in fact I redid part of the last "fix", making it working properly now.
Before we were changing Lamp->la . This is the Blender Lamp, we shouldn't touch it.
So this part of the code is correct now.
Things that could be tackled:
- color attribute is returning negative values when NEGATIVE is toggled
- objects with no material (default gray one) still don't support lamp spots (not spot lamp but the spot of the lamps)
Four new buttons in World settings to control frame rate:
fps: Nominal frame rate in frame per second.
Also sets the physics timestep = 1/fps
phys: Maximum number of physics timestep per game frame in case
the actual fps is less than nominal. This allows the
physics to keep up with real time even if the graphics slows
down the game.
sub: Fixed number of simulation substeps per physic timestep.
Improves the precision of the physics simulation. Useful for
fast moving objects for example.
log: Maximum number of logic steps per game frame in case the
actual fps is less than nominal. This allows the logic
system to follow the physics simulation.
Upper bound = phys
(setting the value higher than phys has no effect).
On games with heavy logic system, it is useful to set this
value to 1, to keep logic time under control.
All these values were already accessible from Python except phys:
GameLogic.getMaxPhysicsFrame():
Gets the maximum number of physics frame per render frame.
GameLogic.setMaxPhysicsFrame(phys):
Sets the maximum number of physics timestep that are executed per render frame.
Higher value allows physics to keep up with realtime even if graphics slows down the game.
Physics timestep is fixed and equal to 1/tickrate (see setLogicTicRate)
maxphysics/ticrate is the maximum delay of the renderer that physics can compensate.
phys: integer
Compound shape control
======================
1) GUI control
It is now possible to control which child shape is added to
a parent compound shape in the Physics buttons. The "Compound"
shape button becomes "Add to parent" on child objects and
determines whether the child shape is to be added to the top
parent compound shape when the game is stated.
Notes: * "Compound" is only available to top parent objects
(objects without parent).
* Nesting of compound shape is not possible: a child
object with "Add to parent" button set will be added
to the top parent compound shape, regardless of its
position in the parent-child hierarchy and even if its
immediate parent doesn't have the "Add to parent" button set.
2) runtime control
It is now possible to control the compound shape at runtime:
The SetParent actuator has a new "Compound" button that indicates
whether the object shape should be added to the compound shape
of the parent object, provided the parent has a compound shape
of course. If not, the object retain it's individual state
while parented.
Similarly, the KX_GameObject.setParent() python function has
a new compound parameter.
Notes: * When an object is dynamically added to a compound
shape, it looses temporarily all its physics capability
to the benefit of the parent: it cannot register collisions
and the characteristics of its shape are lost (ghost, sensor,
dynamic, etc.).
* Nested compound shape is not supported: if the object
being parented is already a compound shape, it is not
added to the compound parent (as if the Compound option
was not set in the actuator or the setParent function).
* To ensure compatibility with old blend files, the Blender
subversion is changed to 2.48.5 and the old blend files
are automatically converted to match the old behavior:
all children of a Compound object will have the "Add to
parent" button set automatically.
Child ghost control
===================
It is now possible to control if an object should becomes ghost
or solid when parented. This is only applicable if the object
is not added to the parent compound shape (see above).
A new "Ghost" button is available on the SetParent actuator to
that effect. Similarly the KX_GameObject.setParent() python function
has a new compound parameter.
Notes: * This option is not applicable to sensor objects: they stay
ghost all the time.
* Make sure the child object does not enter in collision with
the parent shape when the Ghost option if off and the parent is
dynamic: the collision creates a reaction force but the parent
cannot escape the child, so the force builds up and produces
eratic movements.
* The collision capability of an ordinary object (dynamic or static)
is limited when it is parented: it becomes automatically static
and can only detect dynamic and sensor objects.
* A sensor object retain its full collision capability when parented:
it can detect static and dynamic object.
Python control
==============
KX_GameObject.setParent(parent,compound,ghost):
Sets this object's parent.
Control the shape status with the optional compound and ghost parameters:
compound=1: the object shape should be added to the parent compound shape (default)
compound=0: the object should keep its individual shape.
In that case you can control if it should be ghost or not:
ghost=1 if the object should be made ghost while parented (default)
ghost=0 if the object should be solid while parented
Note: if the object type is sensor, it stays ghost regardless of ghost parameter
parent: KX_GameObject reference or string (object name w/o OB prefix)
Previous patch was not sorting the state actuators. This was causing
some problems with YoFrankie that relies on the order of actuators
when multiple state actuators are activated at once.
Active state actuators will now be sorted per object. This doesn't
change the fact that state actuators are executed before all other
actuators as before.
