Alembic stitching issue except when there is an overlapping frame

[GoogleCodeExporter]

This should be all of the information, but if anything is missing please let me 
know.  Thanks.

Alembic Version: 1.1.2
Houdini-Version: 12.0.754
Rendered with SESI-ROP (build with Alembic 1.1.2)
Stitcher used: 1.0.5.beta24

Stitching will fail for:
* 1011-1014 & 1015-1017

Stitching will work with an overlapping frame:
* 1011-1014 & 1014-1017

The error the stitcher displays:
------------------------------------------------------------------
Can not stitch different sampling type for node "/test"
        timePerCycle values are different
------------------------------------------------------------------

The files (alembic & hip) are attached.

All we did was:
-Created a Sphere
-Create transfrom and animate in x and connected them
-Create ROP-subnet
-Create an Alembic ROP (Sesi's). inside the subnet
-Only changed the filename on the ROP.
(btw. there seams to be another bug. Changing the Root-Object from "/obj" to 
"/obj/test" will result in empty alembic-files. Dont ask me why. So I rendered 
it with "/obj")

Original issue reported on code.google.com by shappysf...@gmail.com on 18 Jan 2013 at 3:17

Attachments:

• alembic_stitch_issue.tar.gz

[GoogleCodeExporter]

Sorry, ignore the line:
    Stitcher used: 1.0.5.beta24

We used the 1.0.5 and 1.1.2 abcstitcher executables.

Original comment by shappysf...@gmail.com on 18 Jan 2013 at 3:18

[GoogleCodeExporter]

Can you reattach the file (or send it to me privately)?  It's 0 bytes.

Original comment by miller.lucas on 22 Jan 2013 at 5:42

[GoogleCodeExporter]

Ok, let's try this again...

Original comment by shappysf...@gmail.com on 25 Jan 2013 at 4:02

Attachments:

• alembic_stitch_issue.tar.bz2

[GoogleCodeExporter]

Looks like that worked now

Original comment by shappysf...@gmail.com on 25 Jan 2013 at 4:02

[GoogleCodeExporter]

The time per cycle is indeed subtly different between the two.

I'll look into giving the check a little more floating point leniency but I'm 
worried that the ROP may be introducing floating point error in it's 
calculation of fps.

Original comment by miller.lucas on 25 Jan 2013 at 5:39

• Changed state: InProgress

[GoogleCodeExporter]

A fix was checked in here.
http://code.google.com/r/millerlucas-dev/source/detail?r=3059ed4a7a4758d9fbe9413
20319d4faaab772c3

Original comment by miller.lucas on 25 Jan 2013 at 7:37

• Changed state: PleaseVerify

[GoogleCodeExporter]

Thanks Lucas.  We'll try that next week!

Original comment by shappysf...@gmail.com on 26 Jan 2013 at 4:33

[GoogleCodeExporter]

Original comment by miller.lucas on 14 Feb 2013 at 9:08

• Changed state: Verified