| View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
|---|---|---|---|---|---|---|---|---|---|
| 0005256 | Spring engine | General | public | 2016-05-30 20:04 | 2016-12-23 18:04 | ||||
| Reporter | Anarchid | ||||||||
| Assigned To | Kloot | ||||||||
| Priority | normal | Severity | minor | Reproducibility | always | ||||
| Status | closed | Resolution | no change required | ||||||
| Product Version | 101.0+git | ||||||||
| Target Version | Fixed in Version | ||||||||
| Summary | 0005256: rendering discrepancies between model formats | ||||||||
| Description | Collada (maybe all assimp?), obj (using spring's own pipeline) and s3o models seem to render differently when partially built, and when cloaked using the CUS cloak shader. The cloaking difference seems to affect only Collada of the formats i've tried. Here is a screenshot of armjamt using the same geometry in obj and dae formats alongside, in cloaked, part-cloaked, and part-built states: http://storage2.static.itmages.com/i/16/0530/h_1464630821_7890479_115252caf8.png The build difference seems to affect everything that is not s3o or 3do. Here are two (different) models alongside in partially built state. One is s3o, and another is dae. The dae model jumps to have its entire top half full-textured at about 65%, and, then fills that back again with raw color, and then snaps to being completely built. http://storage1.static.itmages.com/i/16/0530/h_1464631233_4386146_98a543bc6a.png This behaviour seems consistent among GPUs and OSes. | ||||||||
| Tags | No tags attached. | ||||||||
| Checked infolog.txt for Errors | |||||||||
| Attached Files |
| ||||||||
 Relationships |
|
 Relationships |
| Notes |
|
|
Anarchid (reporter) 2016-05-31 17:13 |
I now realize that i cannot edit the bug report, but "reproducibility" should be "always". |
|
Anarchid (reporter) 2016-07-04 11:13 |
The last time i tested, at least one dae model seemed to render correctly during nanoframe phase. I will try to test others and see if this part of the problem went away. |
|
Google_Frog (reporter) 2016-07-04 14:03 |
Nanoframe rendering depends on correctly detecting the upper and lower extremes of the model. Spring is bad at detecting these extremes for dae and that is what results in the broken nanoframe visuals. I fought with Bumblebee for a while to make its visuals somewhat reasonable. The extreme detection seems to depend on a lot of variables so whether an un-fought model works is essentially random. I moved the model around in blender as well as messing with parameters like height, radius and the position of base in its LUS. |
|
Kloot (developer) 2016-12-23 18:03 Last edited: 2016-12-23 18:04 |
"the CUS cloak shader" no such shader exists. there is however a lups "cloaker" particle class which, like several others used by ZK, has either no or broken detection for non-builtin model types ("self.isS3o = ...") because the code predates assimp's introduction and was never properly updated since. why is it that I know these details but active Zk contributors don't? |
| Notes |
| Issue History |
|||
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2016-05-30 20:04 | Anarchid | New Issue | |
| 2016-05-31 17:13 | Anarchid | Note Added: 0016382 | |
| 2016-05-31 20:29 | hokomoko | Reproducibility | have not tried => always |
| 2016-07-04 11:13 | Anarchid | Note Added: 0016467 | |
| 2016-07-04 14:03 | Google_Frog | Note Added: 0016468 | |
| 2016-12-23 18:03 | Kloot | Assigned To | => Kloot |
| 2016-12-23 18:03 | Kloot | Status | new => closed |
| 2016-12-23 18:03 | Kloot | Resolution | open => no change required |
| 2016-12-23 18:03 | Kloot | Note Added: 0016993 | |
| 2016-12-23 18:04 | Kloot | Note Edited: 0016993 | View Revisions |
| Issue History |