I can understand why you make the viewer the same aspect ratio (close anyway) as the input images (so that you have fewer pixels to perform hole filling on around the edges when moving), but it seems to make less sense when in global camera view and geometry is arbitrarily lopped off when there's plenty of screen left for it to reside in when rotating the model.
I'm trying to look forward and imagine these synths embedded in Bing Maps 3D - imagining the areas outside of the canvas where the photos are drawn as possibly being filled in by the map/globe.
Also, why abandon the extracted color background outside of the photos' canvas when in global view if you aren't going to draw the frusta and point cloud outside the smaller canvas?
In other words, if you aren't using that space for anything that you don't normally, why change it to charcoal grey?
I do understand that this is a beta, but wanted to ask what your thinking was here.