Copy link to clipboard
Copied
Hi,
after years of using CS6 I hope 26.2 has the tools I have waited for.
I need to select at least three points on the source photo and then the same points on the target photo and hit the 'go and orient and scale distort' button to have the one photo married to the other one.
typical use, I have two photos of a brick wall that need bringing together, I could take much time to rotate, scale, distort, keep turning layer opacity up and down, try again, until eventually it fits, and my trick is to create a layer, link it then trace a feature, turn off the visibility, repeat for the destination picture, turn off vis on the image and marry the lines drawn, any move of the line takes the photo with it. turn on vis and they marry, or so we hope, but moving the dustort tool is a nightmare , what matched on point 1 then moves, like three puppies on elastic leads, still a lot of faffing about.
Such as I describe above has got to be doable by some clever developer coder. and boy oh boy wouldnt it save time.
Merlin
Copy link to clipboard
Copied
Have you tried Image > Auto-Align Layers?
Please post sample images and corresponding screenshots (including all pertinent Panels) to illustrate the process you are trying to automate.
As a CS6 user you are likely not be familiar with Smart Objects, and those do not magically achieve what you want, but they at least would help avoid unnecessary image degredation for repeated transformations.
Copy link to clipboard
Copied
Yes, a clear case for Auto-Align Layers.
I can't guarantee how well it works for brick walls specifically, it depends entirely on circumstances. Auto-Align works by pattern recognition, and it tends to focus on large-scale patterns rather than small-scale patterns. A regular grid is also not a typical use scenario - but still, try it and see.
Make sure to correct lens distortion in advance, if there is any (barrel or pincushion).
Copy link to clipboard
Copied
It certainly seems difficult to mathematically determine the transformation of a rectangle (as the bounds of a Layer always are rectangular) depending on the position and target position of three points if the difference is not limited to scale and rotation.