Meshroom Photogrammetry Software and Instant Meshes

Meshroom has come a long way is pretty amazing for what it does. You can 3D scan just about anything with a good enough smartphone camera. Combined with InstantMeshes retopology software (also free) you can get game-ready models into Unity. Not to mention Meshroom is open source and free software.

Link to Meshroom: https://alicevision.github.io/?fbclid=IwAR0-HRMDxm59DMu0R2q9X4c8TLdCfRv-x0MpxDh3IeOuVbz4BlgWvPSCBtE



Link to InstantMeshes. https://github.com/wjakob/instant-meshes

 

I have experimented with both Meshroom and VisualSFM+Meshlab and have found that I can get considerably better results with VisualSFM then with Meshroom.

My attempts to tweak Meshroom for better quality usually either end up blowing the processing time way out, or cause it to fail mid way through some steps (particularly the DepthMap step).

Meshroom's interface is far more intuitive though.

 

Thank you for the VisualSFM information. I'm going to look into it and give that a try too.

 

InstantMeshes looks pretty amazing.

 

It's cool and all, but besides things smaller than a car and that won't be animated, how useful is this? Seems like the time put into cleanup and optimization might still be close to time involved to just make a clean low poly blockout.

I could see this making sense in a large studio where a small team of people just churns out dozens of tertiary props with higher complexity than they could sculpt in the same time, but for the small timer...? Any of you guys using stuff like this on the regular?

 

Yeah it is. I think it's basically like a standalone Zbrush ZRemesher.

 

It's useful on a by-use-case scenario. Depends on the needs of your project. If you were to look at the Book of Dead environment assets those were all produced this way, with photogrammetry and they didn't have a very large team on that project. It's also helpful for folks who like to traditionally sculpt their characters as a concept. Once you've got that sculpt you can scan it in and you have a reference to model your game-ready off of. It doesn't differ far from the traditional workflows...only big difference is the 3d scan is pretty much your high poly. In the traditional workflow you'd still need to do your high poly character, then your retopology (game-ready mesh), then your UVs and texturing. Not too much really changes when using a 3d scan, at least in regards to workflow.