This is not necessary: the implicit solver data can keep track instead
of how many off-diagonal matrix blocks are in use (provided the
allocation limit is calculated correctly). Every time a spring is
created it then simply increments this counter and uses the block index
locally - no need to store this persistently.
This is more involved than using simple straight bending targets
constructed from the neighboring segments, but necessary for restoring
groomed rest shapes.
The targets are defined by parallel-transporting a coordinate frame
along the hair, which smoothly rotates to avoid sudden twisting (Frenet
frame problem). The rest positions of hair vertices defines the target
vectors relative to the frame. In the deformed motion state the frame
is then recalculated and the targets constructed in world/root space.
derivatives for stabilization.
The bending forces are based on a simplified torsion model where each
neighboring point of a vertex creates a force toward a local goal. This
can be extended later by defining the goals in a local curve frame, so
that natural hair shapes other than perfectly straight hair are
supported.
Calculating the jacobians for the bending forces analytically proved
quite difficult and doesn't work yet, so the fallback method for now
is a straightforward finite difference method. This works very well and
is not too costly. Even the original paper ("Artistic Simulation of
Curly Hair") suggests this approach.
This returns a general status (success/no-convergence/other) along with
basic statistics (min/max/average) for the error value and the number
of iterations. It allows some general estimation of the simulation
quality and detection of critical settings that could become a problem.
Better visualization and extended feedback can follow later.
These are much better suited for creating stiff hair. The previous
bending springs are based on "push" type spring along the hypothenuse
of 3 hair vertices. This sort of spring requires a very large force
in the direction of the spring for any angular effect, and is still
unstable in the equilibrium.
The new bending spring model is based on "target" vectors defined in a
local hair frame, which generates a force perpendicular to the hair
segment. For further details see
"Artistic Simulation of Curly Hair" (Pixar technical memo #12-03a)
or
"A Mass Spring Model for Hair Simulation" (Selle, Lentine, Fedkiw 2008)
Currently the implementation uses a single root frame that is not yet
propagated along the hair, so the resulting rest shape is not very
natural. Also damping and derivatives are still missing.
code.
The implicit solver itself should remain agnostic to the specifics of
the Blender data (cloth vs. hair). This way we could avoid the bloated
data conversion chain from particles/hair to derived mesh to cloth
modifier to implicit solver data and back. Every step in this chain adds
overhead as well as rounding errors and a possibility for bugs, not to
speak of making the code horribly complicated.
The new subfolder is named "physics" since it should be the start of a
somewhat "unified" physics systems combining all the various solvers in
the same place and managing things like synchronized time steps.
This will allow us to implement moving reference frames for hair and
make "fictitious" forces optional, aiding in creating stable and
controllable hair systems.
Adding data in this place is a nasty hack, but it's too difficult to
encode as a DM data layer and the whole cloth modifier/DM intermediate
data copying for hair should be removed anyway.
responses.
The S matrix together with the z Vector encodes the degrees of freedom
of a colliding hair point and the target velocity change. In a collision
the hair vertex is restricted in the normal direction (when moving
toward the collider) and the collision dynamics define target velocity.
Instead of handling contact tests and collision response in the same
function in collision.c, first generate contact points and return them
as a list, then free at the end of the stepping function. This way the
contact response can be integrated into the conjugate gradient method
properly instead of using the hackish and unstable double evaluation
that is currently used.
This is still using the old BVH tree collision methods to generate
contact points, similar to what cloth does. This should be replaced
by a Bullet collision check, but generating contacts in this way is
easier for now, and lets us test responses and stability (although in
more complex collision cases the BVH method fails utterly, beside being
terribly inefficient with many colliders).
Description:
--------------------------
Use loose edges marked as seams as sewing springs.
Usage:
-------------------------
All this patch does is set the rest length to 0 and the stiffness to 1 for springs for loose edges marked as seams so that during the cloth simulation they will be brought together.
Example Video:
-------------------------
http://www.youtube.com/watch?v=-Y_bC0gjoM0
Original Patch by thesleepless (+ git patch by codemanx)
Thank you!
Fix 1: Pinned vertices were never released when "unpinned" by Dynamic Paint.
Fix 2: When pinning vertices during simulation, they would get "warped" to their original starting position of frame 1.
Thanks to MiikaH for pointing this out and also for providing the regression blend file: http://wiki.blender.org/uploads/a/ab/Cloth_dynamic_paint.blend
Self collision vertex groups enable artists to exclude selected vertices from getting involved in self collisions. This speeds simulations and it also resolves some self collision issues.
- 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?
added some missing functions too - which are not used yep but should be there for api completeness.
* CDDM_set_mloop
* CDDM_set_mpoly
* BLI_mempool_count
- removed deprecated bitmap arg from IMB_allocImBuf (plugins will need updating).
- mostly tagged UNUSED() since some of these functions look like they may need to have the arguments used later.
to specify different spring lengths.
Implementation is quite ugly because the shape key has to be pulled
through the modifier stack somehow, need a more flexible data mask
system to solve this properly.
(commits 27773,27775,27778 by Brecht from render25 branch)
* Convert all code to use new functions.
