“…to see the statement of intent that resides in natural form.” — John Szarkowski describing the work of Edward Weston in Looking at Photographs.
“The world just does not fit conveniently into the format of a 35mm camera.” – W. Eugene Smith
Panoramas are my favorite multi-image approach. I frequently shoot landscapes and very often the landscape elements are not consistent with camera formats (aspect ratios). Panoramas don’t need to be large numbers of frames; sometimes just 2 frames with an over lap of 20% will be exactly what you need. Panoramas are relatively easy to produce if you remember three things:
(1) Level your tripod and your camera. My tripod has a little circular level that helps a lot. There are also leveling heads that can simplify the process instead of alternately adjusting each leg. To level my D200 I use a little bubble level that fits in the hot shoe; it works great. My D800E has an attitude indicator in the camera that I find convenient because I can then have a GPS unit on the hot shoe, so I use that instead.
(2) Focus and exposure must be on manual. You need to set your best average focus point for the entire sequence and not have auto focus decide what is best or the stitching will be obvious; adjust your depth of field appropriately. You also want a consistent exposure, so go to manual exposure. Watch out if you frequently use a polarizer. As you change your camera angle relative to the sun, the polarizing effect will change as well. You can end up with a distinct change in exposure from one shot to the next, especially for skies.
(3) Find the nodal point (aka no-parallax point) in your lens. For telephoto lenses this is generally near the central point of the lens or about where the diaphragm is. For wide angle lenses it gets more complicated rapidly as you go wider and the nodal point can change with the angle of incidence (the amount of rotation). Finding the nodal point is more critical for scenes that are relatively close. If you don’t rotate around the no-parallax point, your images may not line up very well. To rotate around the nodal point, long plates that fit an Arca Swiss mount are helpful, so you can slide the camera back and forth along the lens axis until you are at the nodal point. With my 70-200 mm lens, which has a mounting foot, I am very close to the nodal point without much modification. Usually the images align very well. With my 24-70 mm which doesn’t have a foot, aligning the images is almost impossible if rotating around the camera itself. The less confident you are about the nodal point, the more space you should leave around your subject, so that if the alignment is off you will still have room to crop.
A web search for “finding the nodal point of a lens” will turn up numerous descriptions that don’t need to be repeated here, but very briefly, two vertical objects are set up, one closer the other further away, that align along the axis of the lens. A telephone pole and building edge can work. As you rotate the lens around the nodal point, they should remain aligned. If they don’t then shift the camera back and forth along the lens axis using a long attachment plate, until they remain aligned.
More often than not, I do a horizontal pan to capture the full scope and grandeur of a landscape, but I do occasional verticals as well. Verticals are a little harder
because there is nothing to keep the camera aligned as you tilt upward, instead of rotating with your ball head.
You can also do a horizontal panorama with your camera mounted vertically as I did with this image of the “Gossips” in Arches National Park. Sometimes you don’t need to add much and just two images are sufficient
to get what you want with the appropriate foreground and background.
I always rotate left to right so that the frames are taken in the sequence in which they will be stitched, although the program can manage either way. I usually overlap each frame by 20-25% and as I rotate I try to make sure that right side of preceding frame and the left side of the suceeding frame have some distinctive structure to overlap.
To facilitate processing in Photoshop, I shoot my hand at the beginning and end of each sequence to make it easier to see where the sequence begins and ends. When naming the files I use a scene name, followed by the number of the panorama for that scene, followed by a letter for the frame sequence. This can be done automatically in Batch Rename under the tools menu. So for example The first pan of a tundra scene might be TundraPan1a, TundraPan1b, TundraPan1c. If I shot a second pan sequence it would be TundraPan2a, TundraPan 2b, etc.
I usually shoot in RAW, so to begin I select the images in Bridge and double click on one to open them all in camera RAW. I make lens and chromatic aberration corrections from the lens corrections tab. I then make other global changes such as exposure, saturation or clarity. I don’t make local corrections such as spot removal. Then “select all” on the left and then synchronize. Click OK if you have not made local corrections; if you have just unclick those such as spot removal or local corrections. This will make the same changes to all of the images so they remain consistent. Then click Alt-open to Open copies.
Photoshop will open all of the images. When that is complete, File > Automate > Photomerge. Click “Add Open Files.” Generally the auto layout should work well, but you can experiment. Blend images will be selected by default and click OK. If your lens has some vignetting that can show up as darker areas where the frames were stitched. If it was not corrected by lens corrections, you can also click for vignette removal. Photoshop will automatically place each image on a separate layer with appropriate masks to align them. I then make sure they are all selected and go to Layer > Flatten image. I then close the component images to not take up memory (don’t close “untitled panorama or you will have to do it all over again). Often there are slight inconsistencies in alignment so I crop them out, then process according to my normal workflow.
Keep in mind that panoramas also increase the file size of the combined file and give you more pixels to create large images. The ultimate is the GigaPan that automatically rotates and lowers to capture dozens of images that can be combined to produce sufficient resolution for wall size images.
Keep panoramas in mind when out in the field. They provide the means to more
accurately capture your vision in some situations by changing the aspect ratio of your final image, rather than cropping, letting you keep the resolution you want, and capturing a scene as you see it without the perspective changes inherent in a wide angle lens.
Friday: HDR (High Dynamic Range)