| Anonymous | Login | Signup for a new account | 2013-05-18 17:00 PDT |  |
| Main | My View | View Issues | Change Log | Roadmap |
| View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
| 0000790 | LuxRender | Core | public | 2010-06-04 03:27 | 2010-11-10 04:36 | ||||
| Reporter | heretic | ||||||||
| Assigned To | jeanphi | ||||||||
| Priority | normal | Severity | major | Reproducibility | always | ||||
| Status | closed | Resolution | fixed | ||||||
| Platform | OS | OS Version | |||||||
| Product Version | 0.7RC2 | ||||||||
| Target Version | 0.8RC1 | Fixed in Version | 0.7.1 | ||||||
| Summary | 0000790: mattetranslucent behaving "problematic" and wrong | ||||||||
| Description | Statement: mattetranslucent should by default apply as material for eg normal uncoated paper. Mattetranslucent as it is now does not suit this req. more details and an example: http://www.luxrender.net/forum/viewtopic.php?f=14&t=4014&start=10#p40202 [^] remarks 1. mattetranslucent as it is now tends to burn out pretty fast in exposed areas (despite r+t <= 1) 2. materials with high transmission requires low reflection. Esspecially with textures this results in a dirty, greyish reflection of the matte part in shadowy areas. As a result: gettign mattetranslucent right and balancing r and t is close to impossible with increasing t for anythign that is NOT in front of a light source. This becomes worse with a texture channel applied to both. Suggestion: - drop the r+t rule, but hardcode it dependant on amount of light passing through a mattetranslcuent object | ||||||||
| Additional Information | newbie explanation: Take paper for example. 100g normal white uncoated paper is by default mattetranslucent. Laying flat on a surface, that paper is 100% reflection and 0% transmission. As a result one could choose eg 95% white matte instead. Take the very same paper to the edge of the table and have a light shine through from below. One part of the paper is at app. 80% transmission now (where the light shines through), the other part at 0% transmission (the part flat on the table). The area between both parts would be blended towards the 0% transmission (light leaking from the transmission part over). What do I want to tell by that example? - I dont know how mattetranslucent is defined in a technical sense - as a user for a material I would expect that transmission and reflection is based on amount of light passing through. Solid surface behind the mattetranslucent material=100% reflection channel and 0% transmission channel, at increasing amounts of light passign through, the transmission increases and reflection decreases. mattetranslucent - how I would expect it: - defined as it is now, except both channeles are made as 100% (without the r+t <=1 rule) - add another channel: transmission amount eg in percent, ranging 0 - 1 Example for the "real newbie-mattetranslucent" for one pixel: Situation: x% of light passing through, eg indirect light passing through our paper from somewhere (however this can be measured - I am the newbie) Property of our pixel material in that situation: transmission=(transmission amount)* (x%) * (transmission color) reflection=(1-(transmission amount)* (x%))*(reflection color) | ||||||||
| Tags | No tags attached. | ||||||||
| Mercurial Changeset # | |||||||||
| Requires Documentation Update | No | ||||||||
| Requires Exporter Update | |||||||||
| Attached Files | |||||||||
 Relationships |
|
 Relationships |
|
Notes |
|
|
(0002239) heretic (reporter) 2010-06-04 03:44 |
addendum: - setting r=0 (true black) on mattetranslucent and mixing it with another material eases the problem slightly but doesnt eliminate the base problem. |
|
(0002246) jeanphi (administrator) 2010-06-06 01:51 |
Actually there are 2 use cases with mattetranslucent: volume or sheet surfaces. For sheet surfaces I think it makes sense to keep current behaviour and just scale down colors if the sum is above one. For volume surfaces I think it'd make more sense to just define the reflection color and let all the remaining light go through and be absorbed (maybe keep a transmission value that would multiply this remainder). |
|
(0002249) heretic (reporter) 2010-06-06 02:50 |
Sheet surfaces - agreed Volume surface - there are sufficient mattetranslucent materials with similar reflection, but heavily differing translucency. Can those be handled by that volume surface? |
|
(0002252) jeanphi (administrator) 2010-06-06 13:57 |
Translucency is a result of absorption, so I'd say yes. You can attach volumetric properties to an object with a mattetranslucent material. |
|
(0002351) jeanphi (administrator) 2010-09-04 08:31 |
A new option "bool energyconserving" allows to switch to a new behaviour where the transmission is multiplied by (1-reflection). |
|
Notes |
|
Issue History |
|||
| Date Modified | Username | Field | Change |
| 2010-06-04 03:27 | heretic | New Issue | |
| 2010-06-04 03:44 | heretic | Note Added: 0002239 | |
| 2010-06-06 01:51 | jeanphi | Note Added: 0002246 | |
| 2010-06-06 02:50 | heretic | Note Added: 0002249 | |
| 2010-06-06 13:57 | jeanphi | Note Added: 0002252 | |
| 2010-06-06 14:07 | jeanphi | Target Version | => 0.8RC1 |
| 2010-08-08 11:44 | jeanphi | Assigned To | => jeanphi |
| 2010-08-08 11:44 | jeanphi | Status | new => assigned |
| 2010-09-04 08:31 | jeanphi | Requires Documentation Update | => No |
| 2010-09-04 08:31 | jeanphi | Note Added: 0002351 | |
| 2010-09-04 08:31 | jeanphi | Status | assigned => resolved |
| 2010-09-04 08:31 | jeanphi | Resolution | open => fixed |
| 2010-11-10 04:34 | jeanphi | Fixed in Version | => 0.7.1 |
| 2010-11-10 04:36 | jeanphi | Status | resolved => closed |
|
Issue History |
| Copyright © 2000 - 2012 MantisBT Group |
 |