Then and Now photos - alignment


(Alan Gibson) #21

I suggest you try it and see. Take photos with those lenses, taken from the same spot, pointing in the same direction. Crop the wide photo down to match the same field of view of the long lens. Aside from barrel/pincushion distortion, they will match geometrically.


(srowed) #22

This is true. A good example of this is in using a telephoto lens to shoot multiple photos and stitch them together, for the purpose of increasing the resolution. The same shot can be taken with a single frame using a wide angle lens, but the resolution will suffer when blown up to mural size. The key point here is that the relative size of the nearby and distant objects will be the same using either a wide or tele lens.


(srowed) #23

Today I bought a USB OTG (on the go) cable, which allows me to connect the camera to my phone via USB connections. I installed the “Camera Connect & Control” app from the Google Play and it does work. I haven’t tested it in depth yet, but it does look promising for this purpose as it displays the camera live view on the phone LCD and allows me to change camera settings and take the pictures from the phone. You can even touch the phone LCD to make the camera focus on a given spot.

A few questions and trials remain but this is encouraging. My quick test tonight was with a Nikon D7100 with a Nikkor 24-70 f2.8 attached to a Samsung Galaxy S4 phone. I’ll try the setup in daylight outside to see how well it will work in a real life situation. I don’t know if the phone’s screen will be big enough, or if I should use a tablet. I also need to figure out the best way to hold the phone, such as attached to the tripod with a gorilla pod or super clamp, and the best way of shading the screen (black cloth covering my head and the phone?).

Once I get some reasonable tests done with the same lens I’ll test using different lenses with slightly different focal lengths to see how Hugin handles lens distortions.

The software seems to have a lot of useful features. https://play.google.com/store/apps/details?id=com.rupiapps.cameraconnectcast


(srowed) #24

That is roughly what I did with the shot of my mom and daughter, although it wasn’t perfect. For using an acetate overlay on the LCD, it’s necessary to use a zoom lens in order to adjust the focal length as well as the camera position. And sure, after the camera position is determined as accurately as possible, you can zoom out a bit if necessary to ensure the “now” photo completely covers the “then” photo. It would also help if the zoom lens had as little distortion as possible.


(Alan Gibson) #25

To my previous post, I’ll add that if we are photographing a particular group of objects (such as people) with a wide lens we typically move close to the group, to fill the frame. With a long lens we need to step back to get them all in the frame. Moving the camera forwards or backwards changes the perspective, the relative size of distant and near objects. Changing the lens, or cropping a photo, doesn’t change perspective.

If we change focal lengths without moving the camera, the perspective is unchanged.

(True, a wider angle from the same position shows more context, and this may change the perception of the objects.)

For then-and-now photos, the “now” camera lens is ideally placed in the same position as the “then” camera lens.

On barrel/pincushion: the goal is, of course, to correct one image against the other. I expect the Hugin toolset can do that. Quite possibly the “now” photo could be corrected to to an absolute standard, and the “then” photo corrected to the result, if desired.

After geometric corrections, it may be desirable to change lightness and contrast, and sharpen or blur, so the images perceptually match. Or go the other way, and sepia-tint the “then” photo.


(Pat David) #26

As a small side note, I would love to help you write up an in-depth article on producing your images and how things worked out (what worked, what didn’t, etc) to publish here if you’re up for it.

More related, Hugin can work wonders bending things to fit if you’re ok with some level of pixel bending. It may be some manual work depending on how tight you need the alignment. (Of course, a more automated align_image_stack might be fine for many instances).


(Glenn Butcher) #27

On the narrow gauge railroad forum where I spend time, a couple of folk do then-n-now photography of scenes where railroads did/still run. Regarding these, what I’ve found is that precise alignment isn’t so important as the placement of key structural or geographic components. Here’s an example:

http://ngdiscussion.net/phorum/read.php?1,356301,356847#msg-356847

John Fielder published a book of these, going out to W.H.Jackson locations and imaging the present:

https://www.johnfielder.com/1870-2000/

Note the Fielder page shows him with a view camera and a printed picture, eye-balling it.

FWIW…


(srowed) #28


I’ve done a lot of photography for Rocky Mountaineer Railtours in Alberta and BC. About 20 years ago they wanted me to do a “now” shot of their train entering the Spiral Tunnels in Yoho Park, BC. This was originally shot by the CPR photographer, Nick Morant, on a steam engine coming into the tunnel. Arrangements were made with the train crew to stop outside the tunnel so we could get the shot. I set up two cameras (both Pentax 67s) on tripods on the tracks inside the tunnel. The creative director was outside on the curve so he could communicate between the train crew and me to position the train in exactly the right spot.

