Comments

https://www.facebook.com/ted.whitten/posts/10211017163609195?hc_location=ufi

A rough outline of my process -

1. Upscale the photo to 300 dpi, approximately 8x10 size, and do all necessary touch-ups, repairs, color adjustments etc before starting on the conversion. (Working on a fairly high-res image is always beneficial for multiple reasons, but being consistent and always doing conversions on images that are roughly the same size helps you learn and develop your process, and get a feel for how much you need to move each layer, based on how many layers you expect to do.) Also, see step 6 where I mention that you should make sure to start with an image that has "junk" on the sides, that can afford to be cropped a little.

2. Make a copy of the original and lay the two identical images side by side, with some empty space to work with between them. The original image serves as the "right eye image" and remains untouched. All changes will be made to the "left eye image" to alter it, to simulate a left eye perspective. (Note .. While working, I do keep the "left eye image" on the right side, so I can easily cross-view whenever I want, during the conversion process, to see my progress)

3. Working in Photoshop with the selection tools (usually the Polygonal Lasso Tool) I manually select around whatever is in the closest foreground (in this case, Toby's right leg). I imagine the scene being chopped with multiple vertical planes of glass that come straight down and slice up the scene into multiple layers, and try to make each selection based on the next "slice" and what it includes, so the individual layers are not whole people or objects, rather, they are "slices" of them, about 1 inch thick ... in this case, approximately 1 to 2 inches thick in depth. In this image I did a total of 25 layers to cover the entire depth front to back. Ideally I would spend more time and do a larger number of layers, but in this case I felt sort of rushed to get this done, so I just did 25. The first selection included a slice of Toby's right leg about 1 inch thick.

4. Shift that selected area to the right by 1 to 2 pixels. In this conversion, most of the shifts were 2 pixels, horizontally. I do the shift by holding down (on a Mac) Option + Command and tapping the right arrow key twice. This moves the selected area to the right by two pixels, and automatically fills in the "blank" area left behind, with pixels identical to the pixels directly outside the selected area. This is a convenient way to fill in that empty space automatically, and usually does not create retinal rivalry. Whether to move it 1 pixel, 2 pixels, or more, is somewhat guesswork, but I've found that I get a good result starting with an image the size I described, doing 25-30 layers with 2 pixel shifts each, or 50-60 layers of 1 pixel shift each.

5. After this, I make sure "Add to Selection" is selected, before selecting the next section deep. "Add to Selection" enables you to select a larger area of the image while still keeping the original selected area. So you study the image, decide what to select that represented the next 1-2 inch thick "slice" of the scene, and add that to the selected area, then move that. When you move that, now, you're moving not only the new area, but the original selected area as well, so now you've moved the new area once, and the original area has been moved twice. When you're done with the whole image, the original area will have been moved the greatest number of times. Repeat, and keep doing this over and over as you work your way "deeper" into the image, one slice of depth at a time. In some places, depending on how the people or things are arranged, you may decide you want more layers for a certain area (like the faces, for example), so you do "shallower" slices, and only shift them 1 pixel at a time instead of two. Or you may hit an area that requires more pronounced depth, so you may choose to shift by more than 2 pixels at certain places. Eventually you get to the background. Once you do ...

6. Flatten and save t

As to my "fairly quick" comment, you're right .. 4-5 hours isn't exactly quick. I just mean, doing quality 2D-3D conversions, in my experience, takes a lot longer than that. I rushed through this one because I wanted something to give as a gift.

Thanks Tom. Yes, I've done 2 depth map conversions so far, and been very unhappy with the results, even though I spent like 10 hours on each of them, and made the depth maps extremely accurate and detailed, with various gradients etc. After I apply the depth map the image is so screwed up it requires another 10 hours just to clean it up. Very frustrating. Some people seem to do very well with depth map conversions, so I wish I could find someone who is good at it to instruct me better.

Looks very natural - awesome! I tried playing around with depth maps for the very first time yesterday. It works, but it's not pain free, a lot of tweaking is required. I don't think I could call a 4 to 5 hour job "fairly quick". Obviously you're doing what you love.

Thanks very much Maecky!

Thank you Maecky!

Wonderful!

Fantastic!

Visible just beyond the tree line in the image center, are the two cabins - the one we rented that week, and the one my grandparents used to own. The hike up to this lookout point was a popular annual event for our family.

I framed this photo to intentionally include the cabin we were renting that week (left of center), and the smaller one (between the trees) which two generations of our family grew up enjoying in the summers (my grandparents owned it from approximately 1950 to 1995).

The protective glass over this street mural by the famous graffiti artist 'Banksy' was actually vandalized - shattered, but not broken (https://www.ksl.com/?sid=28221288). Ironically, this shattered look was what made it an interesting stereo subject to me.

Here's an 11-minute YouTube clip showing the full scene this image represents: https://www.youtube.com/watch?v=VFBYJNAapyk

Thank you very much Wim!

Nice serie!

One frustration I recall however, when taking this photo, was that I couldn't get more of the stairway into the picture. It was about twice as long as what you can see. But I settled for what made the nicest overall composition.

Thank you Charles. Stairways are often a great 3D subject, and when I saw this one I was thrilled. I love the way the colors are split, red on the left, green on the right.

Beautiful, great depth.

Yes, indeed. Very peaceful. As you can see from the surface of the lake, she wasn't exactly battling fierce winds. I was surprised she was able to get the boat moving at all. ;-)

very peaceful feeling