Skip navigation
Currently Being Moderated


Mar 15, 2013 7:50 PM

My apologies, but I'm kind of a left brained guy and sometimes it's hard to wrap my mind around right brain spatial problems that others find laughably easy.


I'm interested in creating this simple box with four walls through which one could view a video at the far end.  I don't need all the little videos on the side, I am just unsure how to create that simple box.


Would I extrude that in 3d somehow?  Can I construct it all at once rather than aligning one wall at a time?  I'm not too good with parallax and vanishing points yet.


What basic workflows do I need to master to create this simple box in 3D?


Thanks so much!  Rocket science to me right now.


Here is the underlying video from which it was derived; it was constructed in another program, but I'm certain this is very straightforward in AE.



Thanks!   Matt Dubuque



Screen shot 2013-03-15 at 7.43.03 PM.png

  • Currently Being Moderated
    Mar 15, 2013 8:58 PM   in reply to Matt Dubuque

    Make an open cube out of 5 3D layers (leaving off the sixth side, the one in front). The back layer is the one playing your movie. Move your camera backward out of the cube.


    Since you seem to be struggling with the basics, and you've said that you don't like to read, I very strongly recommend that you get one of the comprehensive video training courses for After Effects and watch it all the way through. Spending a few days doing that will really help you.


    I give several recommendations here: or-after-effects.html


    I especially recommend Chris & Trish Meyer's Apprentice training on and the Learn By Video training on video2brain, in that order.

    Mark as:
  • Currently Being Moderated
    Mar 16, 2013 9:23 AM   in reply to Matt Dubuque

    > Also, I think I will somehow need to use the transform tool to transform the sides of the wall of my screen shot to get that sense of perspective.

    You only need to do things like that if you're faking 3D with 2D effects and transformations. If you actually build a 3D box and move a camera, perspective comes for free. That's the point of working in 3D.

    Mark as:
  • Currently Being Moderated
    Mar 16, 2013 10:38 AM   in reply to Matt Dubuque

    Try this.

    1. Create 5 different colored solids in a new HD composition in AE.
    2. Make solid 1 through 3 1600 X 800 pixels, make solid 4 and 5 1600 X 1600 pixels.
    3. Make all solid layers 3D.
    4. Ctrl/Cmnd + A to select all layers
    5. Press A to reveal the anchor point.
    6. Set the z value of the anchor point for Layers 1 - 3 to -800 pixels
    7. Set the z value of the anchor point for Layers 4 & 5 to -400 pixels
    8. Rotate layer 2 90º on the Y axis
    9. Rotate layer 3 270º on the Y axis
    10. Rotate layer 4 90º on the X axis
    11. Rotate layer 5 270º on the X axis
    12. Add a two node camera to your scene with a 35 mm lens
    13. Select Layer>Camera>Create Orbit Null
    14. Add another null to the scene, make it 3D and name it CubeController
    15. Select all of the layers that form your cube and make the CubeController the Parent
    16. Adjust the perspective you want in your 'theater' by changing the Z value of the camera
    17. Adjust the angle of the camera by rotating the Camera Controller null
    18. Add some lights to the scene to emphasize the perspective


    There you go. Five minutes and you've got your 'theater' and you're ready to start replacing the walls with footage or artwork. The best way to do that is to build each wall as separate composition the same size as the solids, add the elements you are going to use for the walls, and then return to the main comp and select a wall and it's replacement pre-comp in the Project Panel and Alt/Option drag from the Project Panel to the Timeline to replace the 'wall' (layer) with the pre-comp.


    It should look something like this:

    Screen Shot 2013-03-16 at 10.34.07 AM.png

    (Did you catch the math used in setting up the box? AE's 3D space is very easy to use once you've done it a time or two.

    Mark as:
  • Currently Being Moderated
    Mar 16, 2013 5:15 PM   in reply to Matt Dubuque

    If you use solids to create a box you just have to pay attention to the sizes They can be anything. They could be comp size but the math isn't as easy to do in your head. I chose 1600 X 1600 and 1600 X 800 so that the offset in Z for the anchor point would be a nice even number. You use half the horizontal width of the solid for the Z value for horizontal walls and half the vertical height for floor and ceiling. That way, when you rotate them the edges perfectly match. It's just simple arithmetic.


    As far as placing a video on the projection screen you can scale the video to any size you want, but for the layer replacement to work the pre-comp must be the same size as the solids you used to model your theater. I'd probably lay the video on a solid and either crop or scale it to fit so there was a little border.


    This is pretty basic stuff if you just think How would I build what I want out of a deck of cards.

