User tracks are the method I've adopted from the SynthEyes workflow; the concept is; less trackers but more accurate trackers, the better, aiming for ~8 - 15 on screen at any one time.
I'm quite open to learning why hundreds of trackers improves a camera track though. All that has achieved, for me at least, is to add noise and errors to the solve. Matching the 2d tracker to 3d tracker is to reduce the RMS - the closer the 3d tracker projects to the 2d one means a more accurate track. On 16 May 2015 at 12:01, Deke Kincaid <[email protected]> wrote: > I don't understand why you would try to fix a track this way by trying to > match the 2d to the 3d one? Why not work on improving the track instead by > adding usertracks? > ᐧ > > -- > Deke Kincaid > Media & Entertainment OEM Development Manager > The Foundry > Skype: dekekincaid > Tel: (310) 399 4555 - Mobile: (310) 883 4313 > Web: www.thefoundry.co.uk > Email: [email protected] > > On Fri, May 15, 2015 at 2:29 PM, Ned Wilson <[email protected]> wrote: > >> Marten, >> >> What you have described is exactly how I've been doing it, but to me, >> this is a counter intuitive and time-consuming workflow. There are >> many times where 3D points that are part of the solve, even after >> error filtering, are not even close to accurate. With several hundred >> tracking points, it can be excessively complicated and frustrating to >> guess which 2D point corresponds to the 3D point that is in error. >> >> So, I guess this is more of a feature request. Thanks everyone for >> clarifying, and I will forward this on to the Foundry. >> >> -n >> >> On Thu, May 14, 2015 at 11:25 PM, Marten Blumen <[email protected]> wrote: >> > Just select them in 2d and delete ya. >> > >> > Steps are: Track, solve, create linked camera, tab key -select camera >> and >> > lock 3d camera -top red icon, then zoom in, identify the tracks, tab key >> > switch to 2d, select tracker, delete, update . rinse and repeat. >> > >> > On 15 May 2015 at 17:53, Daniel Hartlehnert <[email protected]> wrote: >> >> >> >> Because you can be specific about what you want to delete? Instead of >> >> letting an algorithm decide. >> >> >> >> Am 15.05.2015 um 04:46 schrieb Marten Blumen: >> >> >> >> What's the advantage of selecting/deleting 3d points? just use the >> >> RMS/Track Error slider to set the threshold on the AutoTracks tab. bad >> ones >> >> go red - delete them... >> >> >> >> On 15 May 2015 at 07:28, Daniel Hartlehnert <[email protected]> wrote: >> >>> >> >>> Unfortunately nothing. It doesn't work that way. It could be so easy, >> but >> >>> its not :( >> >>> >> >>> Am 14.05.2015 um 20:34 schrieb Ned Wilson: >> >>> >> >>> > Hello all, >> >>> > >> >>> > I have a plate connected to a CameraTracker, and am attempting to >> get >> >>> > a decent solve. With the camera tracker selected and connected to >> the >> >>> > viewer, I switch into 3D view. From there, I can clearly visualize >> >>> > some of the solved tracking points that are in error. What I want to >> >>> > do is marquee select them, delete them, and update the solve. >> >>> > >> >>> > I tried Vertex Selection mode. Well, frankly, I tried all of the >> >>> > different selection modes. However, try as I might, I am unable to >> >>> > actually select anything in the 3D view. The only way I have done >> this >> >>> > successfully in the past is to switch back into the 2D view, try and >> >>> > guess at which points are in error, and then go back into 3D to see >> if >> >>> > they are selected. >> >>> > >> >>> > What am I missing here? >> >>> > >> >>> > -n >> >>> > _______________________________________________ >> >>> > Nuke-users mailing list >> >>> > [email protected], >> http://forums.thefoundry.co.uk/ >> >>> > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >>> >> >>> _______________________________________________ >> >>> Nuke-users mailing list >> >>> [email protected], http://forums.thefoundry.co.uk/ >> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> >> >> >> >> _______________________________________________ >> >> Nuke-users mailing list >> >> [email protected], http://forums.thefoundry.co.uk/ >> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> >> >> >> >> >> >> _______________________________________________ >> >> Nuke-users mailing list >> >> [email protected], http://forums.thefoundry.co.uk/ >> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> > >> > >> > >> > _______________________________________________ >> > Nuke-users mailing list >> > [email protected], http://forums.thefoundry.co.uk/ >> > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> _______________________________________________ >> Nuke-users mailing list >> [email protected], http://forums.thefoundry.co.uk/ >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> > > _______________________________________________ > Nuke-users mailing list > [email protected], http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >
_______________________________________________ Nuke-users mailing list [email protected], http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
