Creating panoramic images which can be printed or displayed on the screen as a still picture is fairly straightforward. With appropriate software, it is even possible to convert the on-screen image into a movie, which you can pan around, turning left or right, pausing to examine different parts of the scene through a full 360 degrees.
Special wide angle lenses are not needed to make a panorama. In fact a very wide angle lens can induce “fish eye” distortion which is seen in many 360 degree “virtual tours” created with proprietary systems. Panoramas can be made with any camera and lens and movies using them are viewable without plug-in or any special client software except a standard Internet Browser.
There are 3 steps involved: the first is capturing a series of images that covers the entire scene you wish to show; the second step is to merge (“stitch”) the individual images together, creating a single picture; the final step is to apply "panning" effects, which can be quite easily done in various software packages. In this Intel, I will explain the process of capturing the images you need and describe 2 inexpensive software packages which are capable of accomplishing the task of merging and displaying your panoramic video.
One set of software products is from ImageMatics (http://www.imagematics.com/ ); they provide easy to use pan effects which are ideal for the task, and produce flash (.swf) files. PanoramaPlus 4 from Serif automates the entire process from stitch to virtual movie and uses Quicktime.
Most panoramic images are created from a series of still photos that are "stitched together" using the stitch function provided by many Imaging and Paint packages. Typical panoramic video displays are created by stitching together between 2 and over a dozen still images. The tricks to good panoramic images are a level camera and plenty of overlap between the images.
A tripod is very useful for 360 degree panoramas where you want to simulate movement, but a standard tripod head rotates the camera around the axis in a way that makes it difficult to align the images realistically. For these “virtual tour” type movies, a tripod fitted with a special panorama head is a real boon. For still pictures it is helpful, but not essential.
Rather than using the camera in the “normal” horizontal position, consider using the vertical or “portrait” orientation. This provides a better height to width ratio when the images are stitched - you never have to worry about the picture being too short...only too narrow.
The panorama head for your tripod mentioned above incorporates a bracket that rotates the camera on the “Nodal Point”. The nodal point is the point inside your camera where the light rays converge and cross. Making this the rotation point eliminates image mismatch caused by parallax error. It is more important if you are creating panoramas of small spaces with short distances to the camera, because this emphasises the “parallax” error.
These heads can cost as much as your tripod, so it’s fortunate that careful hand-holding can produce very good results! For more information on nodal points and panorama heads, see http://www.edb.utexas.edu/teachnet/QTVR/NodalPoint.htm
Taking the photos:
For best results, overlap the frames for your panorama by about a third. To do this, line up the first shot and note an object in the picture 2/3 in from the left side (assuming you turn to the right). This object should placed at the left-hand edge of your next frame.
Keep the camera level. Turn on the grid in your display if you have one, or use the focussing squares or upper frame edge as a guide: make sure it is on the equivalent point in successive frames. For instance, let the upper focus frames be the same level above or below the horizon in each shot.
Every other guide will tell you to use manual focus and manual exposure to prevent having different points of focus and different exposures in your individual frames. This is because different parts of the scene are lit differently, and you may finish with odd changes in colour, tonal density, and focus shifts which make blending the final image difficult.
Exactly the opposite applies when using dedicated software to align and merge the frames. I use program mode and autofocus for my panos. This allows me to produce panoramas with both near and distant features in sharp focus (infinite depth of field; see my Intel on Depth of Field and Hyperfocal Distance). Focal length is another matter: keep it the same through your shots.
I tend to mix shots which are not intended to become panoramas on the same shoot as my panos; with several hundred pictures to sift through, it is useful to be able to locate the appropriate frames quickly.
To do this, make a blank frame immediately before frame 1 of the pano group - just press the shutter button without taking the lens cap off (you may need to shift to manual exposure first). If your camera won’t allow this, take a picture of your foot to indicate the start of a panorama sequence.
Putting the pictures Together:
Using any editor that supports layers, make a new image with a canvas larger than the length of your final picture. If it is 3 frames wide, make the canvas 3x wider than the standard image. Keep the height the same.
Load each image into a separate layer and, working in pairs, set the transparency of the top layer to about 50%. then rotate and align the one immediately below. Merge this pair and repeat with the merged image as the top layer, working down through the images. the final image will be smaller than the canvas, and needs to be cropped.
The only processing I do with the images before merging is to correct the horizontal or vertical alignment. Other corrections are done on the composite image.
This is not as simple as it sounds. It is seldom easy to get the alignment right, perspective and tonal shifts often make the joins obvious and unnatural, and any motion between exposures has to be hidden (waves, passing cars, or wind in trees). There is a lot of work required with the clone tool and a fair bit of burning and dodging required.
The alternative is to use more specialised tools.
I load all the panorama sequences into PanoramaPlus, in no particular order, even images from several sequences at once. If I'm not sure which were pano sequences and which were not, I add them "just in case". Then I click “stitch”.
The software determines which images will make a panorama and presents all the potential combinations as large thumbnails in a window. It will even suggest alternate versions using the same images.
I then take each pano that I like the look o and fine tune the alignment and cropping, then press “export”; I choose the destination, file type and size and PanoramaPlus does the rest. It aligns, melds and blends the images seamlessly, correcting exposure and tonal problems, and any perspective issues.
Rarely, I may fine tune the output in PaintShop, but usually the result is a ready to print image. Unlike other similar programs, PanoramaPlus works in 2 dimensions – stitching photos together horizontally and vertically!
One of the export options is a Quicktime movie, which provides the virtual tour option mentioned earlier in the outset, and really needs no further explanation.
If you do not have PanoramaPlus or a similar package your next step is to open the final image in StillMotion Personal Edition which you can download free at http://www.imagematics.com.
This is actually a very competent tool for creating multimedia slide shows, web content and audiovisuals. It has a nice demo of the software’s panorama capacity showing a virtual tour for real estate agents on its site.
To achieve the same effect as with the PanoramaPlus example, you simply open the panorama in Stillmotion, apply the pan and zoom effects you want from the menu and export as a movie!
The online tutorials and demos make the process very easy.
