David, Not sure if this will give you exactly what you're after, but this is a technique I've used in the past to replicate motionblur from the motion of a shape.
One way to get motion vectors that represent the motion of your shape is to do a splinewarp that warps between your shape and the same shape offset by 1 frame. Then apply this SplineWarp to a coordinates map (not normalized, but actual pixel coordinates), substract the original map from the SplineWarp result, and use that as your motion vector information inside VectorBlur. Attached is an example script. Hope it helps. On Thu, Aug 15, 2013 at 11:26 AM, Julik Tarkhanov <ju...@hecticelectric.nl>wrote: > I don't think you can do that with roto, since the only place where it can > interpolate motion is at the edges (but not inside the shapes). So in > theory you could get a kind of a vector map per pixel but that map would be > limited to where the roto edge is located, since a roto shape is > post-filled. You could try to extract (guess) the vectors from roto motion > but this is prone to artifacts. If you want to add moblur inside of an area > I would use Kronos with different plates used as source and warp. > > On 15 aug. 2013, at 19:39, "David Yu" <dave...@gmail.com> wrote: > > > I turn on motion blur to see the effect which i want to use as the > "vectors" that will drive the vector blur > _______________________________________________ > 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 >
roto_to_vectorblur_test.nk
Description: Binary data
_______________________________________________ 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
