gpapaioa Posted May 26, 2014 Share Posted May 26, 2014 After searching the forum, I manged to create a motion vector pass and according to some posts, in order to use it correctly in Nuke, the values must be fitted/normalized. I used MV2Nuke (http://www.nukepedia.com/gizmos/colour/mv2nuke) but still cant seem to find a proper workflow. Has anyone came up with some solid workflow on how to proper use Mantra's motion vector pass? Thank you. Quote Link to comment Share on other sites More sharing options...
Neil78 Posted May 27, 2014 Share Posted May 27, 2014 I use this technique - transferring to screen space - check it out. You could normalize the motion vectors in Houdini - but also MV2Nuke works well for this, as it means that when using the vector blur in Nuke - the 'multiply` value makes sense - such as 0.5 for 180 degree shutter etc. I'd be really interested in any other techniques you might come up with. But also I think that the motion blur in Houdini is so fast, more often than not motion vectors are not always needed in Nuke. motionvectors.hipnc Quote Link to comment Share on other sites More sharing options...
Guest tar Posted May 27, 2014 Share Posted May 27, 2014 I use this technique - transferring to screen space - check it out. Isn't shops calculated in screen space automatically? Quote Link to comment Share on other sites More sharing options...
gpapaioa Posted May 28, 2014 Author Share Posted May 28, 2014 Thanx Zweger, I will check your .hip. The thing is that I have to deal with scaled cameras (dont ask why...) and I want to be able to tweak the values in Nuke so I can manually fix them in a way. When you use MV2Nuke, which one do you think is the appropriate preset for Mantra? Quote Link to comment Share on other sites More sharing options...
Neil78 Posted May 28, 2014 Share Posted May 28, 2014 Hi Marty - good point - i didn't know that, i'm assuming that 'P' is not - so overcompensating perhaps - but i've never actually tested the difference. gpapaioa - i've found 'raster space' to be best - getting normalized values into vector blur. Quote Link to comment Share on other sites More sharing options...
eetu Posted May 28, 2014 Share Posted May 28, 2014 Isn't shops calculated in screen space automatically? To be precise, shops are calculated in camera space, not screen space - and Neil is transforming from camera ("current") to screen ("ndc") space, which looks correct to me. Quote Link to comment Share on other sites More sharing options...
Neil78 Posted May 28, 2014 Share Posted May 28, 2014 Good to know - thanks eetu; i remember finding recently that P was not world P but camera P. - and thus great for Z depth pass in Nuke. Quote Link to comment Share on other sites More sharing options...
gpapaioa Posted May 28, 2014 Author Share Posted May 28, 2014 Thanks everyone for the valuable information Quote Link to comment Share on other sites More sharing options...
XRM Posted June 7, 2014 Share Posted June 7, 2014 For scaled cam, I always use an expression in respective scale xyz fields origin(opinputpath(".",0), "", "SX") origin(opinputpath(".",0), "", "SY") origin(opinputpath(".",0), "", "SZ") Check this link http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&p=142497&sid=55ff675427276bd80c80b160b13f1d5f Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.