Panoramas of Washington National Cathedral

July 07, 2011  •  Leave a Comment

Lately, I've been interested in creating panoramas. I've been reading up on the subject and found a wonderful open source application called "Hugin" that can be used to create panoramas by stitching together photos. The latest release is quite powerful and despite the fact that it's free software, it's quite polished.

To test out the latest release of Hugin, I visited Washington National Cathedral. I've always loved cathedrals and the National Cathedral in particular. I also recently finished reading Ken Follett's great book The Pillars of the Earth, which has me even more excited about cathedrals. So, the National Cathedral was a perfect subject for my test.

My goal was to take a 360 degree panorama of the interior of the cathedral from the crossing (the point in a cathedral where the nave, transepts, and presbytery intersect). I think it worked:

Stitched panorama of the crossing at Washington National CathedralWashington National Cathedral Crossing (1:5 panorama)

To take this image, I placed my camera on a tripod and used a spirit level to ensure that the entire setup was perfectly level. I set the focal length to 24mm with an aperture of f/8. I took some test shots and then set my camera to full manual mode using the shutter speed taken from the test shot (to ensure even exposure across every image). Then I began shooting. After each image, I rotated the camera slightly to the right, making sure the next image would overlap with the previous image by at least 25 percent. After rotating an entire 360 degrees, I angled my camera upwards 45 degrees and repeated the process. (This took a long time because I had to wait patiently for people to move out of my shots.)

Each image that I took was actually a series of 3 shots--one evenly exposed, one under exposed, and one over exposed. In total, I took 117 shots. For each set of 3 shots, I later blended them together into one high-dynamic-range (HDR) image using tone mapping software called HDR Efex Pro. I did this because the lights and windows were too bright and the shadows were too dark otherwise. The act of blending the photos toned down the lights and brought out the details in the shadows.

This left me with 39 high dynamic range photos, which I stitched together in two rows using Hugin to form the image above. I think the result is quite spectacular. (The final image is massive, so I reduced the size for purposes of posting it online.)

While I liked the panorama from the crossing, I had been hoping to get an image that featured the sanctuary and the altar more prominently. So, I entered the presbytery and repeated the same process, except that this time, I used a 12mm focal length (very wide) and in Hugin I used a mercator projection instead of a cylindrical projection.

Photo by John Baggaley of the presbytery and sanctuary of the National Cathedral in Washington, DC.Washington National Cathedral Presbytery and Sanctuary I think this one turned out spectacularly. In the center of the image is the sanctuary and on the far left and far right is the choir. Since it's a 360 degree image, I could have set the center of the image to be any point of my choosing. Here's the same panorama, but with the choir in the center and the sanctuary on the far left and right and a bit more of the ceiling cropped out:

I hope you will agree with me that my experiment was a success. I love the unique perspective that the panorama stitching software provides. However, generating the images was quite laborious. In taking the photos, I had to use manual camera settings and ensure that the the camera was perfectly level at all times. Moreover, between the two panoramas, I processed 201 images. Blending and stitching takes a lot of computer processing, so the entire process was very time consuming. However, I think it was worth it, and I look forward to more such experiments.


No comments posted.
January February March April May June July August September October November December (1)
January (1) February March April (1) May (1) June (1) July August (1) September (1) October November (1) December
January February (1) March April (1) May (1) June July (1) August September October November December