Sunflower A.09.02.L

Sunflower A.09.02.L

Description

Jumping a root
Frank is jumping over a tree root.

This is the left eye view of the stereo video. The right eye is rendered in Session 1352.

Big Buck Bunny is copyright 2008, Blender Foundation / www.bigbuckbunny.org available under CC-by v3

Message boards : Comments and discussion : 1351

AuthorMessage
Loup

Send message
Joined: 18 Feb 12
Posts: 18
Credit: 3,051,672
RAC: 0
Message 11206 - Posted: 20 Feb 2012, 20:05:49 UTC

This sessions seems to cause quite a few invalid and/or conflicting results. Some examples:

http://burp.renderfarming.net/workunit.php?wuid=1677083
http://burp.renderfarming.net/workunit.php?wuid=1676994
http://burp.renderfarming.net/workunit.php?wuid=1677020
http://burp.renderfarming.net/workunit.php?wuid=1676961
http://burp.renderfarming.net/workunit.php?wuid=1676932
http://burp.renderfarming.net/workunit.php?wuid=1676899
http://burp.renderfarming.net/workunit.php?wuid=1676898
http://burp.renderfarming.net/workunit.php?wuid=1676888


All finished frames which I checked also had some white dots on the left site of the frame, around the left mushroom and around the dark plants near the stones. Just in case that these aren't supposed to be there.
ID: 11206 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Janus
Volunteer moderator
Project administrator
Avatar

Send message
Joined: 16 Jun 04
Posts: 4571
Credit: 2,100,463
RAC: 8
Message 11207 - Posted: 21 Feb 2012, 10:16:01 UTC - in response to Message 11206.  
Last modified: 21 Feb 2012, 10:28:41 UTC

Yes, there is indeed an issue with this session. It seems that there is some noise on the fur of the rodents (not Frank - just the tall and small one) as well as the nut that the tall rodent is carrying. Here is a crop of a normalized difference map between some of the results returned for frame 23 (from this workunit):


The validator correctly picks this up and increases the session redundancy to make sure it gets consistent results.
Once the session is over all the results will be revalidated using a different validation setting to calculate proper average images. Also, credit will be granted to the machines that did not get included in the first set of results (those currently having a status as "Completed, marked as invalid").

You also noticed an issue regarding white dots on the preview image. This is unrelated and is only present in the web (png and jpeg) previews used on the website. The actual rendered images have a higher dynamic range and are stored in a format that browsers cannot read (OpenEXR) - these images do not have those dots. Exactly why they appear in the previews is a good question.

So in conclusion things are working almost as intended: The input has some issues but these are being mitigated by increasing redundancy and averaging the results.
ID: 11207 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Janus
Volunteer moderator
Project administrator
Avatar

Send message
Joined: 16 Jun 04
Posts: 4571
Credit: 2,100,463
RAC: 8
Message 11222 - Posted: 4 Mar 2012, 10:24:33 UTC - in response to Message 11207.  

Credit for this session has been granted, also to the results previously marked as invalid.
ID: 11222 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Post to thread

Message boards : Comments and discussion : 1351