The process used in this capture is simple and complex at the same time.
It includes 6 pixel layers, 3 level layers and ACR filter used three times on different layers. Only two basic layer masks are used.
Initial perception:
- Crop
- The sky has to change, there is a potential in there.
- The wall is drab, it needs to stand up more, possibly warmer
- The grass has to change if the wall changes but not the same way
- The car? An argument to remove might be there but it is an integral part of the scene as explained in the initial post. It needs to be there and survive 'intact'.
Process making the selections:
- Duplicate the background layer. Correct color defringe using ACR
- Select the sky and isolate it meaning the sky will be a separate layer with its own adjustment.
- Select the wall and only the wall for the same reason as the sky (Layer)
- Select the grass... Actually I did not. I used the sky selection to create the grass selection (inverted the sky selection) Layer
Working on each selection...
- Sky
- All layers are active
- Create a dust removal layer first and use it to remove the flaws non destructively
- Create a sky level tied to the layer below it. Adjust as needed while looking at the overall effect (hence the all layers active)
- Grass
- The wall layer is not active
- Create a level layer tied to grass layer. Adjust as needed.
- Create a mask on the grass layer, not the level layer. Mask out the fringing and the car area.
- Wall
- All layers active
- Use ACR to sharpen the wall
- Create a level layer tied to the wall layer, Adjust as needed.
- On the level layer create the mask to remove fringing
Done
Are there flaws in There?
Yes.
- One is the line of trees in the background. It seems to give depth to the tree line (layered height) in reality these are the same tree with a fringe. It works with the image so I left that alone.
- Second there is some fringing on the wall due to an imperfect sky selection. Even at 100% this is not visible unless pointed out so, is it worth the effort to correct this? Not in my opinion.
- I am sure you will find more flaws but think 'Is it worth the effort to correct them?'
Anyway, from a so-so capture to a relatively decent image one can learn quite a bit.
Link to PSD file (500MB)