I’m thinking about this issue for some time and I haven’t found a satisfying solution yet. Now I’m working on a scan of a small church that is why I will use it as an example. Differences in brightness inside are huge. The difference between lightened by the windows wall and the darkest corner it can be even 10 ev. The color temperature varies. I use three different cameras so each sensor reacts differently on tint etc. Possible solutions:
AUTO MODE (shutter speed) - I’m scared of this. In theory, the camera will choose proper exposure and I will pass smoothly between the darkest and lights part of the church. I expect that texture generated this way can give hard passes that need to be avoided. Even if generate texture from point colors, it can give some unwanted results.
HDR - It will produce nice input but… the quantity of the photos can be just too big. The first day I took 1,5k photos. Due to having smooth passes between stacked exposure I usually use more than 3 stacs. Depending on EV scale of the interior I’ll need from 5 to 11 photos for the stacks. 1,5k x 11 and tomorrow the same? 30k photos to process and I’m gonna use information from max 30% of those. Not so efficient.
ONE exposure - Using top DSLR it is possible to achieve 14ev of dynamic range. In theory, picking a middle value and aligning exposure in postprocessing should give usable photos. The problem will be low quality of the details in the shadows. This solution is also far away from being perfect because the size and weight of FF DSLR can be a huge obstacle to use it for an example from a drone.
Even if exposure problems will be solved there is still color temperature inconsistency and tints. I’m sure that there are some solutions in postprocessing that allows for obtaining decent results. Texturing from the only one serie of photos that are made fast to catch lighting conditions? If it is possible… Texturing from chosen photos? This workflow seems to be painful… or maybe there is some software solution I don’t know. What idea do you have? What are you doing in this kind of objects?