Incidently, made the logic loop faster.
The modules were moved in order to access the camera internal matrixes. It will make then compatible with multiple viewports in a near future.
So far the problem I found was:
1) KX_Camera doesn't store the canvas viewport
2) RAS_ICanvas methods: GetDisplayArea and GetWindowArea are affected by multiple viewports (and they shouldn't).
Test file is here: http://www.pasteall.org/blend/68
use PY_SET_ATTR_FAIL and PY_SET_ATTR_SUCCESS return values so the fake subclassing can know if a value failed to be set or if it was missing from the type. (with PY_SET_ATTR_MISSING)
Also noticed some other mistakes.
- KX_LightObject, setting the type didnt check for an int.
- KX_SoundActuator, didnt return an error when assigning an invalid orientation value
- KX_GameObject, worldOrientation didnt return an error value.
Added an 'Invisible' button to make this more clear, it seems like a display option but its also related to logic because the actuators can toggle this after the game starts.
Without this its annoying to add UV's only to set the invisible flag.
Sensor objects were not clearing the softbody gameflag
Servo control motion actuator did not work as expected when the object
is moving on a moving platform.
This patch introduces a new Ref field in the servo motion actuator
to set a reference object for the velocity calculation.
You can set the object during the game using the actuator "reference"
attribute; use an object name or an object reference.
The servo controller takes into account the angular velocity of the
reference object to compute the relative local velocity.
A new bookmark button is available on the controller UI.
When set, the controller is guaranteed to execute before all
other non-bookmarked controllers, provided it is scheduled
for execution.
This is useful for initialization scripts that run once at startup or
scripts that must set some prerequisite for the other controllers at
the start of each logic frame.
This feature is also available at python level with the "bookmark"
attribute. It can be changed during the game.
Note that if several script are bookmarked, their relative order of
execution is not guaranteed. Make sure they don't depend on each other.
- Size adjustments can be accomplished with warp mesh data now. So we get a free spot in the GUI for a tilt option.
- Tilt option to tilt the camera (for planetarium domes).
Angle is in degree from -180 to +180. It's needed for planetarium domes (as this one http://domejunky.blogspot.com/2009/05/dome-corrected-bge.html ).
- This is the last commit regarding dome code I expected to 2.49. I consider this feature full implemented now. (working on docs now)
- corrections to docs
- disallow calling controller.activate(actuator) when the controller is not active. (Raise a SystemError)
- Added 2 new attributes, CValue.name - deprecates CValue.getName(), KX_GameObject.children deprecated KX_GameObject.getChildren(), (same for getChildrenRecursive()).
A new type of "Sensor" physics object is available in the GE for advanced
collision management. It's called Sensor for its similarities with the
physics objects that underlie the Near and Radar sensors.
Like the Near and Radar object it is:
- static and ghost
- invisible by default
- always active to ensure correct collision detection
- capable of detecting both static and dynamic objects
- ignoring collision with their parent
- capable of broadphase filtering based on:
* Actor option: the collisioning object must have the Actor flag set to be detected
* property/material: as specified in the collision sensors attached to it
Broadphase filtering is important for performance reason: the collision points
will be computed only for the objects that pass the broahphase filter.
- automatically removed from the simulation when no collision sensor is active on it
Unlike the Near and Radar object it can:
- take any shape, including triangle mesh
- be made visible for debugging (just use the Visible actuator)
- have multiple collision sensors using it
Other than that, the sensor objects are ordinary objects. You can move them
freely or parent them. When parented to a dynamic object, they can provide
advanced collision control to this object.
The type of collision capability depends on the shape:
- box, sphere, cylinder, cone, convex hull provide volume detection.
- triangle mesh provides surface detection but you can give some volume
to the suface by increasing the margin in the Advanced Settings panel.
The margin applies on both sides of the surface.
Performance tip:
- Sensor objects perform better than Near and Radar: they do less synchronizations
because of the Scenegraph optimizations and they can have multiple collision sensors
on them (with different property filtering for example).
- Always prefer simple shape (box, sphere) to complex shape whenever possible.
- Always use broadphase filtering (avoid collision sensor with empty propery/material)
- Use collision sensor only when you need them. When no collision sensor is active
on the sensor object, it is removed from the simulation and consume no CPU.
Known limitations:
- When running Blender in debug mode, you will see one warning line of the console:
"warning btCollisionDispatcher::needsCollision: static-static collision!"
In release mode this message is not printed.
- Collision margin has no effect on sphere, cone and cylinder shape.
Other performance improvements:
- Remove unnecessary interpolation for Near and Radar objects and by extension
sensor objects.
