Ahaa! Wow that was a nice and clear explanation, thanks, I guess the conclusion is that maya 2014 has this oldschool alembic (HDF5) going on. No problem, cause I'll be leaving this workflow behind soon. Just would of loved to integrate blender into that pipeline and be more self reliant in shading/rendering. Thanks anyways.
Hi Adam, thanks for joining the discussion and providing a test file. This file is written using the obsolete HDF5 format, which Blender doesn't support. You can convert it to the current Ogawa format using abcconvert (a tool included in Alembic).
Fri, Apr 21
About the alembic file, precisely I was telling you before that I was sure that the file wasn't corrupted and I know this cause Maya imports the alembic file pefectly, so is not the alembic file. I just checked that to be secure that the file is not the problem.
After talk with Andy, things are a bit more subtle than described...
Then attach the crashing file to this ticket, otherwise we can't investigate the issue.
Thu, Apr 20
Oh Im sorry, my fault... But yes, I used the last daily build before report this problem. Just now I have downloaded the daily build again and also I have exported another alembic file from the same Maya project and the problem still persisting in Blender. Maya imports the alembic without problem, so it isn't a corrupted file though...
Well, that's easy then :-)
Given i wasn't able to repro with master in december, i'm pretty sure @Kévin Dietrich (kevindietrich) fixed this one before you took over.
I've traced the problem to missing support for "Simple" child hairs (apparently only "interpolated" is supported at the moment). D2638 should fix this.
The file works fine in current Master, please test with a daily build from builder.blender.org/download/
@Fran (foxist) at least give us the version of Blender you used to test. Those fields are there for a reason.
@Sybren A. Stüvel (sybren), mind having a look? Thanks!
This issue (lagging of imported curves) is a known problem, which is caused by limitations of the current modifier system in combination with Alembic. The modifiers are made to either construct a new object or deform an existing object. Since Alembic curves can change topology from one frame to another, the Mesh Cache modifier needs to do both (i.e. create a new object when the curve topology changes, and deform an existing object when it doesn't). To fix this properly, we'll have to modify the modifier system to allow both. I don't know if I can do this (or work around this issue another way) in time for 2.79; if not, we'll work it out in 2.8 (in which case there will be a different, more elegant approach to handling cached data).
Wed, Apr 19
Two more features just landed in Master:
Here's a simple Alembic cache with some static user properties, some animated one, and a user property on geometry. I made one of these for SideFX when they were upgrading their user property support and it proved useful. Let me know if you have any questions about it. In general, animated values are keyed on frames 1 and 10. Interpolation of these animated values if up to the package.
Tue, Apr 18
Fri, Apr 14
should we use the daily build ?
Thu, Apr 13
For the case teste, the vertex number is constant, so is it possible to calculate proper motion blur.
Anyway, as an additional step, please check if possible to import "v" vector attribute on points (v is velocity vector attribute coming from Houdini alembic) and use as motion vector to make Cycles 2.5D motion blur.
I've improved the importer quite a bit, maybe now is time to re-test using latest master?
Marking as resolved. Feel free to leave a comment when there are still issues (but please open up another ticket if it's a different issue).
Fixed in ebb304511402