Being hardly adept at Nuke’s inner workings, I’ve little to no clue what might be doing it. I get an error message in that… little accompanying Nuke-cmd-ish-window, saying stuff like:
[ 1:02:03] Warning: ShuffleCopy4: get(channels=0x10), but request(channels=0x10)
[ 1:02:03] Warning: Read76: Interest(channels=0x0), but request(channels=0x0)
So I get the feeling Nuke’s upset over my .exr channels, but can’t figure out just why. Worked fine when I set it all up – I’ve switched the files the Read nodes originally read, having rendered out new ones – but it’s still the same channels, and still the same errors when I reimport the sequences with new Read nodes.
A more wordy error report now follows:
For one thing, Nuke won’t View some nodes in my nodetree. Sometimes, I can view the end result, no problem – more often, I have to restart Nuke to get it to at all View another frame. Often, it seems that the only nodes I can view are Read nodes, and if I follow those nodes down the tree, I can view my way down to the bottom result – but if I try to view the bottom result at first, it won’t preview it. I attach the node to a viewer, that white scanline-thingie appears, but nothing happens. Sometimes I can view a certain node successfully, but when I press play, it won’t show the next frame.
So it won’t View consistently; it’s much the same with rendering Write nodes. Sometimes, I can attach a Write node somewhere in the middle of the tree, and write out successfully; but attaching it to one of the nodes farther up in the flow – that is, one of the nodes that contributed to the earlier, successful node – that won’t render.
As a side note, curiously, Nuke seems to’ve decided that a few select channels in one of my .exr-sequences are to be black. At least in this node-tree – opening up a new Nuke and importing in the sequence there fresh, shows the channels in all their colourful glory. How this ties in to them error messages before, I can but imagine.
Aaanyway. So, I’m hoping some clever, well-read fellow will find this in the next few hours and solve all my problems; until then, I’ll simply redo my Nuke script, with frequent test-renders to notice if things go wrong.
Post a Comment