Splicing together two kodachromes from 1966 (thank you Photoshop!)

Photoshop CS5 is so awesome.  I am almost entirely a Lightroom man, but I do go into Photoshop for three things: (1) panorama stitching, (2) the content-aware tool, and (3) cloning.  The first two features are not available in Lightroom and the third feature, cloning, is far superior in Photoshop in my opinion.

So, here is today's story:

Last weekend I was "digitizing" some of my dad's Kodachrome's from a trip to Europe we took together in 1966.  He used a Kodak Retina IIIC rangefinder camera.  With just a 50mm F2.8 lens, and film rated with an ASA of 25,  many today would view his camera as "limited" in the images it could create.  On the other hand, neither of us knew anything different, so it was quite adequate.

Visiting the Cathedral at Chartres, France we actually thought it was pretty cool that the structure was so large and tall that he couldn't get it all into the viewfinder.  I remember enjoying the slide show we created that first showed the bottom half and then showed the top half. 

When I saw last weekend that the two slides had significant overlap, I was hopeful that I might be able to splice them together into one vertical image. I didn't let myself get too hopeful about the outcome, however, once I noticed that the bottom picture was in horizontal orientation and the top picture was in vertical orientation. 

So... without doing any editing to the images, I decided to challenge photoshop.  I exported the images from Lightroom to Photoshop.  Photoshop did all the rest, automatically.  Below are the two original images and the result from Photoshop.





Next, to create a finished product, and since I couldn't crop the bottom half to fit the top half, I used the "content-aware" tool to fill in the sky in the top half.  It just took a minute or two.  I am happy about the photoshop "content-aware" tool, because I know nothing about layers and all that other good stuff that most power-users of photoshop know.

At this point I still had some fill in some additional sky.

Lastly, after saving the file as a .tiff and bringing it back into Lightroom, I made some other adjustments, such as a slight rotation of the image and some color adjustments.  I am sure more can be done, but the final image below is good enough for me (and my Dad likes it too!)

4 image panorama: Sulphur Mountain, Canadian Rockies (Banff)

The building you see here is the terminal for the Banff Gondola.  This pano is the result of 4 images. What you see is a reduced size copy so it can fit in this blog.  However, on my hard drive the original measures 14,000 pixels by 4,500 pixels.  If I did the math right, that's over 60 mp. That will make a huge print!

The camera was hand-held and images were shot in P-mode. Photoshop did a nice job stitching them together.  I did nothing other than click the right buttons.  (See my instructions in a prior post.)

If you want to see a larger view of this pano plus 6 others I included in the prior two posts, I have them on my Web site in larger sizes, here. Once you are there, if you click a couple of times on any of the thumbnails they should get "big", then "bigger".



3 image panorama with Photoshop CS5: Lake Louise

My prior post included several 2 image panoramas from our recent trip to the Canadian Rockies.  The 2 image approach is a good one.  The resulting pano, with a bit of cropping/trimming, will have an aspect ratio of about 2:1, which is very close to the 16:9 aspect ratio of an HD television.  The actual angle of view that you will get depends on the focal length of the lens used.  If you use a wide angle lens of 28mm equivalent field of view, I believe the resulting pano will have a field of view equivalent to about what you get from a very wide angle 14mm lens. (Someone reading this may be able to correct me on this; however in any case the resulting image is nearly twice as wide as what I can get with a single 28mm wide angle image.)  

Below I show a 3 image pano.   What makes this a bit different is that I held the camera vertically so as to get more sky and water.   This technique can result in a slightly more stretched out pano than what you see here.  However, in this case I had a greater than usual overlap in the second and third images, primarily because there was nothing I wanted to capture further right than the person in red.

This approach also makes sense if you want to make big prints.  The camera used has a 16 mp sensor.  However, because of multiple images this pano measures 25 mp.

More Panoramas with CS5: Two shot panos from the Canadian Rockies

I've been working on my Canadian Rockies images and have been very pleased with the panorama shots taken.  This is really very easy.  My cameras are set up to show the 1/3 viewfinder grid.  Basically, the grid looks like  tic-tac-toe lines on the viewfinder that divide the view into horizontal and vertical thirds. 

All my panos are stitched from hand-held shots, typically shot in in P-mode.  With today's advanced software (like Photoshop CS5), auto-mode will work fine, also. 

As I move the camera from left to right (you can also do a vertical pano by moving from bottom up) I merely overlap each picture by one-third of the frame.  In other words, the right 1/3 of the first picture needs to appear in the left 1/3 of the second picture.

The pictures are then exported to Photoshop CS5.  See screenshot below.  I always check "vignette removal" and "geometric distortion correction".  The other items are pre-selected.  Then, click on "OK".  You'll need to do a little cropping on the final result, to even up the edges.  But that is all there is to it.

Here are several samples of two images panos:

From Sulphur Mountain, Banff, Alberta

Bayview, Idaho

Athabasca Glacier from the Columbia Icefields Parkway

From Sulphur Mountain, Banff, Alberta

View of Banff and the Bow River from Sulphur Mountain

The Panorama feature in CS5 is awesome

I have had so much fun stitching two or three (usually no more than that) images together using Photoshop CS5.  CS5 snaps the images together flawlessly (i.e. 90% of the time, by my guess) and even blends different exposures used in each photo. 

Different exposures often result when using any of the automatic exposure modes, because the light invariably hits the scene in different amounts as the angle to the light source changes slightly each time you move the camera to take the next picture. 

Below are two images taken inside the Cabelas store in Idaho.  I used P-mode and auto ISO and simply overlapped two 24mm wide angle shots by.  No manual exposure. No tripod.  No fancy footwork required. 

[Taken with Panasonic LX5.]

The two images

Two image pano (before cropping)