I'm developing a tour that has an object pano it. I want that in as large as possible a resolution to show on HTML5 too, so I'll just chuck all my image frames into PTGui and spit out an equirectangular file that equally spaces the images along the horizon. I then convert to cube face and then simply limit the field of view to the resulting width of one frame (360 / num_frames). Together with a mod of the default object pano template this works really well. Except for the resolution -- the more frames I have, the smaller the resolution of course.
Now, with this technique, however, I don't need the top and bottom cube face, so I just leave these two "useless" cube faces out (and tell KRpano not to complain by using showerrors="false"), but does that mean I can acutally safely use larger cube faces for the remaining 4 assuming that no memory is used for the missing top and bottom faces, or is that not the case?
What's your thoughts on this?