St Mary Redcliffe, Bristol, to be the finest example in Britain. Though it has much to admire I wouldn't go that far, preferring the broach spires of the fourteenth century to later examples such as this one. However, I thought that if I could see it without the distortion that you experiences during a visit as you look up, and as you see it in an uncorrected photograph, perhaps my appreciation of its architecture would change. So far it hasn't.

photograph & text (c) T.Boughen

Camera: Olympus E510
Mode: Aperture Priority
Focal Length: 11mm (22mm/35mm equiv.)
F No: f7.1
Shutter Speed: 1/250
ISO: 100
Exposure Compensation: -0.7 EV
Image Stabilisation: On

4 comments:

Anonymous said...

You've done a good piece of work with this shot. You say it's not up to architectural photograph standards, but it sure makes a fine picture. I've done this kind of photograph myself so I can guess that the original was very tapered. Perhaps you should have posted "before" and "after"!

Tony Boughen said...

Thanks Anon. I did consider doing a "before" and "after" but knew I don't like to post too many doubles images.

Regards, Tony

Tony Boughen said...

Thanks Anon. I did consider doing a "before" and "after" but knew I don't like to post too many double images.

Regards, Tony

Kenny said...

I've tried to get this church in one image and failed miserably the first time, you've done a good job here though! I thought of trying to do it using photomerge, but I'm not sure how that will work!