Talk:Catmull–Clark subdivision surface: Difference between revisions

m
added a section header to the first topic to properly place the table-of-contents (TOC) --- (this happens more often than one would think).
(removed the discussion about the error in the example data (which is now fixed))
m (added a section header to the first topic to properly place the table-of-contents (TOC) --- (this happens more often than one would think).)
 
(4 intermediate revisions by 2 users not shown)
Line 1:
== what's the input for the function? ==
 
What's the input for the function? Example input/output?
 
Line 81 ⟶ 83:
( 4, 18, 25, 14);
]
 
: I propose this image to illustrate the article:
[[Image:Catmull_subsurf.png‎]]
:: Looks good, but I would suggest breaking it into three images and losing the built-in descriptor lines. I would also suggest stretching the contrast ratio such that it's white on black, rather than light-gray on dark-gray. --[[User:Short Circuit|Michael Mol]] 08:08, 9 January 2010 (UTC)
[[Image:CatmullClark_subdiv0.png|thumbnail|left|Original mesh]]
[[Image:CatmullClark_subdiv1.png|thumbnail|left|1 subdivision]]
[[Image:CatmullClark_subdiv2.png|thumbnail|left|2 subdivisions]]
::: I have tryed to make another set with your suggestions, but I have uploaded the original sizes because I thought IM was able to make nice thumbnails, but it fails, so I will have to make the resize through Gimp and re-upload these. --[[User:Blue Prawn|Blue Prawn]] 23:39, 22 January 2010 (UTC)
:::: Here the new set resized with Gimp:
[[Image:CatmullClark_subdiv_0.png|Original mesh (bis)]]
[[Image:CatmullClark_subdiv_1.png|1 subdivision (bis)]]
[[Image:CatmullClark_subdiv_2.png|2 subdivisions (bis)]]
 
== Algorithm Description Improvements ==
Line 104 ⟶ 94:
: Hi, I think candidate for deletion should rather be considering if an implementation is poor at the design point of view. Here the code produces wrong output but the error is probably a minor calculation error somewhere. There is also the Tcl example that produces wrong output on the border of the hole. Borders of holes should be smooth, and on the screenshot we can see that it's not. [[User:Blue Prawn|Blue Prawn]] 23:20, 17 August 2011 (UTC)
 
:: The Tcl solution is now fixed. It was a problem in the code to update point locations on the edge of the wholehole (which to be fair isn't actually discussed ''anywhere'' on the WP page or in the literature that I found with only a small amount of searching). The formula I picked seems to give nice-looking results. –[[User:Dkf|Donal Fellows]] 20:39, 7 November 2011 (UTC)
::: This current RC page explains how to handle this (which was found by examining results from blender.org). Could you check that what you have done is the same? [[User:Blue Prawn|Blue Prawn]] 01:04, 9 November 2011 (UTC)
:: In general though, only delete a solution if it is “considered harmful”, i.e., actively promoting bad practices not necessary to the solution of the task. Having a bug is not a sin, and there's no way to persuade people to fix things on any exact schedule. (You could try commenting on the original submitter's talk page; they ''might've'' configured email notifications.) Or try to learn enough OCaml to be able to fix it; the algorithm is complex enough that it remains itself the major challenge. –[[User:Dkf|Donal Fellows]] 20:39, 7 November 2011 (UTC)
 
== Should have another example geometry? ==
 
Looking at the examples, I can't really be sure that code handles meshes with holes properly.
 
Perhaps (a) we should have a task which renders these geometries, and (b) we should have a task example which has a hole in it? (Perhaps a cube with the top and bottom surfaces removed?) --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 19:59, 15 May 2015 (UTC)