- Use direct matrix copy instead of quaternion to synchronize orientation.
Other bug fix:
- Fix Near/Radar position error on newly activated objects. This was causing
several detection problems in YoFrankie
- Fix margin not passed correctly to gImpact shape.
- Disable force/velocity actions on static objects
KX_GameObject.getVelocity() would set an error but nor return an error value when an non vector argument was given.
KX_PythonSeq_Type was not initialized with PyType_Ready which could crash blender when inspecting the type.
[#18701] Issues with camera.pointInsideFrustum method
- note in docs that the projection matrix is not correct for first logic tick.
Renamed...
KX_Camera.isViewport -> KX_Camera.useViewport
KX_Lamp.quat_attenuation -> KX_Lamp.quad_attenuation
Deprecated KX_Camera.getProjectionMatrix(), KX_Camera.setProjectionMatrix() for projection_matrix attr
Added most missing docs reported by the doc-checker script
This commit completes the support for modifiers in the BGE.
- The physic shape is generated according to the derived mesh.
This is true for all types of shapes and all types of
objects except soft body.
- Optimization for static derived mesh (mesh with modifiers
but no armature and no shape keys). Replicas will share
the derived mesh and the display list: less memory and
faster rendering. With this optimization, the static
derived mesh will render as fast as if the modifiers were
applied.
Known Limits:
- Sharing of mesh and display list is only possible between
in-game replicas or dupligroup. If you want to instantiate
multiple objects with modifiers, use dupligroup to ensure
best memory and GPU utilization.
- rayCast() will interact with the derived mesh as follow:
Hit position and hit normal are the real values according
to the derived mesh but the KX_PolyProxy object refers to
the original mesh. You should use it only to retrieve the
material.
- Dynamic derived mesh have very poor performance:
They use direct openGL calls for rendering (no support
for display list and vertex array) and they dont't share
the derived mesh memory. Always apply modifiers on dynamic
mesh for best performance.
- Time dependent modifiers are not supported.
- Modifiers are not supported for Bullet soft body.
scene.active_camera can now be set so you can more easily set the current camera from python scripts without using an actuator.
ConvertPythonToCamera utility function to get a camera from a python string or KX_Camera type.
This commit extends the technique of dynamic linked list to the logic
system to eliminate as much as possible temporaries, map lookup or
full scan. The logic engine is now free of memory allocation, which is
an important stability factor.
The overhead of the logic system is reduced by a factor between 3 and 6
depending on the logic setup. This is the speed-up you can expect on
a logic setup using simple bricks. Heavy bricks like python controllers
and ray sensors will still take about the same time to execute so the
speed up will be less important.
The core of the logic engine has been much reworked but the functionality
is still the same except for one thing: the priority system on the
execution of controllers. The exact same remark applies to actuators but
I'll explain for controllers only:
Previously, it was possible, with the "executePriority" attribute to set
a controller to run before any other controllers in the game. Other than
that, the sequential execution of controllers, as defined in Blender was
guaranteed by default.
With the new system, the sequential execution of controllers is still
guaranteed but only within the controllers of one object. the user can
no longer set a controller to run before any other controllers in the
game. The "executePriority" attribute controls the execution of controllers
within one object. The priority is a small number starting from 0 for the
first controller and incrementing for each controller.
If this missing feature is a must, a special method can be implemented
to set a controller to run before all other controllers.
Other improvements:
- Systematic use of reference in parameter passing to avoid unnecessary data copy
- Use pre increment in iterator instead of post increment to avoid temporary allocation
- Use const char* instead of STR_String whenever possible to avoid temporary allocation
- Fix reference counting bugs (memory leak)
- Fix a crash in certain cases of state switching and object deletion
- Minor speed up in property sensor
- Removal of objects during the game is a lot faster
controller.actuators[name] and controller.sensors[name]
Made a read-only sequence type for logic brick sensors and actuators which can access single items or be used like a list or dictionary.
We could use a python dictionary or CValueList but that would be slower to create.
So you can do...
for s in controller.sensors: print s
print controller.sensors["Sensor"]
print controller.sensors[0]
sensors = list(controller.sensors)
This sequence type keeps a reference to the proxy it came from and will raise an error on access if the proxy has been removed.
Commit 20099 started using a FBO way too big.
