August 21th, 2013 at 5:00pm PDT, Agenda and Minutes

From Web3D.org
Jump to: navigation, search

Attendees: Myeong Won Lee, Joe Williams, Don Brutzman, Dick Puk, William Glascoe, and Kwan-Hee Yoo

Agenda and Minutes:

1. Regular meeting times

   On August 7, Joe mentioned:
   Let's meet every Wednesday.
   Once a month is not enough for what we need to do.
   
   Based on Joe's request, the regular H-Anim meeting will be held every Wednesday at 5pm PDT. 
   I may or may not be able to attend each week, due to the inflexibility of my work schedule.  If I have no other meetings (e.g. Korea Chapter, CAD, AR, e-Learning) to attend in a given week, I will attend the H-Anim WG meeting.
   However, I encourage any H-Anim WG member to give prior notice and hold an H-Anim WG meeting whenever they feel it is necessary to speed up the progression of agenda items.  

2. Th e 10 H-Anim game characters

   - updated and uploaded to the H-Anim Wiki 
   
   http://web3d.org/wiki/index.php/H-Anim#Resources
   http://www.web3d.org/membership/login/documents/navega.php?PHPSESSID=fe89f0184573c6c3409ecbd4a9b3697f&dir=.%2FH-Anim%20WG
   http://www.web3d.org/membership/login/documents/accion.php?PHPSESSID=fe89f0184573c6c3409ecbd4a9b3697f&dir=.%2FH-Anim%20WG%2FAll10Characters   
   
      All characters have been scaled with (scale 0.0225 0.0225 0.0225).  
      All characters display properly in both BS Contact and Octaga.  
      However, I still don't fully understand why we have to manually insert this scaling into the x3d files.       
   - personalitiies requested by Don ==> under consideration.  These will be fabricated as the characters were generated as generic avatars for a VR edutainment game with no distinct personality traits at the time of generation.           
    

3. Resolution problem in wrl and x3d

   - resolution comparison between max, wrl, and x3d 
   - Ideally, general graphics tools should generate wrl files with high resolution 
   - or, wrl and x3d browsers should display them in high resolution 
   - or, the wrl-to-x3d converter should insert the CreaseAngle value? 
   
   Should the wrl-to-x3d H-Anim conversion include a value for CreaseAngle?
   How should this value be determined? It up to the viewer to determine. The resolution of the figure is not affected by any setting of the crease angle.

http://www.web3d.org/x3d/content/examples/Conformance/Geometry/IndexedFaceSet/_pages/page21.html

   <IndexedFaceSet colorPerVertex='false' creaseAngle='6.27' colorIndex='0 0 0 0 0 0 0 0 0 0 0 0 0 ...

http://www.web3d.org/x3d/content/examples/Conformance/Geometry/IndexedFaceSet/_pages/page20.html

   <IndexedFaceSet colorPerVertex='false' creaseAngle='1.5' colorIndex='0 0 0 0 0 0 0 0 0 0 0 ... 

4. H-Anim behaviors

  - Questions about pitch, yaw, and roll among the behaviors: 

MWL wonders if these aviation terms appropriate for humanoids? Joe thinks they are appropriate. Dick thinks it is not a good idea to use the terms because they are context dependent. Joe rationalizes his diction by explaining how the shoulder moves. Don's non-value judgement is 1) humanoids are unique so the behaviors are unique and an author's choice. 2) it's not they are wrong; it's just confusing to others 3) Dick recounts the discovery of the use of various pitch, roll and yaw interpretations during technical exchanges with mathematicians for another ISO specification.

Don recommends with come up with different names because pitch, roll and yaw will not work for the specification.

5. H-Anim presentations at the SC24 Sydney Meeting, August 26-30, 2013

   1) H-Anim specification revision (NWIPs and WD Outlines) and current status of H-Anim developments, Myeong Won Lee
   2) H-Anim facial animation, Jung-Ju Choi
   3) H-Anim hands and feet animation, Kwan-Hee Yoo  (Kwan-Hee Yoo will talk about this topic) 

WISH LIST ITEM: We want sign language animations for every alphabet and/or character set in the international character set (http://en.wikipedia.org/wiki/International_character_set).

WISH LIST ITEM: Kwan-Hee Yoo to lead the identification of the unicode for all medical terms germane to H-anim's skeleton(s).

