I have been using RC for more than a year for 3D scanned people.
Since a couple weeks I have been getting a lot of problems when I reimport the cleaned up scan for retexturing.
I use 2 sets for my 3D scans, first I work with a photoset that includes projection patterns shooted at the person.
For retexturing the reimported mesh I use a photoset without this projection to achieve the final color.
These 2 photosets get taken in a few milliseconds to make sure there is a minimum movement possible between the 2 sets.
However, since a couple of weeks a lot of my scans come out very blurry after retexturing, resulting in dramatic results which I can not use to finish my 3D scan projects.
Is it possible that since a last update RealityCapture became less tolerant between movement? (read->mesh is not in 100% the same place as the 2nd photoset would render a mesh, thus movement)
I have tested the same 3D scan in other software and their the 2 sets remain sharp so I assume that RC is less tolerant for movement.
Any way I can tweak these settings to allow me to make sharp results, more tolerant way of working?
I instruct people to stand as still as possible but breathing, blinking etc results in minimum movements anyway.
Also, I noticed that the settings for imported models state PREVIEW quality when I check the selected model's settings.
May this also be a cause of the problem?
I can't seem to change this setting however.
Your help is much appreciated, thank you!
Below are some screenshots to show the results and my settings.
Please sign in to leave a comment.