According to Paul Bourke this is how it's done in other Engines:
Projectors HD:
1920x1050 - buffersize = 1024; FBO size = 2048
1400x1050 - buffersize = 1024; FBO size = 2048
Projectors XGA:
1024x768 - buffersize = 512; FBO size = 1024
Now in Blender Game Engine we are using:
Projectors HD:
1920x1050 - buffersize = 1050; FBO size = 2048
1400x1050 - buffersize = 1050; FBO size = 2048
Projectors XGA:
1024x768 - buffersize = 768; FBO size = 1024
(I guess I should be committing code to the ge_dome branch instead of the trunk. I feel bad doing all those adjustments in a hurry to 2.49 final release in the trunk. That is ok, right?)
*) a small note:
In the end it turned out that we have upright and downright domes out there.
So I may rearrange the order of the gui later:
(1 = fisheye, 2 = truncated up, 3 = truncated down, 4 = envmap, 5 = spherical panoramic)
I don't plan to do a doVersion() for that, so if you are using it already keep in mind that the modes may change before 249 final release.
After last commit (20099) warping meshes got slower (more quality == less performance). Since we don't need an extra warping for truncated domes, It's better to handle them directly in openGL without the need of warping it.
I'll talk with some Dome owners to see if we need both Upright and Downright modes. I may remove one of them by 2.49 them.
*) also: a proper GLEW_EXT_framebuffer_object check before generating FBO (for warping meshes).
**) next in line (maybe after RC2): tilt option to tilt the camera up to 90º upward.
We are using an image twice as big to render the fisheye before warping.
It'll slow down warping meshes a little, but we get way more resolution.
Therefore I will bring Truncated Dome mode back in order to avoid using warping mesh for that.
not fixed but the problem is now less bad when projection painting, bilinear interpolation was rounding down.
- added gameOb.attrDict to get the internal gameObject dict.
- mesh.getVertex wasnt setting an exception.
This commit extend the technique of dynamic linked list to the mesh
slots so as to eliminate dumb scan or map lookup. It provides massive
performance improvement in the culling and in the rasterizer when
the majority of objects are static.
Other improvements:
- Compute the opengl matrix only for objects that are visible.
- Simplify hash function for GEN_HasedPtr
- Scan light list instead of general object list to render shadows
- Remove redundant opengl calls to set specularity, shinyness and diffuse
between each mesh slots.
- Cache GPU material to avoid frequent call to GPU_material_from_blender
- Only set once the fixed elements of mesh slot
- Use more inline function
The following table shows the performance increase between 2.48, 1st round
and this round of improvement. The test was done with a scene containing
40000 objects, of which 1000 are in the view frustrum approximately. The
object are simple textured cube to make sure the GPU is not the bottleneck.
As some of the rasterizer processing time has moved under culling, I present
the sum of scenegraph(includes culling)+rasterizer time
Scenegraph+rasterizer(ms) 2.48 1st round 3rd round
All objects static, 323.0 86.0 7.2
all visible, 1000 in
the view frustrum
All objects static, 219.0 49.7 N/A(*)
all invisible.
All objects moving, 323.0 105.6 34.7
all visible, 1000 in
the view frustrum
Scene destruction 40min 40min 4s
(*) : this time is not representative because the frame rate was at 60fps.
In that case, the GPU holds down the GE by frame sync. By design, the
overhead of the rasterizer is 0 when the the objects are invisible.
This table shows a global speed up between 9x and 45x compared to 2.48a
for scenegraph, culling and rasterizer overhead. The speed up goes much
higher when objects are invisible.
An additional 2-4x speed up is possible in the scenegraph by upgrading
the Moto library to use Eigen2 BLAS library instead of C++ classes but
the scenegraph is already so fast that it is not a priority right now.
Next speed up in logic: many things to do there...
also deprecated getActuators() and getSensors() for 'sensors' and 'actuators' attributes.
an example of getting every sensor connected to an object.
all_sensors = [s for c in ob.controllers for s in c.sensors]
Use dynamic linked list to handle scenegraph rather than dumb scan
of the whole tree. The performance improvement depends on the fraction
of moving objects. If most objects are static, the speed up is
considerable. The following table compares the time spent on
scenegraph before and after this commit on a scene with 10000 objects
in various configuratons:
Scenegraph time (ms) Before After
(includes culling)
All objects static, 8.8 1.7
all visible but small fraction
in the view frustrum
All objects static, 7,5 0.01
all invisible.
All objects moving, 14.1 8.4
all visible but small fraction
in the view frustrum
This tables shows that static and invisible objects take no CPU at all
for scenegraph and culling. In the general case, this commit will
speed up the scenegraph between 2x and 5x. Compared to 2.48a, it should
be between 4x and 10x faster. Further speed up is possible by making
the scenegraph cache-friendly.
Next round of performance improvement will be on the rasterizer: use
the same dynamic linked list technique for the mesh slots.
