The Most Detailed Photo of Tokyo Ever Taken
io9 scored some interview time with the talented, patient and hard working, Jeffrey Martin. How do I know Mr. Martin is all these things? Simple. Because that’s what it takes to make a photo so large it took 2 days to shoot, 4 months to edit and is comprised of more than 8,000 frames, at a resolution of 600,000 pixels wide. Watch this video that demonstrates the awesomeness that is the largest, most detailed, and zoomable, panoramic photo of Tokyo ever taken. Then go play with the photo for yourself.
Jeffery Martin explains his “transhuman aspect” for photography:
“It’s the idea of creating a view that literally extends our senses far beyond what we can sense on our own. This image shows you orders of magnitude more stuff than you can see when you are actually there. Even if you are on Tokyo Tower with binoculars or a telescope, this image shows you more than you can possibly take in, in person.”
“I use a Canon SLR, in this case, a 7D and a 400mm lens, in this case, a Canon L f/5.6 lens mounted on a Clauss Rodeo on gigapixel robot which is controlled by a tethered laptop to move the camera and fire the shutter while the camera is moving. This requires setting a very fast shutter speed obviously, and tuning everything so that the image is shot correctly without any motion blur and a minimal amount of noise. The image is 600,000 pixels wide. The maximum size of an image in photoshop is 300,000 pixels in any dimension. This means that to create a single seamless image, it was a big challenge. The image never has actually existed as a single file. However, the two files do fit together seamlessly, so it can definitely be considered a single image.”
“The image is 600,000 pixels wide. The maximum size of an image in photoshop is 300,000 pixels in any dimension. This means that to create a single seamless image, it was a big challenge. The image never has actually existed as a single file. However, the two files do fit together seamlessly, so it can definitely be considered a single image.
Also, disk space. Various iterations of the image in progress were tens of gigabytes each. The final versions of the image are 100GB each. I had to convert that into cube faces, each of those are 80GB each. This adds up – people say that disk space is cheap – well, it’s not!
I also needed more than 1.5 terabytes of free disk space to use as the “scratch disk” while doing the final assembly and retouching of the image.
The computer I was using had 192GB of RAM, and still lots of stuff was fairly unresponsive.”