Log In now to add this Gigapan to a group gallery.
Log In now to add this Gigapan to a gallery.
About This GigapanToggle
- Taken by
- Jeff Cremer
- Explore score
- 15.95 Gigapixels
- Date added
- Oct 19, 2012
- Date taken
- Sep 05, 2012
- landscape, travel
photo by Jeff Cremer
Visit me in the Peruvian Amazon: www.perunature.com/amazon-rainforest-photography-tours-photo-workshops.html
Gigapixelperu.com - At 15.9 gigapixels (one gigapixel is equivalent to 1000 megapixels), the image of the Inca citadel allows viewers to see the site in impressive detail.
The online image was taken using a Canon 7D with a 400mm lens. It actually consists of 1,920 pictures taken using a robotic camera mount, which were then stitched together. The photographer was Jeff Cremer, the photo tour director of Rainforest Expeditions, a tour operator that specializes in trips to the Peruvian Amazon. He believes his image is the highest resolution picture of the World Heritage site that has ever been taken.
“In 2008, the World Monuments Fund placed it [Machu Picchu] on its Watch List of the 100 Most Endangered Sites in the world due to environmental degradation mostly from tourism,” he said.
“I believed that this extraordinary site deserved a remarkable photo to hopefully raise awareness and help in its preservation.”
How it was done:
CAMERA SELECTION AND SETTINGS
I chose to use a Canon 7D because of the crop factor and the pixel density that it provides. Using a 400mm lens the crop factor gave me the equivalent of around 640mm at 18MP per shot. Canon 7D has a linear pixel density that is approximately 49% greater than that of the 5DII. (reference: www.robsphotography.co.nz/crop-factor-advantage-7D-5DII.html ) I set the camera at f/10 because I felt that it would give me adequate depth of field while limiting the effects of diffraction. I also wanted to have a fast shutter speed to reduce any effects of vibration but keep the ISO low to reduce noise. The final shutter speed was 1/640.
I mounted the camera and lens set up on a Gigapan Epic Pro and leveled everything out. I switched the shooting setup from the default setting of “columns down” to “rows up”. This was done for two reasons:
1. Changes in lighting conditions - If one is using the columns down mode and the light changes, due to a cloud passing overhead, the picture will look unnatural because it will create a vertical band going from light to dark. I have found that it looks more natural if the “rows across” is used. That way any change in light is confined to one band going across similar parts of the image, such as sky or a city. It is also easier for the stitching software to fuse the exposures if similar parts of the image have different exposures.
2. Focusing and Depth of Field - The second reason that it is important to use rows up is for focusing. If you look at the image there are mountains in the distant background, far from the camera while Machu Picchu itself is closer to the camera. If I used “rows down” I would have to stop the camera at almost every frame to refocus due to the subjects in the frame getting closer to the camera. Using “rows up” I started taking photos of the subjects nearest to the camera, at bottom of the image. I focused the camera on these first then as the camera moved up a row I would pause and refocus the shot then continue shooting.
Another reason I did this was because of depth of field. Depth of field goes deep into the background so if I started shooting the parts of the image nearest to camera first the depth of field would have those areas in focus as well as the parts of the next rows in focus. This made it easier for the software to stitch as well as helped eliminate any focusing errors.
I tethered the camera to my laptop using and transferred all the shots to my computers hard drive. I had to take 1920 photos in RAW and didn't have a memory card big enough to store all that information.
One of the problems that I had with this is that the software froze on me a couple of times and I had to pause the shooting and move the camera back to where I thought that the last image was taken. It was a little nerve racking because if I missed one image there would have been a hole in the final image.
Once the entire image was captured I processed all the RAWS in Adobe Bridge and then sent them to my friend Eric Hanson of Xrez studios in California. Eric has some pretty serious processing power and was able to stitch and edit the image with no problem. He used a MacPro Hexacore 2.67Ghz with 32GB Ram and OCZ 960GB RevoDrive. The RevoDrive is basically a solid state hard drive that plugs directly into a PCI slot giving extremely fast read/write times. The render time was 1.5hrs and created a 47GB .PSB file.
From there I contacted Susan Thesing at Gigapan who said to send the send the image to Paul Heckbert, a computer science professor at Carnegie Mellon University. I have a relatively low and unstable internet connection here in Peru so Paul was able upload the image with no problem. The image took 11.5 hours to upload to gigapan.com which is the largest upload of a .PSB file to date.
The software required 10.5 hours and 120 gigabytes of free disk space to read and convert the 47 GB .PSB file, and 1 hour to do the upload of 6.9 gigabytes of data with about a 7 megabits/sec upload throughput.
After it was all done I created a simple website and embedded the image. The rest is history.