This commit adds the ability to normalize patterns by their
average value while tracking, to make them invariant to global
illumination changes.
To see this in action, check out the "Lobby" scene from Hollywood
VFX. If you track the markers that are shadowed by the actress,
previously they would not track. With the scale adaption on, the
tracker would shrink the area to compensate for the changed
illumination, losing the track. With "Normalize" turned on, the
patch is correctly tracked and scale is maintained.
A remaining problem is that only the Ceres cost function is
updated to handle the normalization. The brute translation search
does not take this into account. Perhaps "Prepass" (see below)
should get disabled if normalization is enabled until I fix the
prepass to normalize as well.
There are a few other changes:
- Cleanups in tracking RNA comments.
- Bail out of the sampling loop early if the mask is zero; this
saves expensive samples of the image derivatives.
- Rename the wordy "Translation initialization" to "Prepass" at
Sebastian's suggestion.
- Fix a bug where the mask was ignored when sampling in the cost
functor.
This replaces the old style tracker configuration panel with the
new planar tracking panel. From a users perspective, this means:
- The old "tracking algorithm" picker is gone. There is only 1
algorithm now. We may revisit this later, but I would much
prefer to have only 1 algorithm. So far no optimization work
has been done so the speed is not there yet.
- There is now a dropdown to select the motion model. Choices:
* Translation
* Translation, rotation
* Translation, scale
* Translation, rotation, scale
* Affine (Not implemented yet)
* Perspective
The most stable is the "translation" parameterization. The
others work but still require some tweaking.
- The old "Hybrid" mode is gone; instead there is a toggle to
enable or disable translation-only tracker initialization. This
is the equivalent of the hyrbid mode before, but rewritten to work
with the new planar tracking modes.
- The pyramid levels setting is gone. At a future date, the planar
tracker will decide to use pyramids or not automatically. The
pyramid setting was ultimately a mistake; with the brute force
initialization it is unnecessary.
Also made BKE_tracking_get_search_imbuf use space conversion utility functions,
so now it's not so annoying that search area calculation is happening differently
in different paces.
Also allow even sizes for search area.
Another small fix is about flipping search area dumping by libmv-capi.
It used to be flipped since in blender Y axis is up-aimed.
This commit removes the use of the legacy RegionTracker API from
Blender, and replaces it with the new TrackRegion API. This also
adds several features to the planar tracker in libmv:
- Do a brute-force initialization of tracking similar to "Hybrid"
mode in the stable release, but using all floats. This is slower
but more accurate. It is still necessary to evaluate if the
performance loss is worth it. In particular, this change is
necessary to support high bit depth imagery.
- Add support for masks over the search window. This is a step
towards supporting user-defined tracker masks. The tracker masks
will make it easy for users to make a mask for e.g. a ball.
- Add Pearson product moment correlation coefficient checking (aka
"Correlation" in the UI. This causes tracking failure if the
tracked patch is not linearly related to the template.
- Add support for warping a few points in addition to the supplied
points. This is useful because the tracking code deliberately
does not expose the underlying warp representation. Instead,
warps are specified in an aparametric way via the correspondences.
- Remove the "num_samples_xy" concept and replace it with
automatic determination of the number of samples. This makes the
API easier for users.
- Fix various bugs in the parameterizations.
There remains a bug with subpixel precision tracking when in
"keyframe" mode; this will get fixed shortly.
- adding new points de-selects previous so Ctrl+Click+Drag works usefully.
- cyclic curves now extrude properly
- adding new points inbetween existing now use the surrounding points handle length.
Added four new functions as shortcuts to creating GHashes that use the
standard ptr/str/int/pair hash and compare functions.
GHash *BLI_ghash_ptr_new(const char *info);
GHash *BLI_ghash_str_new(const char *info);
GHash *BLI_ghash_int_new(const char *info);
GHash *BLI_ghash_pair_new(const char *info);
Replaced almost all occurrences of BLI_ghash_new() with one of the
above functions.
