Hello,
I was wondering if I have one big project with 2500 photos + scans (multiple streets), all aligned together to one component, and If I create multiple models (one model per FOV, defined by reconstruction region) from this one alignment - Is RC “smart” enough to use only photos wich are used in this particular reconstruction region or it calculates depth maps from all photos from the project?
In theory this should be faster than separating photos and creating new project for every FOV.
Thanks 
its pretty good about only using photos it needs in depth maps.
on the other hand you will grow your cache folder quite a bit larger this way. and you can’t clear it until your done.
I can post alignment result when I’m done with field work. For cach folder I will have M2 ssd 2TB. As I see it, I can align all the photos, create reconstruction region only for FOV that I need, full process it, and after that FOV is done, delete the cache, and proceed with next one.
What do u think?
What exactly would use up cache space? Are there also temp files for model creating? Because the depth maps will only be calculated once - if another model in the same component (same alignment), then RC will use what’s already in the cache and only add what’s not there yet.
you don’t want to delete the cache because it will reuse depth maps from the other views.
the problem is that each time create the models they keep adding up on top of the depth maps. so it grows each time.
you can get around that by doing all the depths maps first. and then backup the cache. and you can go back to that one each time you create a new model.
but if your not running out of space. then don’t worry about it. and just clear it all when done.
Wouldn’t it be nice if RC would sort the cache files or at least put a marker in the name, so that they could be deleted selectively? I think there is already a feature request about that…