* Branch maintainers may want to skip this commit, and run this
conversion script instead, if they use a lot of math functions
in new code:
http://www.pasteall.org/9052/python
- Hair dynamics have their own panel in particle settings with the settings from cloth panel that apply to hair.
- Basic internal friction force to quickly emulate self collisions and volume preservation. (Still very early code, but gives some idea of what's possible).
- Softbody simulation is no longer used for hair.
* Old files with sb dynamics should just load the hair without dynamics so new dynamics can be applied.
* Invasion of particles exceptions in sb code is finally over.
- Collisions with other objects are disabled for now and will be worked out in the future.
Other changes/fixes:
- Particle mode editing flag wasn't saved properly.
- Some old files with edited hair didn't load correctly.
- Disabled delete & specials menu in particle mode for non-hair editing.
- Fixed yet one more cloth & softbody pointcache update issue.
- Disconnect/connect hair now uses only the deformed mesh so it works correctly also for subsurfed emitters.
- Hair editing now updates correctly with a moving emitter.
- HEADER (beginning of each file)
* general header:
+ 8 char: "BPHYSICS"
+ 1 int: simulation type (same as PTCacheID->type)
* custom header (same for sb, particles and cloth, but can be different for new dynamics)
+ 1 int: totpoint (number of points)
+ 1 int: data_types (bit flags for what the stored data is)
- DATA (directly after header)
*totpoint times the data as specified in data_types flags
- simulation type
soft body = 0, particles = 1, cloth = 2
- data types (more can be added easily when needed)
data flag contains
----------------------------------------
index (1<<0) 1 int (index of current point)
location (1<<1) 3 float
velocity (1<<2) 3 float
rotation (1<<3) 4 float (quaternion)
avelocity (1<<4) 3 float (used for particles)
xconst (1<<4) 3 float (used for cloth)
size (1<<5) 1 float
times (1<<6) 3 float (birth, die & lifetime of particle)
boids (1<<7) 1 BoidData
Notes:
- Every frame is not nescessary since data is interpolated for the inbetween frames.
- For now every point is needed for every cached frame, the "index" data type is reserved for future usage.
- For loading external particle caches only "location" data is necessary, other needed values are determined from the given data.
- Non-dynamic data should be written into an info file if external usage is desired.
* Info file is named as normal cache files, but with frame number 0;
* "Non-dynamic" means data such as particle times.
* Written automatically when baking to disk so basically a library of particle simulations should be possible.
- Old disk cache format is supported for reading, so pre 2.5 files shouldn't break. However old style memory cache (added during 2.5 development) is not supported. To keep memory cached simulations convert the cache to disk cache before svn update and save the blend.
- External sb and cloth caches should be perfectly possible, but due to lack of testing these are not yet enabled in ui.
Other changes:
- Multiple point caches per dynamics system.
* In the future these will hopefully be nla editable etc, but for now things are simple and the current (selected) point cache is used.
* Changing the amount of cached points (for example particle count) is allowed, but might not give correct results if multiple caches are present.
- Generalization of point cache baking etc operator & rna code.
- Comb brushing particle hair didn't work smoothly.
* Based on what happens during simulation the cache is marked (also in cache panel, this could possibly be extended to 3d view as well) as:
- exact (not marked)
- outdated (simulation is not done completely with current settings)
- non-exact (frames were skipped during simulation)
* The parameter "cache step" effects the number of frames between saved cache frames.
- This can save a lot of memory (or disk space) if absolutely frame accurate simulation is not required.
- Speeds up the "quick caching" very much.
- Frames between cached frames are interpolated from the cached frames.
- Current default value of 10 frames works nicely with up/down-arrows (skip 10 frames forwards/backwards on timeline), but can be changed if wanted.
* The caching can work in normal or "quick" mode:
[Normal cache]
- Basic: Calculate what even happens (settings change, big frame steps etc.) and cache results, if possible try to use "cache step" when saving cache frames.
- Becomes non-exact: After larger than 1 frame steps.
- Becomes outdated: After any change effecting the simulation other than frame steps.
- Pros/cons: Freedom of doing anything and playing with particles, but exact results have to calculated from the beginning.
[Quick cache]
- Basic: Calculate simulation up to current frame automatically on changes with cache step sized jumps in simulation. With multiple "quick cached" simulations the smallest cache step is used.
- Becomes non-exact: Always from frame 1 (unless cache step = 1).
- Becomes outdated: Never.
- Pros/cons: Not very accurate, but super fast!
- Todo: Transform of any animated (non-autokeyed) object is locked! Probably needs some tinkering with anim sys overrides.
* The simulation can be run forwards or backwards even if it's cache is outdated or non-exact, the following rules apply in these situations:
- step forwards (to unknown) -> simulate from last exact frame, store result
- step backwards (to known) -> result is interpolated from existing frames, store result, clear cache forwards if current frame is after last exact frame
* "Calculate to current frame" runs the simulation from start to current frame with a frame steps of 1.
- Baking does the same, but runs the simulation all the way to the end of simulation.
- Rendering does this automatically if the simulation is outdated of non-exact, so all rendered simulations will always be updated and exact.
* Every cache panel also holds buttons to "Bake all dynamics", "Free all dynamics" and "Update all dynamics to current frame".
* Cloth simulation supports the new cache too.
