Mesh to Volume: Changes in Fog Volume #107279

Erik Abrahamsson · 2023-04-24T03:55:38+02:00

Erik Abrahamsson commented

2023-04-24 03:55:38 +02:00

This patch fixes a few problems with the current implementation of Mesh to Volume.

Main Problem

Currently Mesh to Volume creates a volume using openvdb::tools::meshToVolume which is then filled with the specified density and the class set to Fog Volume.
This is wrong because meshToVolume creates a signed distance field by default that then needs to be converted to a Fog Volume with openvdb::tools::sdfToFogVolume to get a proper Fog volume. Here is the description of what that function does:

The active and negative-valued interior half of the narrow band becomes a linear ramp from 0 to 1; the inactive interior becomes active with a constant value of 1; and the exterior, including the background and the active exterior half of the narrow band, becomes inactive with a constant value of 0. The interior, though active, remains sparse.

This means with this change old files will no longer look the same. I don't see any way around that as the options for external band width and not filling the volume is removed (they don't make any sense for a fog volume).

Before

The current behavior, visualized with a Volume to Mesh node to the left and the volume to the right.

After

The most noticable difference is the gradient looks better and is adjustable using the interior band width.
Scrubbing the Volume to Mesh Threshold between 0 and 1 shows the size of the gradient.

This patch fixes a few problems with the current implementation of `Mesh to Volume`. ### Main Problem Currently `Mesh to Volume` creates a volume using `openvdb::tools::meshToVolume` which is then filled with the specified density and the class set to Fog Volume. This is wrong because `meshToVolume` creates a signed distance field by default that then needs to be converted to a Fog Volume with `openvdb::tools::sdfToFogVolume` to get a proper Fog volume. Here is the description of what that function does: > The active and negative-valued interior half of the narrow band becomes a linear ramp from 0 to 1; the inactive interior becomes active with a constant value of 1; and the exterior, including the background and the active exterior half of the narrow band, becomes inactive with a constant value of 0. The interior, though active, remains sparse. This means with this change old files will no longer look the same. I don't see any way around that as the options for external band width and not filling the volume is removed (they don't make any sense for a fog volume). #### Before The current behavior, visualized with a Volume to Mesh node to the left and the volume to the right. ![bild](/attachments/646ccccf-2757-47b3-bba1-9cc1444578fe) #### After The most noticable difference is the gradient looks better and is adjustable using the interior band width. Scrubbing the Volume to Mesh `Threshold` between 0 and 1 shows the size of the gradient. ![bild](/attachments/2dba0f6d-26ee-4d40-9f08-6b3f574a433e)

bild.png

383 KiB

bild.png

322 KiB

❤️ 1

Erik Abrahamsson added 1 commit 2023-04-24 03:55:58 +02:00

8828c6f8a7 Mesh to Volume: Changes in Fog Volume

Erik Abrahamsson requested review from Hans Goudey 2023-04-24 21:53:28 +02:00

Hans Goudey requested changes 2023-04-25 04:05:03 +02:00

Hans Goudey left a comment

It's not clear why the sdf volume function is changing in this commit-- it's getting a new option and a new transform argument. Both of which seem reasonable, but again don't really need to be part of this PR. Sorry if it feels like I'm playing dumb, but I think it will be helpful/important to resolve this process thing so we're on the same page.

source/blender/geometry/intern/mesh_to_volume.cc Outdated

						
				@ -96,3 +96,1 @@

				  const float voxel_size = approximate_volume_side_length / res.settings.voxel_amount /

				                           volume_simplify;

				  return voxel_size;

				  const float voxel_size =

Hans Goudey commented

2023-04-25 03:52:14 +02:00

There's a lot going on in this statement. Do you think it could be split into two, with a descriptive variable name?

Erik Abrahamsson added 1 commit 2023-04-25 19:24:03 +02:00

590415278f Revert some of the changes and add comments

Erik Abrahamsson requested review from Hans Goudey 2023-04-25 19:28:20 +02:00

Hans Goudey approved these changes 2023-04-25 20:22:08 +02:00

Hans Goudey left a comment

Looks good now, thanks for following up. For the commit title, I'd suggest something like "Volumes: Generate proper fog volume in Mesh to Volume node" to be a bit more specific.

source/blender/geometry/intern/mesh_to_volume.cc Outdated

						
				@ -75,2 +77,2 @@

				  const float volume_simplify = BKE_volume_simplify_factor(depsgraph);

				  if (volume_simplify == 0.0f) {

				  const float simplify_factor = BKE_volume_simplify_factor(depsgraph);

				  if (simplify_factor <= 0.0f) {

Hans Goudey commented

2023-04-25 20:11:39 +02:00

Feel free to commit volume_simplify -> simplify_factor and 0 -> 0.0f separately without more review, but let's avoid combining the cleanups with a patch with functional changes.

Feel free to commit `volume_simplify` -> `simplify_factor` and `0` -> `0.0f` separately without more review, but let's avoid combining the cleanups with a patch with functional changes.

Erik Abrahamsson added 1 commit 2023-04-25 21:40:13 +02:00

b2496e1f63 Remove even more unnecessary improvements

Erik Abrahamsson added 1 commit 2023-04-25 21:51:02 +02:00

834826e621 Merge branch 'main' into mesh-to-vol-update1

Erik Abrahamsson merged commit 700d168a5c into main

2023-04-25 21:56:20 +02:00

Erik Abrahamsson referenced this issue from a commit

2023-04-25 21:56:20 +02:00

Volumes: Generate proper fog volume in Mesh to Volume node

Erik Abrahamsson deleted branch mesh-to-vol-update1

2023-04-25 21:56:21 +02:00

Iliya Katushenock referenced this pull request