- Triangulate Cloth Mesh for collisions
- Speed up collisions
- Remove EL Topo code
- Prepare code to incooperate El Topo self collisions (TODO next commits)
TODO:
----------
- Triangulation: Is custom data/uv preserved correctly?
- Use MPoly not tessface?
- reuses auto handle calculation
- keeps handle length relate to surrounding handles
todo
- only works well when cyclic is off
- isnt aspect aware (looks like other tools are not aspect aware too)
Added support of tracks to have their own grease pencil datablock.
This is controllable by Clip/Track toggle in Grease Pencil panel
in Clip Editor.
This toggle shows whether grease pencil stroke should go to movie clip's
datablock or to active track. As soon as this toggle is switched to Track,
all further strokes would be added to active track and only strokes which
belongs to active track would be displayed on the screen.
This strokes would be relative to track's position, which means if one
made some strokes to track and started moving it, stokes would be moved
together with track.
Different tracks might share the same grease pencil datablock.
Also hide draw modes which are not supported by space clip.
Initial idea why tracks should have grease pencil datablocks was to
support easy mask creation for tracking. This seemed fastest and
easiest way to achieve mask editing for tracks.
To get mask for specified track there's function BKE_tracking_track_mask_get
which requires MovieTracking structure, track for which mask is creating and
also width and height of current overall frame. it'll return ImBuf with
mask (probably it's more like overkill, because internally it works with
single-channel value buffer and probably that's one should be passing to
libmv).
Probably hacking grease pencil in such way is not ideal solution, but that
part of code might be cleaned up further and it seems to be straightforward
to use grease pencil for such a purpose.
Note: this is commit to tomato branch, not to trunk.
This starts converting markers in the movie clip editor / 2D
tracker from using pat_min and pat_max notation to using the
a more general, 4-corner representation. There is still
considerable porting work to do; in particular:
- The preview widget does not respect the new
representation
- The corners cannot yet be edited individually.
- The clamping does not do the right thing for general
corners; e.g. detecting a convex quad.
- The tracking code in Blender does not actually invoke the
corner tracking version of libmv's tracking API.
Next steps:
- Convert libmv's legacy ESM tracker and brute tracker to
work under the new TrackRegion() API.
- Make Blender use the new TrackRegion() API; this will
allow real planar tracking tests.
- Everything else (UI, etc).
This patch is mostly the work of Sergey Sharybin.
... instead add scene.sequencer_editor_create / clear, these match id.animation_data_* functions.
- refactor for names, for scene level functions call them BKE_sequencer_*
--- Changes to File Output node ---
* Flat layer names in EXR multilayer files.
For a socket with name "AAA" the previous resulting EXR layer name would be "AAA.AAA", i.e. the render layer as well as render pass would use the socket name.
Now the "render_layer.render_pass" scheme is ignored in multilayer files, socket names are directly written to EXR layers (EXR layer name is "AAA" in this example). If sockets should have a notion of "render layer" this can still be achieved by explicitly adding a separator, e.g. "AAA.BBB". When loading such layers into a Blender Image struct, the name is interpreted as a "render_layer.render_pass" again (although the image node does not care about it, see below).
* Socket sub-paths (for singlelayer) or layer names (for multilayer) are stored in dedicated string variables in the socket storage data. This way the RNA can define precise string subtypes (PROP_FILEPATH) and length. The file/layer slots are defined as separate structs with own name properties in the RNA as well, so they can be used nicely with the list template.
* Ensure unique socket paths/layer names to prevent overwriting of files and layers respectively.
--- Changes to Image node ---
* Loading multilayer OpenEXR files has improved layer name splitting into render layer + render pass names now. This properly supports arbitrary EXR layer names now.
Example:
OpenEXR layer name: AAA.BBB.CCC
is split into
Render layer name: AAA.BBB
Render pass name: CCC
If the layer name has no '.' separators the render layer name is empty.
* Image node ignores the selected render layer in the image user data. Instead all existing layers are displayed at the same time by combining the render layer names with render pass names again, to reconstruct the original EXR layer name. This avoids the problem that render layers with empty name are not selectetable in the dropdown and allows using all image layers at the same time without duplicating the node.
