Time Lapse with 2.5D effect

[Bob]

Here is an excellent video combining time lapse with parallax shifts using the 2.5D technique. All the work was done with still images. To get the parallax effects, the stills were divided into separate planes in Photoshop and the layers were arranged in 3D space in After Effects and the virtual camera animated. Very well done! Enjoy!

[momoffduty]

Nice find Bob! Love the parallax work. The narration was well done.

[Steve Grisetti]

Beautiful photography and parallax technique!

[Peru]

Indeed, very well done.

[_Paz_]

This is truly beautiful country and a fabulous job of editing the stills together. I was particularly impressed by the movement over the crevasse, around 124. It really feels like you are slowly looking down and about to step across the edge.

The concept of doing something like this is new to me. Had no idea it could be done from stills, in fact. And it seems like some of the clouds moving footage must be video instead of stills. I guess that is a time lapse group of shots. I'd love to see more info on how to go about a project such as this.

[BuddyB]

Quite dynamic and well done, but pardon my ignorance, do not understanding the meaning and what Parallax is and how it works in the production. I'm also curious, we have the black border at the top and bottom, which to me means a special camera..and why all of a sudden do I feel Parallax coming on?

[Bob]

I'm also curious, we have the black border at the top and bottom, which to me means a special camera..

According to the author's post, this was shot entirely as still photos using a Nikon D7100 DSLR camera. The photos were prepared in Photoshop and brought into After Effects to produce the video. The ultra wide frame aspect ratio gives it a panoramic feel. This could have easily been done by cropping the images in either Photoshop or After Effects.

pardon my ignorance, do not understanding the meaning and what Parallax is and how it works in the production.

Parallax is the displacement in the apparent position of an object when viewed along two different lines of sight. Closer objects have a larger parallax than distant objects. This is the basis for stereopsis (3D vision). The brain uses the parallax from the left and right eye views to gain depth perception and estimate distances to objects.

When you take a photograph, you end up with a 2D representation of a 3D scene along a single line of sight. Single line of sight means no parallax differences thus no depth. It's a flat image. You can move the viewing position, but it remains a flat image. What the author did here was divide the photos into multiple layers, each layer containing elements of the photo at a particular distance. The most distant objects being in the bottom layer and each subsequent layer containing closer objects than the proceeding layer with the closest objects in the top layer. When brought into After Effects, the layers can be assigned positions along the "z" or depth axis in After Effects' 3D space. By assigning the layers different depths, the After Effects virtual camera can be moved in 3D space and parallax differences can once more be observed. Instead of an obvious pan of a flat photo, we have a pan in a 3D environment. It has a more dimensional feel.

[_Paz_]

Bob,

Sorry, I still don't get it.

each layer containing elements of the photo at a particular distance.

Does that mean portions of some layers are erased? For example, in the opening image, would the figure and the first 'triangle' of land at the bottom of the image be on the top layer with everything above it, up to and including the sky, be erased so it is transparent so that areas at further distances could show through from behind (below)?


If so, wouldn't portions of other layers, I'm guessing other 'triangles' of land, need to be enlarged so they can fit behind and offer a little room for peeking over the edge?

(I'm not sure my question makes sense! )

[Bob]

Your question makes perfect sense.

Yes, you've got the idea. Each layer contains things at roughly the same distance with everything else erased or masked off -- foreground objects (that triangle of land you mentioned) in the top layer, background objects (sky and distant mountains) in the bottom layer, and intermediate objects in the middle layers.

Because we will be moving the camera in 3D space, there will be some areas that become visible where the image was erased or cut out. Other areas may overlap and need to be feathered to blend seamlessly. There are ways to handle the blank areas. One way is to scale the layers so that there will be overlap that hides the missing areas. Another is to clone in something to fill in the exposed areas. In cases where you have two photos -- one with a foreground object and one without -- you can extract the foreground object and place it over the background photo. What technique you will choose will depend on the individual photo and what you are trying to do with it. Cutting out the pieces and any clone work will be done in Photoshop. Scaling will need to be done in After Effects as you won't be able to tell how much to enlarge until you position the layers in 3D space and move the camera.

This tutorial should help you:

[youtube]http://www.youtube.com/watch?v=IqYuoznrhf4[/youtube]

[momoffduty]

Does that mean portions of some layers are erased? For example, in the opening image, would the figure and the first 'triangle' of land at the bottom of the image be on the top layer with everything above it, up to and including the sky, be erased so it is transparent so that areas at further distances could show through from behind (below)?


If so, wouldn't portions of other layers, I'm guessing other 'triangles' of land, need to be enlarged so they can fit behind and offer a little room for peeking over the edge?

Yes. Usually you would have to make the background larger in AE than the foreground pieces. If not, you would see the edges. The trick is to place the cut out pieces so they appear in proportion.

When I cut up a photo for this effect, I think about how I will move the camera through the scene, how much room I will have and cloning work. For example: a photo of person standing in front of a tree with a mountain in the back. For the tree layer I would have to clone in where the person was standing. Same thing for the mountain. I would clone in where the tree was. Not all photos are good to work with. Sometimes you can take a person from one photo and use a completely different background. The lighting and the tone would have to match to sell the realism.

EDIT: Looks like Bob posted the same time as me. Great explanation Bob.

[_Paz_]

Great explanations from both of you! Thanks!!!

I don't have time to watch the video now, but I'm sure it will go a long way towards explaining the process. Thanks for posting it, Bob.