PFTrack, already established as a great 3D tracker, moved to provide a new level of functionality when it provided added object tracking in version 3.0. Now with version 3.5 this is refined even further. With Version 3.5 is due out soon, fxguide sat down with Richard Spöhrer Managing Director of PixelFarm to discuss the product in this week’s podcast.
Having already established a great 3D tracker, Pixel farm have moved to provide a new level of functionality when it provided object tracking in version 3.0 of PFTrack. Now with version 3.5 this has been refined even more. With Version 3.5 just out, fxguide sat down with Richard Spöhrer Managing Director of PixelFarm to discuss the product in this week’s podcast.
PFTrack, is a comprehensive image analysis tool, with both camera reconstruction and object tracking. PFTrack has gained popularity with matchmovers for its tracking, imbedded image based modeling tools and automatic z depth extraction. The 3.0 camera solver provided a fast solution and it has enabled Match movers to handle even more complex shots. One of the main improvements in PFtrack 3.5 is segmented tracking. This allows the user to take sections of the same shot into linked but separate solutions. This can be great for real world complex movements that can often prove difficult to obtain as one universally solved move.
While many people have assumed that a 3D tracker is ideal to track objects, in fact most 3D trackers do not solve object tracking, it is not a natural by-product of camera solutions and often would not be possible at all with conventional 3D trackers. PFTrack was one of the first 3D trackers to explicitly add the ability can now track many separate moving objects within a sequence, as well as solving camera motion issues – you simply export data from tracking groups, whether it’s camera data, object movement or motion-capture sessions.
Points can now be converted to a mesh – allowing the match mover to create terrain models quickly. With the addition of the unique geometry tracking, pftrack allows 3D geometry to be imported into PFTrack and then matched to sequences, making this an ideal tool for head and facial replacement meaning artists can now be confident that animated CG, when rendered and composited, will exactly match the live action. Complex Geometry can be imported and used to solve cameras in addition to the use of automatic or user placed features. This is perfect for problematic shots or for tracking set extension models into footage.
New fully editable F-Curves within PFTrack make cleaning up tracking data extremely easy and accurate. You can quickly view and edit each individual property of the camera within a familiar graph editor in the timeline â€“ curves can be represented by both linear & splines and then simply smoothed or edited using the tools provided. Once edited, the new camera solver quickly produces a new camera without the need to totally re solve every parameter in the sequence, which can save a lot of time as a shot iteratively is improved.
The new, fully integrated image based modeling capabilities allow users to quickly build up 3D environments by using the background image as a guide. Previously this has only been available as a stand alone system, so the improved image based modeling tools in V3.0 provide an integrated environment which delivers a much more flexible solution. Perfect for pre-visualization and animation guidance, textured models can be exported to all major 3D applications.
Integrated Z depth extraction tools assign depth data, on a per pixel basis, to sequences for exporting as a grey map. This data can be viewed within PFTrack, in 3D, using the new orthographic view options which can then be broken down into any number of layers. The program is now fully 64bit running on PC/LINUX/IRIX/Mac-Intel or PowerPC.
We've been a free service since 1999 and now rely on the generous contributions of readers like you. If you'd like to help support our work, please join the hundreds of others and become an fxinsider member.