Thanks all for your points everyone. I have again re-written a lot of the points in response to this feedback. It was actually the students who asked for this do / don't list. Nuke is difficult to teach to busy undergrad students who cant afford the time to devote themselves to it completely... it allows anything. This is a mark of its greatness, but also of its high expectations of the user. I was, I confess, also motivated by a certain desperation on my part. I get utterly worn out editing paths on Read nodes in student projects when I want to be looking at what they have actually achieved in the script. In fairness, I do make the point at the top of the page that all rules are there to be broken.
Thorsten: Again, this was the result of my desperation with the students. They were placing Reformat nodes after Read nodes without considering how the reformatted image will conform to its new bounding box. I have re-written this point to: >> Dont over use the reformat node >> If you wish to re-format a still image why not just do so in Photoshop? It >> requires less processing from Nuke and looks tidier. It also, in my >> experience, encourages a more careful evaluation of how the image might >> relate to the rectangle of the new format. > > Thorsten > Just out of interest, in the PS Workflow page you are saying not to rely on > reformat to size-conform assets and rather do it in PS. I wonder why? I can > only see problems doing so actually, so i am a tad confused. Howard: This is a good point. I could elaborate... perhaps work backwards from a diagnosis of the result. > The main principle I teach is to understand the tool/ methodologies so that > the user can make an intelligent decision based on a full understanding. A > list of do's and dont's will always fall foul of exceptions and in the end > only leads to parrot fashion learning. An example 'never composite an > unpremultiplied image' or 'only composite a premultiplied image' *which ever > it was... > well > 1) you can - a keymix expects an unpremultiplied image. The matte tool is > there for that purpose too. Why not anyway? what would be the tell tale signs > of compositing an image with premult set wrong? (light edges, dark edges > etc..., some of which aren't even premult issues anyway) What if you are > applying any filtering to an image (translate of 0.5 pixels for example, let > alone blur etc...), what would happen if this wasn't premultiplied? I have elaborated, with examples, on the linked page: http://opticalenquiry.com/nuke/index.php?title=Premultiplication#What_happens_if_premultiplication_is_not_handled_well > What happens if you colour correct a premultiplied image? What would you look > out for? "Always check your edges for premult artifacts" would be the only > guide I would suggest here. The artifacts will always depend on a number of > mistakes, subtle or blatant and instilling the need to check this the > important thing. Fixing the cause(s) requires an understanding of why. The students projects are guaranteed to travel away from their 'pathed' drive. However, I could place the point it in a fuller context. > 'Never use explicit paths in read nodes', simply depends on your setup and > pipeline. A well set up pipeline/ network and this shouldn't be an issue > (hasn't been for me in 15 years) _______________________________________________ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
