Changes between Initial Version and Version 1 of Ticket #15899, comment 19


Ignore:
Timestamp:
Sep 11, 2024, 10:42:22 PM (14 months ago)
Author:
Tom Goddard

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #15899, comment 19

    initial v1  
    1 {{{
     1
    22That all sounds right!
    33
    … …  
    2323
    2424
    25 
    26 
    27 
    28 
    29 
    30 -----Original Message-----
    31 From: ChimeraX <ChimeraX-bugs-admin@cgl.ucsf.edu>
    32 Sent: Wednesday, September 11, 2024 2:35 PM
    33 To: Browne, Kristen (NIH/NIAID) [C] <kristen.browne@nih.gov>; goddard@cgl.ucsf.edu
    34 Cc: gregc@cgl.ucsf.edu
    35 Subject: [EXTERNAL] Re: [ChimeraX] #15899: WRL vs. GLB output
    36 
    37 
    38 
    39 #15899: WRL vs. GLB output
    40 
    41 ---------------------------------------+-------------------------
    42 
    43           Reporter:  kristen.browne@…<mailto:kristen.browne@…>  |      Owner:  Tom Goddard
    44 
    45               Type:  defect            |     Status:  assigned
    46 
    47           Priority:  normal            |  Milestone:
    48 
    49          Component:  Input/Output      |    Version:
    50 
    51         Resolution:                    |   Keywords:
    52 
    53         Blocked By:                    |   Blocking:
    54 
    55 Notify when closed:                    |   Platform:  all
    56 
    57            Project:  ChimeraX          |
    58 
    59 ---------------------------------------+-------------------------
    60 
    61 Comment (by Tom Goddard):
    62 
    63 
    64 
    65 Hi Kristen,
    66 
    67 
    68 
    69 Thanks for your immense patience explaining this to me!  I think I get it  now.  Let me try explaining the way I understand it and if I have it right  we can come up with a solution.
    70 
    71 
    72 
    73 You have a scene in a GLTF file that is a tree of nodes.  Each node can  have a transformation matrix, for instance, in the current ChimeraX GLTF  the top centering node has a translation so that the bounding box of the  entire scene is centered at 0,0,0 in the world coordinates.  Let's imagine  my scene has two atomic models that are a ways apart from each other.  I  might want to rotate one of those two atomic models in the scene about its  center.  So in Blender I could click on its node and rotate it.  But that  works poorly now because Blender takes the pivot for rotation at 0,0,0 in  the local coordinates of that node.  That atomic model came from  crystallography and its bounding box is far away from 0,0,0.  So it ends  up rotating around a point far from the center of the atomic model which  is bad.  It surprises me quite a bit that Blender and other 3D scene  software would work that way.  I would expect it to rotate not about 0,0,0  in the node local coordinates.  I would expect it by default to rotate  about the center of the bounding box of the node.  Since it doesn't work  that way the tree of nodes ideally should be created so that for every  node its bounding box is centered at 0,0,0 in the local coordinates of  that node.  The node can then have a transformation on it that places the  node in the correct position within its parent node.  This is a  restrictive constraint on how the transformations of all the nodes are set  to maintain this property that every node has its geometry centered at
    74 
    75 0,0,0 in the node's local coordinate system.
    76 
    77 
    78 
    79 ChimeraX and molecular data in general is not restricted in that way.  But  I see that any node hierarchy can have its transformations replaced and  vertex buffers shifted to have that property.  Basically it amounts to  going to the bottom of the tree where there are vertex buffers, shifting  them to center them, and then undoing the shift using the transform in the  node that references those buffers.  And you just walk from the leaves of  the tree up to the root modifying the transformations all the way to the  top node doing the same trick, shift all the child nodes of a parent so  the parent is centered at 0,0,0 and then reposition the parent using its  transform so it maintains its original position in the scene.  The end  result is every node is centered at 0,0,0 in its local coordinates and the  scene looks identical, you just juggled the positioning transformations  between child nodes and parent nodes.
    80 
    81 
    82 
    83 I've never had to work with a restrictive scene graph rule like this.  But  from what you say it is the norm that 3D graphics software like Blender  and C4D rely on this assumption and if it is violated the user interface  in those programs becomes very challenging to use.
    84 
    85 
    86 
    87 If I understand the situation correctly I can see if I can do the  transformation propagation from vertex buffers up the tree to make a more  usable GLTF file.  One casualty of that is that you lose the original  crystallographic coordinates for say an x-ray atomic model node.  The node  no longer has those coordinates and you only know where it is in the  overall scene.  But for use of these GLTF files where knowing the original  source data coordinates is not likely to be important, that is no loss.
    88 
    89 
    90 
    91     Tom
    92 
    93 --
    94 
    95 Ticket URL: <https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.rbvi.ucsf.edu%2Ftrac%2FChimeraX%2Fticket%2F15899%23comment%3A18&data=05%7C02%7Ckristen.browne%40nih.gov%7C5978dce96456443722c008dcd2906537%7C14b77578977342d58507251ca2dc2b06%7C0%7C0%7C638616764816772467%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=bboKexpOOFNgtW6qiplzolBWDbqzqoBEZqV%2FR4RWvvE%3D&reserved=0<https://www.rbvi.ucsf.edu/trac/ChimeraX/ticket/15899#comment:18>>
    96 
    97 ChimeraX <https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.rbvi.ucsf.edu%2Fchimerax%2F&data=05%7C02%7Ckristen.browne%40nih.gov%7C5978dce96456443722c008dcd2906537%7C14b77578977342d58507251ca2dc2b06%7C0%7C0%7C638616764816784326%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=7bpyAMR%2FHccYNApPlf9ldia8h5bEO%2BSMJCk2xABxV5s%3D&reserved=0<https://www.rbvi.ucsf.edu/chimerax/>>
    98 
    99 ChimeraX Issue Tracker
    100 
    101 CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and are confident the content is safe.
    102 }}}
    103 
    10425[[Image(image001.jpg)]]