3

I am attempting to take 2 photos of the same scene with different settings and would like for them to be as aligned as reasonably possible. I'm doing the following to help in that goal:

  • I'm keeping the aperture the same and varying only the ISO and shutter speed.
  • I am using a (midrange) tripod and a wireless remote, but I fear I may be nudging the camera slightly when I change the settings
  • Photographing scenes with no moving objects (trees, animals, etc.)
  • Shooting indoors where there's no wind. It's possible the floorboards move slightly as I move around, though I'm usually not moving a lot when taking the photos.

I'm shooting with a Canon 7D and have Mac I could tether it to if that would allow me to change settings without touching the camera.

Currently, I find that when I pixel peep them at 100% on the computer there's a shift. On some of the image pairs it's a fairly small shift, but on a few it's quite large. Because I'm going to be running them through my own software, I won't be able to use something like Photoshop's HDR alignment tools first.

What are some things I can do to improve the alignment of the images I'm shooting?

Update: I gave some thought to what the comments and answers said and made some changes and got some great results! I wasn't able to switch out my head just yet, though in doing this experiment realized there are other reasons why I should anyway. So hopefully I'll do that soon.

I gave some thought to vibration isolation and ended up not only tethering the camera to the computer to control with the EOS Utility, but then controlled the computer remotely from a computer in another room. This completely removed me from the room and seems to have been enough to get the alignment I needed. Thanks to everyone for their help!

  • As much aligned as reasonably possible yet complaining noticing it via pixel-peeping and worried about floor boards moving... I think your far into the unreasonable frontier. I'd say thats fine, but maybe you are looking for "lab control" and not reasonable control – AthomSfere Mar 11 '18 at 1:48
  • Could be. I'm asking because I'm not sure. Some of the misalignment was probably visible without pixel peeping, which surprised me. I was expecting small variations, but not large ones, so just trying to get some ideas for improving it. – user1118321 Mar 11 '18 at 2:01
  • A pixel size of 4 micrometers and a focal length of 50 mm gives a rotation angle needed for a one pixel shift of 4 micrometers/(50 mm) radians = 8*10^(-5) radians = 4.6*10^(-3) degrees. That's why you really can't do without alignment tools. It's still useful to limit shifts as much as possible, as the remapping needed for the alignment of pictures will introduce interpolation artifacts and lead to unsharpness the more misaligned the pictures are. – Count Iblis Mar 11 '18 at 17:27
  • Excellent point! I hadn't thought about how small each sensel actually is! – user1118321 Mar 11 '18 at 18:56
4

The greatest single improvement you can do is to stabilize your camera mount. That's were the highest amount of variability is likely occurring.

  • You need a solid floor that doesn't shift.
  • You need a much more solid mount than a 'mid-range' tripod. Use a heavy duty video tripod or a permanently installed mount that doesn't shift.
  • You need a solid tripod head that can be securely locked down and doesn't wobble.

As the old saying goes, a chain is only as strong as its weakest link. It doesn't matter how well you do the tripod and the head, if the floor isn't solid you won't get the desired results.

Tethering and operating the camera from the computer will help to a degree with your current floor/tripod/head only if there is no physical movement in the area between shots. That is, if you remain seated in the same position at a desk and operate the system from the Mac for all of the shots in the sequence without moving you'll see some improvement.

  • Thanks very much for your help! This all makes sense. I will see what I can do to improve the situation. – user1118321 Mar 11 '18 at 3:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.