Right on schedule I could hear the train coming, although it was around the corner so I couldn’t see it. It didn’t sound like it was slowing down. The creative director started yelling at me “He’s not stopping. GET OUT!” I grabbed the cameras and ran for the tunnel entrance. I couldn’t get completely out and had to press up agains the rocks while the train went past. The train crew never even saw us.

We then decided that it would be a lot smarter to wait until the train stopped, then go into the tunnel and take the picture, which is what we did a few days later. The final shot was used extensively for their advertising.


(srowed) #29

Hi Pat, yes let’s do that. I know we talked maybe doing a couple of articles some time ago - one my workflow for scanning the old negatives with a DSLR, then using open source software to process then. The other you mentioned you’d like to do would focus on my father’s photography, which covered an incredible variety of material dating from the mid 1930s through the 1970s. It seems that his first major assignment was to cover the 1936 Berlin OIympics - really haunting images.


#30

Glad you weren’t hurt. :slight_smile:


(Alan Gibson) #31

Love the train shot. The cloud top-right really makes it. Was that luck, or good planning, or darkroom trickery?

Which reminds me that back in the day, many roll-film (ie larger than 35mm) cameras, and certainly plate-film (eg 5x4) cameras had lens movements: tilt and shift. Even quite cheap cameras often had movements.

So, beware that “then” photos might have used movements, eg using Scheimpflug to get distant mountains and close-up ground both in focus, or rising lens so mountains don’t shrink when the camera points upwards.

If this happens, a modern camera may not give an image that can be aligned without digital tinkering.


(srowed) #32

Yes, that will make it more complicated. For those shots I’ll probably be satisfied with just approximate alignment. For my father’s photography, he used a variety of cameras: 35mm, 6x6 cm, 6x9 cm, 3x4 in, 4x5 inch and 5x7. At least with the 35mm and 6x6 he wouldn’t have been using tilts and shifts. The Vallarta shot, for instance, is 6x6, probably from a Rollei twin-lens reflex.

This was back before digital cameras so I shot it on film and scanned it. The sky and the mountain were not manipulated other than maybe a bit of brightness, contrast and colour, but I did a fair bit of digital imaging in the rest of the shot. First, the tunnel had bushes grown up along some of the side and top that weren’t there in the original Nick Morant shot, so I cloned them out. Then the front of the train, as it was close to the tunnel, was about five stops darker than the mountain scene. I bracketed the exposure and cloned in a properly exposed engine into the scene from a different scan.

Then the creative director mentioned that the interior walls of the tunnel had no texture as they had in the Morant shot. In the intervening years the soot from the diesels had blackened the walls so the cement was covered. I scouted around for a shorter tunnel where the soot didn’t accumulate, walked in about the same distance, shot that tunnel and replaced these tunnel walls from my original shot.

These were the early days of Photoshop. No digital cameras, no raw files - you shot film and scanned it. In any case, it worked and they loved the result. Ironically, a member of the public complained that the shot was digitally manipulated because he claimed that a dark line along the edge of the mountain against the sky was a seam between two shots. I laughed at that because the line was an actual shadow on the mountain and that area of the picture was one of the few areas that WASN’T manipulated.


(Flössie) #33

I’ve been following this thread with half an ear, but like to point you to

  1. Ghost images in Magic Lantern
  2. Aligning images with Hugin

HTH,
Flössie


#34

Moin @floessie!

MagicLantern is a good idea!
ML is one of the few things I really miss from
my Canon era.

MfG
Claes in Lund, Schweden


(Mica) #35

Aligning the image with hugin might work really well if you manually choose your points. Hugin will take care of all the tilt, great idea


(Luc Moreau) #36

With some cameras (mine included) you can pre-load a picture on the memory card and use that as a start for a multi-exposure sequence. In that mode the “then” photo can be displayed in an overlay on (or under) the live view or EVF “now” image to be taken. Of course once the alignment is done you can lock that down on a tripod, skip the multiple exposure and just take a standard picture. In theory that would work (the fact that it works on the EVF might allow more precise work than the back LCD screen). I have to check to see if this works in practice, I never tried. FYI this info is based on my Lumix G80 instruction manual. The pre-loaded image has to be taken in RAW by the same camera so you might need to re-shoot your image negative (or a positive printout) in RAW with the same camera for this to work?


(srowed) #37

That’s the kind of solution that would be ideal. I played around a bit with my Nikons but there doesn’t seem to be a way of viewing the live view over top of the picture on the card. So far it seems the best solution with my gear is to print off the “then” picture onto transparent film and put it over top of the LCD screen on my phone or a tablet. The Android app Camera Connect and Control seems to work well and is only $10 Cdn. It would be even better if it had an overlay feature or even the ability to instantly switch from the “then” to the “now” pictures instead of sliding the image to the next one.