    Mark as:
  • Currently Being Moderated
    Apr 2, 2013 4:06 AM   in reply to Matt Dubuque

    Position the anchor points is purely mathematical. Think of a cube with all of the anchor points for each layer exactly in the center of the cube. If the sides were 400 pixels by 400 pixels then 1/2 of the width or height would be 200. Offset the anchor point by 200 pixels in Z and you've moved all of the layers back by 1/2 the width of the layer. Since rotation always happens around the anchor point when you rotate the in increments of 90º they will always end up with aligned edges.


    Now to the problem of your edge seams. First of all, work with even pixel values for the size of your layers. 200 X 200 works, 205 X 205 or 300 X 145 won't work as well. This is because AE does sub pixel positioning and the edges at half pixel values will be interpreted with different values than edges at whole pixel values.


    The second problem is that any time there's any angle to a line or edge the edge must be interpreted to fix aliasing (jaggy edges) that are caused when you try to draw a diagonal line on a checkerboard. It is called anti-aliasing. This gives you different pixel values at the edge with the image is at an angle. After Effects makes it's transparency calculations using straight alphas. In most cases this is by far the best option. When it comes to matching up edges of these 3D planes it isn't the best. It would be better to add the alpha values together rather than use a straight calculation which combines the values.


    Fortunately AE has a blend mode called Alpha Add. It only effects the alpha channels are calculated.  Changing the blend mode to Alpha Add eliminates the problem until you start adding lights to your scene.


    As soon as you start adding lights to the scene you run into additional problems calculating alpha edge values. There are some workarounds to eliminate light leaks in these situations that involve scaling the layers slightly that are nearest the camera so that the problem is hidden. In spite of what your wife says trying to be perfect in every frame isn't a very good way to make a living in the visual effects business. You have to be able to make your scene look right with the least amount of effort so you can move on the the next of the thousands of frames it takes to complete a project. That is the biggest challenge faced by anyone working with AE.


    Unfortunately, unless you use ray traced rendering to extrude a box in AE you cannot fix the problem completely, but by carefully designing your animation to hide these problems you can make things look perfect without being so. Take these two cubes with a point light set in the center. The one on the left is perfect but looks wrong, the one on the left has the back left wall moved 1 pixel to the right and the right front wall is scaled 101% in X to hide the seams. If I needed to spin the cube and add motion blur I would need to pre-compose the spinning cube and add CC Force Motion blur to the scene to hide the seams. It's prety easy to see which version of the spinning cube works better.



    I hope this long winded explanation helps. Take a look at his image. It shows two solids separated by 2 pixels in the center. The Alpha levels of the pixels that are going to match up are shown. The panel on the left shows the edges matching with the blend mode set to Alpha Add. The image on the right shows the blend mode set to normal. The differing shades of blue are caused by combining rather than adding the alpha values of those pixels.



    Mark as:
  • Currently Being Moderated
    Apr 2, 2013 9:25 PM   in reply to Matt Dubuque

    Matt Dubuque wrote:


    One thing I will be doing, which seems implicit from an earlier post you made, is to put a border around each of the sides, which also masks these small issues that arise.


    My question about anchor points was more of a conceptual one, seeking kind of a view from 1000 feet.


    I was wondering how you can tell whether any problem in After Effects is likely to require manipulation of the anchor point in its workflow......

    Putting a border around a file does not help with the anti-aliasing issue very much because a diagonal line, whether or not it is a border is still a diagonal line. Also, when adding lights you may be able to help a bit but the shift and scale options I described work much better and you can animate them to keep the sides toward the camera looking perfect.


    Manipulation of anchor points is required when you want to change the center of rotation, the center of scale, or easily line up layers in 3D space. The anchor point is also a great tool in 2D when doing things like slide shows. You'll find that it's easier to animate scale and anchor point to do a move on a still in 2D than it is to manipulate scale and position. There's not enough time to go into it here, but before 3D in AE everyone who knew what they were doing changed the anchor point (which changes the relative position, and adjusted the scale to move in or out of a photograph.


    The whole key to understanding AE's 3D space is to look at the problems like you were building a house of cards. What is the easiest way to line things up. Unless I'm building cubes or walls nearly all of my layers have the anchor point set to 0 Z space but the anchor points are set to my desired center of rotation, center of scale.


    I hope this helps.

    Mark as:
  • Currently Being Moderated
    Apr 2, 2013 11:37 PM   in reply to Matt Dubuque



    Off topic, but what resources do you have for compositing and green screening? I can make a few suggestions. I started on AE roughly a year ago, and have built a nice library. I have a few titles for greenscreening and compositing. I can list them if you are interrested. Keeping in mind - they are what I have found helpfull, there would probably be better resources out there (or here in the forum of course!), but it might save you a little researching time.



    Mark as:

More Like This

  • Retrieving data ...

Bookmarked By (0)

Answers + Points = Status

  • 10 points awarded for Correct Answers
  • 5 points awarded for Helpful Answers
  • 10,000+ points
  • 1,001-10,000 points
  • 501-1,000 points
  • 5-500 points