6. Schedule next meeting

  We will skip next week’s meeting due to the SC24 Sydney meeting so the next H-Anim WG meeting will be (1st Wed) September 4th at 5:00pm PDT (5th, 9:00am KST).

Added by Don

1. the plan listed here sounds great.

Joe and I may send avatars to meetings on off days. 8)

2. thanks for the wonderful model updates. astounding progress: Chul passed all validaton tests!! Wow!! Very impressive. 8)

from discussion, here are some translator comments:

- please use .x3d (not .X3D) as filename suffix

- please use X3D version="3.3"

- do not use rotation='0 0 0 0' since that is not a legal value

- please do not use DEF="hanim_etc." rather use character name DEF="Chul" etc. this will let authors copy them into one file.

I will work on adding these to the model archive and provide further translator comments. Again thank you.

3. not sure yet about the resolution problem... will think about it. wondering if some 3DS Max values or settings are not getting translated into X3D properly, perhaps a misunderstanding exists.

creaseAngle discussion..

- it is an authoring choice. yes, sometimes tools offer this as a feature. it is hard to come up with a single rule that is always correct.

- here is tooltip:

http://www.web3d.org/x3d/content/X3dTooltips.html#IndexedFaceSet

creaseAngle [creaseAngle accessType initializeOnly, type SFFloat CDATA "0"] [0,∞) creaseAngle defines angle (in radians) for determining whether adjacent polygons are drawn with sharp edges or smooth shading. If angle between normals of two adjacent polygons is less than creaseAngle, smooth shading is rendered across the shared line segment. Hint: creaseAngle=0 means render all edges sharply, creaseAngle=3.14159 means render all edges smoothly.

http://www.web3d.org/x3d/content/X3dTooltipsKorean.html#IndexedFaceSet

creaseAngle [creaseAngle accessType initializeOnly, type SFFloat CDATA "0"] [0,infinity) creaseAngle은 인접한 폴리곤이 날카로운 변두리 혹은 유연한 음영이 그려질때 각(라디안)을 정의 한다. 만약 두 인접한 폴리곤의 법선이 creaseAngle 보다 작다면, 유연한 음영은 공통의 라인 구분을 지나면서 표현한다 .

Interchange profile Hint: only 0 and _o radians supported. Hint: creaseAngle=0 은 모든 변두리를 예리하게 표현하는것을 의미하고, creaseAngle=3.14는 모든 변두리를 유연하게 표현하는것을 의미한다 .

creaseAngle has nothing to do with the geometry, it changes how the geometry is drawn.

example use in an H-Anim model: Jin might have creaseAngle='0' for showing every crease and fold in the Shape nodes for the shorts, then use creaseAngle='3.14159' for showing smooth legs.

Further discussion available in our _X3D for Web Authors_ book, slides and videos.

http://x3dgraphics.com/slidesets/X3dForWebAuthors/Chapter06-GeometryPointsLinesPolygons.pdf

4. human behaviors: we had a spirited discussion about names, confusion, roll/pitch/yaw, etc.

Joes extension of some basic behaviors in the original NPS interface for Nancy are good. Even if some are confusable, they are valid X3D and legal authoring choices.

Maybe we can improve them to avoid confusion... Joe I'm happy to help.


William mentioned medical animations - this is a promising direction to explore.

Congratulations everyone, we have begun working on the future challenges!

5. Preparations for ISO Sydney meeting are well along and nearly complete.

Kwan Hee Yoo has a wonderful slideset. He presented highlights to us. After the meeting, Myeong Won Lee will ensure that all presentation slidesets are on wiki.

- lots of interesting discussion

- small note: your presentation .pdf slide 11 appears to be missing an image. perhaps slide 16 is missing an image too??

- of interest: slide 12 shows a foot with all bones labeled similarly to H-Anim specification, with annotations both in English and Korean... Kwan Hee Yoo is searching for Unicode translations in many different languages. that will let us automate skeleton annotations for any language, both in tools and in H-Anim models themselves if desired.

We communicated that full-fidelity models are very important, we are striving for anatomical correctness. Any subset models for animation must be strict subsets and not different. Then the H-Anim Level Of Articulation (LOA) models can be added cumulatively.

It was not clear from this slideset what joints are missing from H-Anim 1.1 so please list the changes if possible.

Thanks from everyone for the great work.

6. We agreed to follow schedule listed below.

Corrections to minutes are welcome.