Using mask to prevent projecting some detail

[arumiat]

I have an original model of around 10 million polygon model from a scan (A).

I then performed extensive cleanup in Zbrush using sculpting & dynamesh.

I have then re-meshed the dyanmeshed version using Zremesher & subdivided (to 6million polygons) (B).

I would like to project a lot of the very fine detail from the original scan model A onto the model B. However model A has a lot of bad areas that I don’t want to project onto the cleaned up version.

My options it seems:

1. it seems like I can use the Zproject brush & manually go & draw in the details from A to B. It is a slow process however & I feel like the detail doesn’t transfer over well (I’ll post a video)

2. ideally I’d ‘Project All’ from A to B but somehow prevent the unwanted areas from projecting . I tried masking the areas of model A I didn’t want projected, but this didn’t seem to prevent these regions from projecting. I know I can mask model B for areas I want to ‘protect’, but it’s hard to know which areas on B I should protect without seeing model A?

Are there any other tips for using masking to prevent projection of the undesired areas?

The video shows the areas that I do not want to project.

[arumiat]

This is me trying to use the Zproject brush. The original mesh A is 10 million & the one I am projecting to B I have tried at both 6 & 24 million. It seems the brush doesn’t have that strong of an effect.

I have tried increasing the Zintensity to 100 but doesn’t seem to make a difference.

Any tips would be awesome.

[Spyndel]

Hello @arumiat

@arumiat

ideally I’d ‘Project All’ from A to B but somehow prevent the unwanted areas from projecting . I tried masking the areas of model A I didn’t want projected , but this didn’t seem to prevent these regions from projecting. I know I can mask model B for areas I want to ‘protect’, but it’s hard to know which areas on B I should protect without seeing model A?

Masking the areas you want to protect on your target mesh is definitely the way to go. Could you perhaps use the transparency feature to project a ghosted version of the source mesh over top of the target mesh in order to reference the areas you want to protect?

If the color information is part of the determining factor for identifying which areas you want to protect, you could first project the color onto the mesh without sculpting information just to identify those areas for masking, then re-project both the sculpting and color when masked.

Alternatively, you could duplicate your source mesh subtool, fill the duplicate with white polypaint, and paint the problem areas in black polypaint. Then project that polypaint (color only) onto your target mesh, and use the Masking> Mask by Color> Intensity to establish masking in those areas. Then hide that tool, and project the full detail (color and sculpting) from your source mesh onto your target mesh.

[tobor8man]

Just some random thoughts on things to try.

• Delete the parts of the original that you don’t want projected when you project all.

• Hide parts of the original you don’t want projected, set a history point and then use Project History on the new model or use the History Recall Brush for local projection.

• If the topology isn’t important then set a history point on the original, activate Scupltris Pro and locally refine the new mesh in areas where you want the detail, then Project History on the new model or use the History Recall Brush for local projection. (Tomas Wittelsbach uses this method a lot for keeping very fine detail in important areas in his jewellery pieces. Search his Pixologic live streams).

• If the topology isn’t important then decimate the new model, and refine locally with Sculptris Pro as above.

• Break your model up into smaller manageable pieces. Then Decimate, makign sure to check the Freeze Borders option. Remesh locally with Sculptris Pro and Project History as above, making sure the Mask Borders first. Decimate again with Freeze Borders if you want to cut the poly count down after this. If a single subtool is needed then merge subtools and Weld Points. The Freeze Borders ensures a perfect match.

Good Luck.

[arumiat]

Thank you very much both. These were all excellent ideas. As it happened the most elegant solution was simply to project just the colour info to the mesh B first, mask out the bad areas using that, & then project the details. Great.

I have a few more workflow related questions I’ll go ahead and ask here if it’s okay. If I should be asking them in separate threads let me know.

1a) mesh A =10mil polygons & mesh B=6mil polygons so I am losing some small amounts of detail when I project. Is there a way to set a target polygon count for subdivide somehow? If I subdivide B again it will reach 24 million polygons which is overkill & slows everything down a great deal. Was wondering if I can ‘subdivide’ to 10 million to try to match A polgyon count & thus maximise possible the detail from A.

1b. After projecting once from A to B, is there any benefit to projecting a second time (without changing anything else) ? Does projecting again project more somehow?

2. it seems like storing a morph target prior to Project All is a neat way to clean up badly projected areas. I also notice some people use a ‘Layer’ concurrently, but I can’t really figure out the difference between the two, & what extra flexibility it offers. Is anyone able to advise?

Thanks.

[Spyndel]

@arumiat

1a) mesh A =10mil polygons & mesh B=6mil polygons so I am losing some small amounts of detail when I project. Is there a way to set a target polygon count for subdivide somehow? If I subdivide B again it will reach 24 million polygons which is overkill & slows everything down a great deal. Was wondering if I can ‘subdivide’ to 10 million to try to match A polgyon count & thus maximise possible the detail from A.

Subdividing a mesh will always subdivide by a factor of 4. You can, however, use ZRemesher to remesh the base level topology at a slightly higher polycount, which will result in a greater polycount when subdivided to the same level, without requiring an additional subdivision. You can also draw/edit your topology in such a way at the base level to deliver more polygons to high detail areas, and away from low scrutiny/ low detail areas. This will result in those portions of the mesh being much more responsive to detail when fully subdivided.

You can also partially subdivide sections of the mesh at the base level by hiding or masking areas you don’t want subdivided. This will create some fairly ugly topology, but may be a quick and dirty solution if your output doesn’t require clean topology–for instance if you intended to Decimate.

---

@arumiat

1b. After projecting once from A to B, is there any benefit to projecting a second time (without changing anything else) ? Does projecting again project more somehow?

There is no benefit I can think of, but it does introduce new opportunities for things to go wrong

---

@arumiat

2. it seems like storing a morph target prior to Project All is a neat way to clean up badly projected areas. I also notice some people use a ‘Layer’ concurrently, but I can’t really figure out the difference between the two, & what extra flexibility it offers. Is anyone able to advise?

I would stay away from Layers unless you specifically require their functionality. They will complicate and limit what you can do with the mesh.

This video from Joseph has good tips on cleaning up problem areas in the projection.

Good Luck!