Exploring the natural and built environments through photography
There are many ways to create panoramas. If you have a smartphone, it’s likely the native “camera” application has an option allowing you to pan your camera and automatically stitch together an approximation of a panoramic scene. Alternatively, you can manually take a series of photographs and merge them together in post-processing using any number of applications. Whatever camera you use, best results are achieved by keeping the camera level as you pan to capture the images. Using a tripod that has been leveled is ideal for this technique. By keeping the camera level, you will minimize the amount of vertical shift that will be needed between the images when stitching together the photos.
I captured this photo of the landscape at Mt. St. Helens in Washington using a handheld point and shoot camera. You can see the significant steps between the six photos merged in the panorama because I didn’t hold the camera level as I rotated. Unfortunately this significantly limited the maximum available vertical dimension of the image.
Try to overlap each image by about 1/3 or so to provide enough common information for the stitching software to use to piece together the panorama. I have used a tripod that has markings showing the full 360 degrees of rotation, and will use the guide to relatively precisely divide the scene into even increments.
Even when maintained perfectly level, the rotation of the camera introduces parallax distortion which will negatively impact the final image. There are sophisticated (and expensive) specialized panoramic tripod heads that will allow you to either manually or automatically re-position the camera in order to eliminate the parallax distortion for those committed to making the best possible panoramas.
My lens kit includes a tilt-shift lens. A shift lens provides another alternative for creating relatively distortion-free panoramas. By positioning the camera on a tripod, you can utilize the shift movement in the lens to capture a series of images without rotating the camera. Simply set up the lens so the shift movement will move the lens horizontally rather than vertically. Shift the lens to the maximum to one side, then center, and then to the opposite side. The tilt-shift lens has markings to indicate the amount of shift in each direction, so you can also shoot more than three images with precise overlap if you so desire. In my experience, this hasn’t seemed necessary, as there is enough overlap with three images for the stitching software to be able to piece together the final panorama.
I captured this image of San Francisco using this technique. You can see only very minimal vertical shift (slivers of white area in the upper left and lower right corners). I suppose this is because I didn’t perfectly level my little table-top tripod sitting on the hotel room ironing board, but it’s obviously pretty minimal.
Interestingly, the majority of this panorama was created only from the far left and far-right photos. The images below shows the Photoshop layers in the merged file that were created from the left and right photos, with the hole in the sky being the only part that was used from the center photograph.
One benefit of multiple photograph panoramas is the ability to create much higher resolution photos that a single shot could provide. As an example, my final cropped image of San Francisco at night at 3:1 aspect ratio is 9,140 pixels x 3,047 pixels. If the long dimension had been cropped from a single 4:3 aspect photograph, this would correlate to an almost 63 megapixel original image. Prior the recent introduction of the Canon 5Ds with it’s 50.6 megapixel sensor, no DSLR would even come close to providing this resolution. At this resolution, there is incredible detail as seen in this 1:1 crop showing the area around the Ritz Carlton hotel from center right area of the above image… (click for full-resolution display)
Another option when using lens shift movements is to position the camera in portrait rather than landscape orientation, and then use the lens movements to shift from left to right. The resulting merged image will have an aspect closer to a standard 4:3 image, but with a much higher resolution.
For my workflow, I process my initial RAW files in Lightroom. Note that if you shot your original photographs using auto white balance, make sure before merging that the color temperature is the same for all of your photos. If making any other adjustments to exposure or color, they should also be synchronized between the photos to be merged. I then merge to panorama in Photoshop CC, then make final adjustments to the 16-bit TIFF file in Lightroom.
For additional reading, check out Really Right Stuff’s “Panoramas Made Simple“. Obviously they’re in the business of selling specialized panorama equipment, but there’s still a lot of good technical information included that can be of use to anyone trying to create panoramas, regardless of the techniques or equipment used.