ICE | Arnold Render Errors With Strands

844 views
Skip to first unread message

Mitchell Lotierzo

unread,
Jun 29, 2012, 4:53:09 PM6/29/12
to soft...@listproc.autodesk.com
Has anyone out there ever had problems getting strands to render in Arnold? I get this error message when I draw a render region:

"""
# WARNING : [arnold] ignoring parameter points on Strands_PntCld.strands (found: nans or infs)
# WARNING : [arnold] [curves] Invalid number of points (expected 40800, found 0)
"""

I can see my strands perfectly fine in the viewport though. I've noticed that when I disable my "Slide on Surface" node in the ICETree, I can get the strands to render perfectly fine. I tried tweaking every setting in the node as well, but nothing seems to work besides disabling it. There must be something on the inner workings of the node that I can't wrap my head around.

Any thoughts?

Much appreciated,
Mitch

Steven Caron

unread,
Jun 29, 2012, 4:54:53 PM6/29/12
to soft...@listproc.autodesk.com
post this to the sitoa list, include which version you are using.

Todd Akita

unread,
Jun 29, 2012, 4:57:32 PM6/29/12
to soft...@listproc.autodesk.com
It may be that your strand arrays (position, velocity, color, etc) are not all the same size. That would be the first thing I would check.

-T

Mitchell Lotierzo

unread,
Jun 29, 2012, 7:02:36 PM6/29/12
to soft...@listproc.autodesk.com
Thanks guys. I appreciate your responses.

@Todd: I logged the size of all my arrays and they are all identical. The only one that wasn't was Orientation, but I'm pretty sure that should be returning a size of zero since I have "Loft Shape Along Strand" checked in the "Generate Strand Trails" node. I'll keep debugging the tree, but thanks for the tip. :)

@Steven: I wasn't aware there was a sitoa list. I'm pretty new to the world of Arnold. I'm sure my sup. is a member, so I'll have him post a blurb for me. Thanks!

Steven Caron

unread,
Jun 29, 2012, 7:23:18 PM6/29/12
to soft...@listproc.autodesk.com
ya, definitely send it there to make sure to include the version you are using. stefano has done a lot of work on ICE implementation and it constantly is improving.

s

Steven Caron

unread,
Jun 29, 2012, 7:26:53 PM6/29/12
to soft...@listproc.autodesk.com
correction... send it there too (as in also)...

Sam Cuttriss

unread,
Jun 29, 2012, 9:13:52 PM6/29/12
to soft...@listproc.autodesk.com
I was getting this error whenever i attempted to resize ltk_lightning strands.

fine in viewport/ mental ray, not so much in arnold

Sandy Sutherland

unread,
Jun 30, 2012, 1:47:27 AM6/30/12
to soft...@listproc.autodesk.com
I can reply on the correct list.

S.

_____________________________
Sandy Sutherland
Technical Supervisor
sandy.su...@triggerfish.co.za
_____________________________





From: softimag...@listproc.autodesk.com [softimag...@listproc.autodesk.com] on behalf of Sam Cuttriss [tea...@gmail.com]
Sent: 30 June 2012 03:13
To: soft...@listproc.autodesk.com
Subject: Re: ICE | Arnold Render Errors With Strands

Mitchell Lotierzo

unread,
Jul 3, 2012, 11:14:31 AM7/3/12
to soft...@listproc.autodesk.com
So, just an update. After getting in touch with Stefano, it was determined that the "Surface Force" node in my tree was spitting out invalid point IDs. My emitter geo was sitting pretty much right on top of the geo I was plugging into that node, so I just gave the emitter geo a slight push as a simple workaround for now.

Here's Stefano's response as well:

"Arnold fails because there is at least one strand point position with Nan coordinates, caused by the Add Force node in your compound. We don't do any particular check on the coordinates before sending them to Arnold. It's Softimage that returns the invalid values to the exporter. The first invalid strand is #10. If you emit 10 strands, you should see the render working. If you raise to 11, it fails, because of strand #10. If you display the points id in the viewport, you'll see no #10 (with 15, also #13 is gone, as showing below).

So, I guess that the display skips #10 because invalid, yet it's sent to the exporter by the sdk. If you manage to get rid of these invalid cases, it should work again. It's probably the Surface Force node that pushes out some invalid vector in the first place.

Let me know. Ciao. Stefano"
debug_strands.png
Reply all
Reply to author
Forward
0 new messages