Just thought I'd give you an update about this... Mr MB had a look at our scenes and encountered the same problem with reading the cache - the feather instances appeared on frame 1 but disappeared on subsequent frames.
As there were some nested models under the main chicken model it was thought that there was confusion with ICE recognising certain elements of the setup... I ran a test by removing the nested models and there was no change - still a problem reading the cache. His suggestion was it was a soft cache reading error as it was a little sporadic with the errors we were getting. After much faffing and eliminating every conceivable thing from the list of things that could cause an issue we started over. Removed EVERYTHING other than the meshes to be surfaced with feathers. The feathers were remade and then I reapplied the rig with saved out envelope weights. So far so good. Maybe the issue was with the initial build of the feather system? No answer but problem solved. Although the re-grooming of the feathers obviously adds time, at least we have a setup that works including writing the cache out from referenced animation models and applying it to referenced render models. Thanks for your thoughts. Jonny On 4 November 2013 19:53, Sandy Sutherland <sandy.mailli...@gmail.com>wrote: > You might be right Jonny, been a while since I used it. I am sure you can > look see what attributes are being written and find the instance one - it > is most likely an integer that can then be fed into an instance node. > > S. > > > On 2013/11/04 9:01 PM, Jonny Grew wrote: > >> >> >> The problem is, Sandy, that I cannot recreate the instancing section of >> the ice tree. The compound for MB feather tools is not explorable. You >> simply point it at an instance group but the allocation of each individual >> feather instance is defined by an envelope weight on a duplicate of the >> render mesh (called the envelope mesh) Under the hood this will give the >> particles a custom attribute (I'm guessing) that is used to inform which >> instance each particle should use-the same goes for size and rate but >> driven by a weight map that is local to the render mesh. Determining which >> attribute is being used and how is what I would need to know in order to >> use on a secondary ice tree by using the set instance geometry node in my >> 'cache read' particle cloud. >> >> MB looking at it but cache issue could be down to how the feather tools >> were set up originally or some kind of internal confusion due to the fact >> we have 3 separate feather tool setups all referencing the same instance >> feather group. >> >> Will report back when I know more. >> >> Jonny >> >> >> >
