#15620 closed defect (nonchimerax)

Wrl outputs have bleeding color

Reported by: kristen.browne@… Owned by: Tom Goddard
Priority: normal Milestone:
Component: Input/Output Version:
Keywords: Cc: Eric Pettersen, bhinnata.piya@…, phil.cruz@…
Blocked By: Blocking:
Notify when closed: Platform: all
Project: ChimeraX

Description

Tom:

I was just looking at a ribbon-print model that has small molecules attached, and it has some color bleed:
[cid:image001.png@01DAD76E.208480E0]
[cid:image002.png@01DAD76E.208480E0]
These look just fine from the previous GLBs.  Small molecules on their own are fine with crisp boundaries.  This is the only place I've seen a problem so far.

Any idea whether this might be fixable?

Thanks!

Kristen

Added by email2trac

Added by email2trac

Attachments (2)

image001.png (528.5 KB ) - added by kristen.browne@… 15 months ago.
Added by email2trac
image002.png (274.6 KB ) - added by kristen.browne@… 15 months ago.
Added by email2trac

Download all attachments as: .zip

Change History (6)

by kristen.browne@…, 15 months ago

Attachment: image001.png added

Added by email2trac

by kristen.browne@…, 15 months ago

Attachment: image002.png added

Added by email2trac

comment:1 by Eric Pettersen, 15 months ago

Cc: Eric Pettersen bhinnata.piya@… phil.cruz@… added
Component: UnassignedInput/Output
Owner: set to Tom Goddard
Platform: all
Project: ChimeraX
Status: newassigned

comment:2 by Tom Goddard, 15 months ago

Hi Kristen,

Was some mesh decimation or editing done on the per-vertex .wrl after being output from ChimeraX? Because it looks as if two coincident vertices at the midpoint of the bond between the red and yellow atoms got combined into a single vertex. The original two vertices would have one red and the other yellow to give a crisp color boundary. But if they were merged and one of the two colors was then lost then that would explain this artifact. Any mesh editing would have to take account of the per-vertex colors and not merge vertices of different colors.

If no mesh editing was done and the images you attached are directly rendered from the ChimeraX .wrl then I would want to know what VRML renderer made the image. It may be that that VRML render does the vertex combining that messes up the colors.

I am pretty sure ChimeraX is outputing a sharp color boundary and if you showed the .wrl it outputs in an faithful VRML viewer it will show the sharp boundary.

Tom

comment:3 by kristen.browne@…, 15 months ago

Tom:

I just went into ChimeraX and output a WRL directly from there and the color boundary is ok. I'm guessing that the workflows have a bug and are over-writing the wrl from ChimeraX with a processed one in our testing environment. I'll check with Mike. Our mesh processing stuff will all be replaced in not too long and will watch out for this happening at the time.

Thanks for the clarification!

Kristen

Last edited 15 months ago by Tom Goddard (previous) (diff)

comment:4 by Tom Goddard, 15 months ago

Resolution: nonchimerax
Status: assignedclosed

Glad you have a lead on the problem. Let me know if other issues arise with the VRML or GLTF.

Note: See TracTickets for help on using tickets.