OCCT3D
DiscussionsIssue archiveOCCT:Visualization

Archived issue #0024521

Visualization - automatic back face culling is not turned on for Solids packed into compound

Open CASCADEOCCT:Visualizationclosed7 public notes

Search issues

Description

Automatic back face culling (V3d_View::SetBackFacingModel (Visual3d_TOBM_AUTOMATIC)) does not work for Solids packed into compound (StdPrs_ShadedShape::Add()).

Additional information

Back faces culling reduces rendered number of triangles (improve performance) and eliminates artifacts at shape boundaries. However this option might be used only for solid objects, where interior is actually should not be visible from any point of view. Automatic back-face culling mechanism is turned on by default, controlled by V3d_View::SetBackFacingModel().

The following issues have been fixed in StdPrs_ToolShadedShape::IsClosed() (used for definition of back face culling in Shaing Aspect):
- disable culling for free closed Shells (not inside the Solid) since reversed orientation of free Shell is valid case
- enable culling for Solids packed into compound
- ignore Solids with incomplete triangulation

Back face culling is now turned off at TKOpenGl level in the following cases:
- clipping/capping planes are in effect
- for translucent objects and with hatching presentation style

Public activity

7 archived notes

Participants are labeled by their role within this record.

01Author
Patch is ready for review in CR24521 branch.

In addition to described problem (culling is not turned on for solids in compound) this patch also contains back-face culling disabling when clipping plane is in effect.
02Commenter 2
Dear Commenter 1,

The branch CR24521 is reviewed. No remarks.

Please test.
03Commenter 3
Dear Commenter 1,

Branch CR24521 (and products from GIT master) was compiled on Linux and Windows platforms and tested.
[revision removed]

Number of compiler warnings:

occt component :
Linux: 55 (55 on master)
Windows: 1 (1 on master)

products component :
Linux: 20 (20 on master)
Windows: 2 (2 on master)

Regressions/Differences:
bugs vis(004) bug23227: FAILED
http://occt-tests/CR24521-master-occt/Debian60-64/bugs/vis/bug23227.html
http://occt-tests/CR24521-master-occt/Windows-32-VC9/bugs/vis/bug23227.html

Testing cases:


Testing on Linux:
Total MEMORY difference: 358038616 / 357038992
Total CPU difference: 43800.63000000024 / 42990.21999999986

Testing on Windows:
Total MEMORY difference: 413653100 / 412391816
Total CPU difference: 29865.53125 / 32600.59375

There are not differences in images found by testdiff.
http://occt-tests/CR24521-master-occt/Windows-32-VC9/diff-Windows-32-VC9.html
http://occt-tests/CR24521-master-occt/Debian60-64/diff-Debian60-64.html
IMAGE mesh standard_shading X1: X1.png differs
IMAGE mesh standard_shading T8: T8.png differs
IMAGE mesh standard_shading X2: X2.png differs
IMAGE bugs vis bug288_5: bug288_5_Driver1_Viewer1_View1.png differs
IMAGE bugs vis bug24224: bug24224_Driver1_Viewer1_View1.png differs
04Author
The following issues have been observed by tests:
- Automatic culling should probably been disabled when Solid contains not triangulated faces (mesh standard_shading X1/X2).
- Disable culling for both capping and culling (capping might be specified with translucent properties, bugs vis bug24224).
- Disable culling for Shading aspect with translucent properties (caf presentation F3/F4).

Test case mesh/standard_shading/T8 produces broken shape (Shell within Solid has wrong orientation which can be fixed by fixshape, but positive output of checkshape might be considered as additional bug).

bugs/vis/bug288_5 differs due to z-fit issues (this test case is fixed by #0024001)

05Author
Updated patch is ready for review in CR24521_1 branch (CR24521 contains history).
06Commenter 6
Dear Commenter 1.

The branch CR24521_1 is reviewed,

please test.
07Commenter 7
Dear Commenter 1,

Branch CR24521_1 (and products from GIT master) was compiled on Linux and Windows platforms and tested.
[revision removed]

Number of compiler warnings:

occt component :
Linux: 54 (55 on master)
Windows: 1 (1 on master)

products component :
Linux: 20 (20 on master)
Windows: 2 (2 on master)

Regressions/Differences:
No regressions/differences

Testing cases:
Not needed

Testing on Linux:
Total MEMORY difference: 358085764 / 357062448
Total CPU difference: 41731.36999999996 / 42989.819999999854

Testing on Windows:
Total MEMORY difference: 413540996 / 412457984
Total CPU difference: 31390.5 / 32601.609375

There are following differences in images found by testdiff.
http://occt-tests/CR24521-1-master-occt/Debian60-64/diff-Debian60-64.html
http://occt-tests/CR24521-1-master-occt/Windows-32-VC9/diff-Windows-32-VC9.html
IMAGE mesh standard_shading X1: X1.png differs
IMAGE mesh standard_shading T8: T8.png differs
IMAGE bugs vis bug288_5: bug288_5_Driver1_Viewer1_View1.png differs

Related records