In these two versions, the edges look fine.
But, in these two versions, as if there is an error, even the hidden part of the edges are visible.
Could you anyone please reply?
Thanks!
This is a good question for @Aparajit_Pratap Any ideas on why this may be happening?
Also - @Lynn_Kim would it be possible to share your graph so that we can test? If you donβt want to share in public you can always direct message us.
yeah probably these changes:
DynamoDS:master
β mjkkirschner:linezfight
opened 12:23AM - 04 Feb 22 UTC
### Purpose
A continuation of https://github.com/DynamoDS/Dynamo/pull/11400 -β¦ now we also add depth bias to Dynamo point and line geometry.
This PR also removes the code related to vertex colors as it's even harder to figure out if points and lines need custom coloring or not - I've just removed it as it did not work anyway and would have been confusing to anyone who ventured here.
This PR assigns a default depth bias based on geometry type, and offsets this during selection, so that the depth bias ranges do not overlap.
This improves z fighting in most cases, though there are still some situations it does occur. Especially with curves - I've found that in some cases the issue is worse than z fighting - the lines are actually inside the surfaces because the faceting engine produces fewer straight line segments than the tessellated mesh surfaces, I think to fix this more robustly we would need to improve the tessellation algorithms in LibG for curves, or use higher precision for curves than for meshes, both potentially hurting performance.
IMO this PR is a pretty good improvement for a small change. Open to other suggestions though for sure.
the following images show this PR on the left and 2.13.1 on the right.
β οΈ note that for a pathologically bad case there are two copies of the tsplinesurface overlapping in the images.
![Screen Shot 2022-02-03 at 5 17 26 PM](https://user-images.githubusercontent.com/508936/152451342-cf63f2c3-e4ba-4970-a542-26c38bcc79ff.png)
![Screen Shot 2022-02-03 at 5 18 27 PM](https://user-images.githubusercontent.com/508936/152451343-2b32d015-393a-4bdb-8a62-67c375b0c19c.png)
![Screen Shot 2022-02-03 at 5 18 06 PM](https://user-images.githubusercontent.com/508936/152451344-6d6effae-9f3b-4462-b8ff-265bd781a1d3.png)
![Screen Shot 2022-02-03 at 5 17 46 PM](https://user-images.githubusercontent.com/508936/152451345-53886d4c-c3fc-48d9-abbe-1e42a865ed7d.png)
Adds a few image comparison tests, I created them on a citrix machine so we'll need to see how/if they pass on the test machines.
### Declarations
Check these if you believe they are true
- [ ] The codebase is in a better state after this PR
- [ ] Is documented according to the [standards](https://github.com/DynamoDS/Dynamo/wiki/Coding-Standards)
- [ ] The level of testing this PR includes is appropriate
- [ ] User facing strings, if any, are extracted into `*.resx` files
- [ ] All tests pass using the self-service CI.
- [ ] Snapshot of UI changes, if any.
- [ ] Changes to the API follow [Semantic Versioning](https://github.com/DynamoDS/Dynamo/wiki/Dynamo-Versions) and are documented in the [API Changes](https://github.com/DynamoDS/Dynamo/wiki/API-Changes) document.
- [ ] This PR modifies some build requirements and the readme is updated
### Release Notes
improve rendering of coincident geometry
unfortunately to tackle all of these cases at arbitrary scales is quite difficult. I assume that if you scale up this geometry the edges are hidden again?
2 Likes
I scaled up the geometry but it still has the issue.
I attached my script file and related SAT file.
Please provide a fundamental solution.
Thank so much @Michael_Kirschner2 @solamour @jacob.small .
TriRandomExtrude.dyn (52.4 KB)
model_export.sat (11.9 KB)
2 Likes
Thank you @Lynn_Kim - We have filed a task to look into this
1 Like
thanks @Lynn_Kim for the repro case, this should be fixed move near clipping plane further away for more depth buffer precision. by mjkkirschner Β· Pull Request #13338 Β· DynamoDS/Dynamo Β· GitHub - the graphical results are still not EXACTLY the same as before, but they are much improved to support more cases.
the PR was just merged so a daily build of 2.17 should contain it soon for you to test.
3 Likes