Over the past week whilst trying to catch up I decided that I should move on from trying Projection mapping in Blender, tough it is the weaker skill of mine out of what I have learnt so far I figured that I would attempt this in both Blender and After effects, tough in hindsight Ishould have used the same image.
Now to speed up the test I decided to only include five materials, meaning that I missed the roof on purpose as I couldn’t work out how it should look (that explains the grey area)
= Blender VS After effects =
In Blender this process was reasonably difficult, however now IUnderstand how it works I may be able to do this faster next time. As Blender has a 3D render engine (two in fact counting both Cycles and Blender Render) the effect created is true 3D. So projection mapping is using an image as a material to project onto 3D geometry. The initial set up of a project is key as the settings regarding camera lens size etc have to match or else it’ll all look a wrong either in aspect ratio or barrel distortion etc. Basically the idea is to match the camera.
Once the settings match, I added a cube and aligned both that and the camera to match the shot. Then, separated the faces that made up the cube and deleted the face closest to the camera and the top face.
Next it was time to go into Photoshop and separate an image into parts to work as the materials for the mapping.
After importing te now cut up images it is time to bring it all into Blender and making sure I was using Cycles Render I would use the nodes to connect up the various textures and shaders to make up a new material then rinse and repeat the process until I had as many materials as I had faces to work with.
After a lot of tinkering and matching the faces of the cube up, I would apply scale and rotation then enter edit mode press ‘U’ and then select project from view after pressing ‘0’ on the Numpad to toggle viewport to the active camera.
Unlike Blender, after effects lacks a 3D render engine meaning it can only simulate 3D though it may do a good job it is important to note what each software can be used for. To try learning this I watched a few tutorials on how to do camera mapping and it seems to be a little more fiddly to set up but gives faster results. The faster render times are likely due to lacking the 3D engine, where as Blender has to render out the geometry of the 3D meshes and then apply the material textures, after effects doesn’t it works from 2D material instead.
Where I had to build up a scene in Blender, in After effects I just had to create a few solid shape layers, then add the grid effect, rotate the grid to match the scene. Then add a new camera that matched the lens that the image was shot with. Next I added a spot lamp and duplicated the image I was working with. The next step is to give the lamp, duplicate image and camera the same position in 3D space and scale the image accordingly. After changing the solids to only accept light data and making sure the lamp casts shadows, pre-render the scene and it seemed to work. Next is just adding camera movement to the shot to try selling the effect.
(Add after effects test… Next time don’t forget to bring it back from digilab)
Unfortunately this effect didn’t work quite as planned as I had set the planes too low and only used two rather than four like I should have done. Also the shot had curved fences that also affected the final product.