Feature Request: Focus Stacking

[OldRadioGuy]

Some photographers now use a technique called Focus Stacking to achieve phenomenal depth of field in their images.

The process is achieved by shooting several versions of a scene, each exposed with manual focus set for a different distance, and then combining those shots as layers in a photo editor.

The processing in the editor then requires:

1. automatic alignment of the layers;
2. automatic merging of the focused sections of the layers (using masking).

Here's a link to a video showing how this is done: https://www.youtube.com/watch?v=yqAXR94X0T8

Has anyone here ever tried to do this in PhotoLine?

It seems to me the two automatic features would be of great benefit to photographers using PhotoLine.

Bob

[SCX]

At least auto-align would be a great feature, because it is helpful for a lot of techniques. Panorama, group images (at least on person always will have their eyes closed), DRI/HDR, removing tourists which always will run into your images ...

For the example video, you would not need auto-align, because of the tripod. But you would need a auto-de-ghost-feature, because of the waves like in DRI-Software.

I guess is would be very difficult to develop such a feature. Maybe there are opensource filters which can be implemented in PhotoLine like the great healing tools or the liquid rescale tool?

[Hoogo]

I've done this one manually in PL.
8 photos taken with a tripod, 500mm/6.3 mirror lens and all my M42 spacers. That DOF was really narrow.

[greenmorpher]

Wonderful image, Hoogo.

Can you explain a little how you isolated the sharp bits and blended them?

Cheers, geoff

[Hoogo]

Basically I've loaded all my images and sorted them by their DOF. My tripod and the attachment for the camera weren't the best, there was always a little movement between each shot. But luckily that was always just a little tilt, I did not need to rotate or rescale the pictures.

I made all layers but the background invisible, then I surely turned layer modes to "difference", enabled one layer at a time and aligned it to the background. I think I focused on the flare in the eye. Then I turned all layer modes to normal again.

Next I enabled background and first layer only. Clicking on the eye again and again and decided which parts of the frog got worse and erased them simply using the eraser. Then I added the next layer and so on...

I think I have somehow treated the ugly bokeh of the mirror lens, but I can't remember anymore.

[greenmorpher]

Many thanks for the file, Hoogo. Interesting stuff.

Not quite sure of the procedure here. So when we look for "Difference", which part of what do you erase?

And if PL cn find the differences, should it not be a smallish step to automatically erasing all else then flattening the image? Or am I being simplistic here?

Cheers, geoff

[Hoogo]

Not quite sure of the procedure here. So when we look for "Difference", which part of what do you erase?

And if PL cn find the differences, should it not be a smallish step to automatically erasing all else then flattening the image? Or am I being simplistic here?

All done manually.
In the attached file you find a 100% crop and only 2 layers. The blending mode of the upper layer is "difference". There was quite some movement of the camera, so the appearance of the result looks more like a strange 3D-Effect. Now choose the layer tool and move the upper layer to a position where you get maximum darkness. Now these 2 layers are aligned.

In the original I turned off the upper layer and repeated the process for all layers.

Next turn the blending mode of the upper layer to "normal". Click on the eye symbol in the layer palette again and again and see where the picture improves. Here some areas in the bottom right corner look better in focus. Erase all the rest.

[maliha]

I guess they are only acting in 16 bit if they are so set. How can I make the default of new adjustment layers 16 bit instead of 8 bit? It's probably rather psychological rather that specifically real, I suppose, but it would make me feel better that my images start 16 bits, and are modified only in 16 bits, and saved in 16 bits.

[bkh]

I guess they are only acting in 16 bit if they are so set. How can I make the default of new adjustment layers 16 bit instead of 8 bit? It's probably rather psychological rather that specifically real, I suppose, but it would make me feel better that my images start 16 bits, and are modified only in 16 bits, and saved in 16 bits.

When your images are 16 (or 32) bit, adjustments always work in that bit depth, regardless of the bit depth of the adjustment layer. The bit depth of the adjustment layer only refers to the mask which is applied to the adjustment. So if your mask is just black and white, you can safely use "monochrome", and you only need a 16 bit layer if you really need more than 256 mask values (e.g., for a very wide transition). I therefore think that 8 bits is a good compromise which will fit (almost) all cases.

Cheers

Burkhard.