> I really want to avoid seeing computation from node.execute() in my toolset, > it's not looking clean for artist. I don't quite understand. What's "not clean" about it? Nuke is for about working with nodes and their functionality, no?! I have used node.execute() on many a pipeline tool for years and wouldn't want to cut it out of my options.
On Apr 6, 2011, at 10:00 PM, Bertrand Lempereur wrote: > Thanks Frank, > > I didn't know the Dilate/Erode method, it work fine for me, slow but fine ;o) > I really want to avoid seeing computation from node.execute() in my toolset, > it's not looking clean for artist. > > cheers > > 2011/4/6 Frank Rueter <[email protected]> > Here is a script that I just put together for such purposes. > http://pastebin.com/Hp7WUtT5 > > This examples uses a Grade node to auto-normalise the selected node but > obviously you can do with the output whatever you want (tuple of min and max > colour) > > I used to mess with Dilate/Erode nodes in the past to get an images min and > max colour but found it's a lot easier to utilise Nuke's MinColor node for > this, > > > Cheers, > frank > > > On Apr 5, 2011, at 9:03 PM, Bertrand Lempereur wrote: > >> Hi everyone, >> >> I'm looking for a fast way to find the minimum and maximum color value of a >> buffer. The goal is to normalize a picture of a sequence. >> I've tried the CurveTool but it need to be compute manually and it's not >> very stable. For instance, if i create it like this: >> >> C.knob('go').execute() >> With nothing connected to it, Nuke crash. >> >> The HistEQ divide the picture by it's maximum value. But for very little >> color range, the histogram must be very large and it still not very >> reliable. Also, it update on every frame and it's not what i want. >> >> I also tried the sample command in python/TCL that did off-course exactly >> what i wanted but it's very slow to compute on HD picture. >> >> What i'm looking for is a very simple command like the "Stat()" node in >> Shake that find a maximum or minimum value and return dynamically a filled >> buffer with that value. >> >> Regards, >> Bertrand >> >> _______________________________________________ >> Nuke-python mailing list >> [email protected] >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python > > > _______________________________________________ > Nuke-python mailing list > [email protected] > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python > > > _______________________________________________ > Nuke-python mailing list > [email protected] > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
_______________________________________________ Nuke-python mailing list [email protected] http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
