# Demonstration in 2D of what's going on: ![Demonstration][2]
Mathematically it should be correct to render with a orthographic camera, and later transform the image. I think this is the easiest method when you have multiple objects (so you do not have to [add modifiers to each][1]). First you should shear in 2D (skewing) the rendered image, then (if needed) stretch to get the correct aspect.

0. Original:

![Original][3]

0. Sheared:

![Sheared][4]

0. Streched:

![Streched][5]

Most [raster graphics editors][9] have the tools to preform the above transformations. This can also be done in the compositor using a [Blend texture][7] and the [Displace node][8]:

![Node setup][6]

  [1]: http://blender.stackexchange.com/questions/319/
  [2]: https://i.sstatic.net/tBcPS.png
  [3]: https://i.sstatic.net/ZXtYX.png
  [4]: https://i.sstatic.net/VIpAn.png
  [5]: https://i.sstatic.net/D9wpY.png
  [6]: https://i.sstatic.net/61JVJ.png
  [7]: http://wiki.blender.org/index.php/Doc:2.6/Manual/Textures/Types/Procedural/Blend
  [8]: http://wiki.blender.org/index.php/Doc:2.6/Manual/Composite_Nodes/Types/Distort#Displace_Node
  [9]: https://en.wikipedia.org/wiki/Raster_graphics_editor