It appears that the optimal texel size and texture quality is determined by an average distance of the camera to the subject for a given lens. Is there any way to force reality capture to calculate and render at a the maximal resolution based on the closest camera rather than the average distance?
There is no way to change the calculation for the optimal texel size, but you can manually make it smaller if you like. Would you like us to make a feature request for different options for optimal texel size calculation?