First, my English can be strange using a translator.
I studied the process of making photoscan to game asset.
I don’t know if I can post it here, but I have no place to talk and I write here because I studied it based on this program.
417 pictures were used and no control points were created.
In the previous 1.1 version, Mesh reconstruction continued to fail, but it was successful when using Beta 1.2.
It first produced 5.6 million meshes and reduced it to 2,000.
It used the Simply tool.
There are some cases where a message like this comes out if it is too shortened, but the mesh is generated, so you could ignore it.
Texture Reprojection was used to create Low poly mesh and Low poly Texture.
It attempted to convert from Diffuse to Albedo with the Agisoft’s De-lighter program.
I couldn’t make the pants perfectly.
Low poly mesh extracted from RC was not a neat form.
So I organized it neatly.
The UV was also torn like the picture above and re-produced to increase resolution.
Using the RTT, which is one of the 3ds max, the De-light working Texture has been moved.
I tried to adjust the color using Photoshop.
The reason why I used Photoshop first was that I wanted to change the pattern like the picture above.
Based on the mesh extracted and modified from Reality Capture in the substance painter, it was made based on high poly criteria initially and produced normal map.
Due to lack of detail, Substnace Designer produced normal map using diffuse map.
We designated the area using substance painter and partially adjusted the color.
At the same time, roughness and metal were added and adjusted in part.
Finally, it was added to Unreal.
It’s on Sketchfab. It was used for about 20 hours until completion.
I had a lot of worries while producing this process.
In the future, we are considering whether to actively use substance designer or produce texture, which is produced only by high poly mesh, and produce texture together.
I really need someone to look at this process and give me an opinion. I’d appreciate it if you could help me.