Tiles & Splits: A Projector Blend Case Study

This case study assumes some understanding of projection engineering and that the blend is happening on the Media Server side. It is possible to bake blends, but it’s not something I recommend unless you have to. 

Splitting content up after a content render is more efficient then rendering splits at the source.

Say that I’m going to be projecting a 3x1 4k projector array at a 25% blend. This means that my overall canvas is 9600 px wide x 2160 px high. You could call this my “content canvas.” A template for creative folks would just be a blank canvas at that resolution. Typically, at higher resolutions, you may need to separate out your media into “splits” or “tiles” so that you are playing back efficiently on your show machine. 

The obvious thing to do would be to configure your template so that you render out the slices from there. This is the easiest, less thinking required, method to rendering splits – but it’ll take a ton of extra time as the computer needs to process the whole canvas three times, instead of once! In some cases, say in Premiere, that might make sense – but in After Effects, this could really add a lot of render time, particular to high resolution renders. 

Instead of processing your composition three times, you render it once and use a different tool to process it into splits. It requires less thinking to process your composition three times, but it requires less time to process it into splits as a post-creative-render process. Bottom line: It is faster to render the whole surface than pre-render splits in After Effects. 

For doing this efficiently, I recommend FFMPEG – it absolutely rips through a splits process. That said – you can achieve similar ends using Premiere (but will require more QC, as there are greater human error possibilities), or Shutter Encoder (a great GUI for FFMPEG).