I just viewed an image that had a "stitch of vertical takes" and I've noticed that many people have done this regularly. I know this is done digitally but...why? Don't wide angle lenses do this? What are the advantages of this procedure? How is it done? Thanks!
Well, i dont do stitches cause i dont feel the need to, but to answer your question it gives you a much larger image, pure and simple (a stitch of 4 shots gives you an image almost 4 times as big!) Its done using either specialized equipment or just panning/tilting the tripod, then stitching the image together with software that does just that.
Chris - In addition to what Stan said, you can also go much wider than with any lens, and without the distortion of using something like a fisheye lens. I have panoramas stitched that are 180 degrees or more of the view. Plus they didn't cost me anything extra to make, like a new super-wide lens would.
As far as how it's done, you simply take several shots that have some overlap, and let the software have at them. It can get much more complicated, of course, but I've taken handheld panoramas with a point and shoot, and used a simple stitching program with all defaults (Photoshop will do it very well in CS4), and gotten something I was very very happy with.
Now to make higher quality, prevent alignment errors, etc., you can get into specialized equipment, and techniques, but they aren't always necessary.
Posted by Kent Staubus on June 22, 2009 at 10:42 AM
A couple of months ago I saw a train parked on a siding that had 14 engines on it. Each engine is 80 ft. long. If I had back up far enough to even get a shot with my 11mm lens, there would have been a few trees in the way. So, I used a 30mm lens and took a bunch of shots to stitch together.
Stitching multiple shots to create a single final image has almost become my standard way of shooting lately, I do it everytime I can. By doing this you got the advantages Stan and David had already mentioned, and even a few more. Yes, the first and more obvious is the gain in image size which allows you to print at much bigger sizes with outstanding image quality and details. For example, if you stitch together 12 shots from a 15 Mpxl DSLR - like my eos 50D - placed in three rows of 4 pictures each, you can come out with a final image of about 70-90 Mpxl (depending on the amount of overlapping), which I guess can beat the image quality of film large format too.
But by doing big composite images you also have an increased field of view, you can reach the field of view of a very, very ultra wide lens such a fish eye, but without the distortion this kind of lenses produce because you actually shoot with a rectilinear lens.
For example, my tokina 12-24 covers about 85° of FOV in the longest size and 57° on the shortest. Stitching 4 verticals with this lens at 12mm, allowing the necessary overlapping between frames just to prevent stitching errors that easily occurs with wide angles, you can easily produce a final horizontal image covering about 85° vertically and 130° horizontally. A fish eye field of view indeed, but perfectly rectilinear!
And then again, if you want to compose multiple rows, you get a gain in the apparent depth of field in the final scene, using wider apertures. This is easy to understand because you can properly focus each single row, so you will do close focus on the lower row, longer focus on the mid-row, and infinite focus on the top row. This is great when you want to shoot at faster shutter speeds without sacrificing dof.
Finally, doing composite images is just like having a much bigger sensor within your affordable consumer and/or prosumer level camera. Surely full frames camera are better cameras, better dynamic range, better noise control at higher ISOs and so on. But honestly, mid range Canon cameras are already amazing and incredibly cheaper, and if I can equal, or better, surpass the resolution of their sensors at a fraction of the cost, well, I'm simply the happiest guy around for that...
Posted by Steve Sieren on June 22, 2009 at 12:27 PM
Wide angle lenses and stitching can lead to difficutlties though if not correctly.
You really want to take note of how close your foreground subjects are.
Too close of a fg subject and you won't be able to stitch easily.
Pano heads and certain software can help you take advantage of using a NODAL POINT for easier panoramics. That would be a little more advanced for a beginner at panos. Just trying 2 or 3 images would be very easy to start with.
Your imagination becomes far less limited when you can think of how to stitch an image together while out in the field. The rectangular box view just confines the imagination.
I have a little experience doing stitches. Sometimes it is just to get a slightly wider angle of view (my shortest lens that works with my dSLR at the time was a 20 mm, that ended up looking like a 32 mm due to the smaller sensor). In other cases, I just wanted to play with truly panoramic views. I made a 12x48" print of several frames spanning almost 180 degrees of a vast cliff face in Capitol Reef. I actually sold that print (to my amazement). I still have to get a replacement made for my own wall!
Discussion
6 repliesAs far as how it's done, you simply take several shots that have some overlap, and let the software have at them. It can get much more complicated, of course, but I've taken handheld panoramas with a point and shoot, and used a simple stitching program with all defaults (Photoshop will do it very well in CS4), and gotten something I was very very happy with.
Now to make higher quality, prevent alignment errors, etc., you can get into specialized equipment, and techniques, but they aren't always necessary.
Kent in SD
But by doing big composite images you also have an increased field of view, you can reach the field of view of a very, very ultra wide lens such a fish eye, but without the distortion this kind of lenses produce because you actually shoot with a rectilinear lens.
For example, my tokina 12-24 covers about 85° of FOV in the longest size and 57° on the shortest. Stitching 4 verticals with this lens at 12mm, allowing the necessary overlapping between frames just to prevent stitching errors that easily occurs with wide angles, you can easily produce a final horizontal image covering about 85° vertically and 130° horizontally. A fish eye field of view indeed, but perfectly rectilinear!
And then again, if you want to compose multiple rows, you get a gain in the apparent depth of field in the final scene, using wider apertures. This is easy to understand because you can properly focus each single row, so you will do close focus on the lower row, longer focus on the mid-row, and infinite focus on the top row. This is great when you want to shoot at faster shutter speeds without sacrificing dof.
Finally, doing composite images is just like having a much bigger sensor within your affordable consumer and/or prosumer level camera. Surely full frames camera are better cameras, better dynamic range, better noise control at higher ISOs and so on. But honestly, mid range Canon cameras are already amazing and incredibly cheaper, and if I can equal, or better, surpass the resolution of their sensors at a fraction of the cost, well, I'm simply the happiest guy around for that...
You really want to take note of how close your foreground subjects are.
Too close of a fg subject and you won't be able to stitch easily.
Pano heads and certain software can help you take advantage of using a NODAL POINT for easier panoramics. That would be a little more advanced for a beginner at panos. Just trying 2 or 3 images would be very easy to start with.
Your imagination becomes far less limited when you can think of how to stitch an image together while out in the field. The rectangular box view just confines the imagination.
Pete D.