mutter/clutter/clutter-offscreen-effect.c

497 lines
16 KiB
C
Raw Normal View History

effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
/*
* Clutter.
*
* An OpenGL based 'interactive canvas' library.
*
* Copyright (C) 2010 Intel Corporation.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*
* Authors:
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
* Emmanuele Bassi <ebassi@linux.intel.com>
* Robert Bragg <robert@linux.intel.com>
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
*/
/**
* SECTION:clutter-offscreen-effect
* @short_description: Base class for effects using offscreen buffers
* @see_also: #ClutterBlurEffect, #ClutterEffect
*
* #ClutterOffscreenEffect is an abstract class that can be used by
* #ClutterEffect sub-classes requiring access to an offscreen buffer.
*
* Some effects, like the fragment shader based effects, can only use GL
* textures, and in order to apply those effects to any kind of actor they
* require that all drawing operations are applied to an offscreen framebuffer
* that gets redirected to a texture.
*
* #ClutterOffscreenEffect provides all the heavy-lifting for creating the
* offscreen framebuffer, the redirection and the final paint of the texture on
* the desired stage.
*
* <refsect2 id="ClutterOffscreenEffect-implementing">
* <title>Implementing a ClutterOffscreenEffect</title>
* <para>Creating a sub-class of #ClutterOffscreenEffect requires, in case
* of overriding the #ClutterEffect virtual functions, to chain up to the
* #ClutterOffscreenEffect's implementation.</para>
* <para>On top of the #ClutterEffect's virtual functions,
* #ClutterOffscreenEffect also provides a <function>paint_target()</function>
* function, which encapsulates the effective painting of the texture that
* contains the result of the offscreen redirection.</para>
* <para>The size of the target material is defined to be as big as the
* transformed size of the #ClutterActor using the offscreen effect.
* Sub-classes of #ClutterOffscreenEffect can change the texture creation
* code to provide bigger textures by overriding the
* <function>create_texture()</function> virtual function; no chain up
* to the #ClutterOffscreenEffect implementation is required in this
* case.</para>
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
* </refsect2>
*
* #ClutterOffscreenEffect is available since Clutter 1.4
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "clutter-offscreen-effect.h"
#include "cogl/cogl.h"
#include "clutter-actor-private.h"
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
#include "clutter-debug.h"
#include "clutter-private.h"
#include "clutter-stage-private.h"
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
struct _ClutterOffscreenEffectPrivate
{
CoglHandle offscreen;
CoglMaterial *target;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
ClutterActor *actor;
ClutterActor *stage;
gfloat x_offset;
gfloat y_offset;
gfloat target_width;
gfloat target_height;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
};
G_DEFINE_ABSTRACT_TYPE (ClutterOffscreenEffect,
clutter_offscreen_effect,
CLUTTER_TYPE_EFFECT);
static void
clutter_offscreen_effect_set_actor (ClutterActorMeta *meta,
ClutterActor *actor)
{
ClutterOffscreenEffect *self = CLUTTER_OFFSCREEN_EFFECT (meta);
ClutterOffscreenEffectPrivate *priv = self->priv;
ClutterActorMetaClass *meta_class;
meta_class = CLUTTER_ACTOR_META_CLASS (clutter_offscreen_effect_parent_class);
meta_class->set_actor (meta, actor);
/* clear out the previous state */
if (priv->offscreen != COGL_INVALID_HANDLE)
{
cogl_handle_unref (priv->offscreen);
priv->offscreen = COGL_INVALID_HANDLE;
}
/* we keep a back pointer here, to avoid going through the ActorMeta */
priv->actor = clutter_actor_meta_get_actor (meta);
}
static CoglHandle
clutter_offscreen_effect_real_create_texture (ClutterOffscreenEffect *effect,
gfloat width,
gfloat height)
{
return cogl_texture_new_with_size (MAX (width, 1), MAX (height, 1),
COGL_TEXTURE_NO_SLICING,
COGL_PIXEL_FORMAT_RGBA_8888_PRE);
}
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
static gboolean
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
update_fbo (ClutterEffect *effect, int fbo_width, int fbo_height)
{
ClutterOffscreenEffect *self = CLUTTER_OFFSCREEN_EFFECT (effect);
ClutterOffscreenEffectPrivate *priv = self->priv;
CoglHandle texture;
priv->stage = clutter_actor_get_stage (priv->actor);
if (priv->stage == NULL)
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
{
CLUTTER_NOTE (MISC, "The actor '%s' is not part of a stage",
clutter_actor_get_name (priv->actor) == NULL
? G_OBJECT_TYPE_NAME (priv->actor)
: clutter_actor_get_name (priv->actor));
return FALSE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
}
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
if (priv->target_width == fbo_width &&
priv->target_height == fbo_height &&
priv->offscreen != COGL_INVALID_HANDLE)
return TRUE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
if (priv->target == COGL_INVALID_HANDLE)
priv->target = cogl_material_new ();
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
texture =
clutter_offscreen_effect_create_texture (self, fbo_width, fbo_height);
if (texture == COGL_INVALID_HANDLE)
return FALSE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
cogl_material_set_layer (priv->target, 0, texture);
cogl_handle_unref (texture);
/* we need to use the size of the texture target and not the minimum
* size we passed to the create_texture() vfunc, as any sub-class might
* give use a bigger texture
*/
priv->target_width = cogl_texture_get_width (texture);
priv->target_height = cogl_texture_get_height (texture);
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
if (priv->offscreen != COGL_INVALID_HANDLE)
cogl_handle_unref (priv->offscreen);
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
priv->offscreen = cogl_offscreen_new_to_texture (texture);
if (priv->offscreen == COGL_INVALID_HANDLE)
{
g_warning ("%s: Unable to create an Offscreen buffer", G_STRLOC);
cogl_handle_unref (priv->target);
priv->target = COGL_INVALID_HANDLE;
priv->target_width = 0;
priv->target_height = 0;
return FALSE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
}
return TRUE;
}
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
static gboolean
clutter_offscreen_effect_pre_paint (ClutterEffect *effect)
{
ClutterOffscreenEffect *self = CLUTTER_OFFSCREEN_EFFECT (effect);
ClutterOffscreenEffectPrivate *priv = self->priv;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
ClutterActorBox box;
CoglMatrix projection;
CoglColor transparent;
CoglMatrix modelview;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
gfloat fbo_width, fbo_height;
gfloat width, height;
gfloat xexpand, yexpand;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
if (!clutter_actor_meta_get_enabled (CLUTTER_ACTOR_META (effect)))
return FALSE;
if (priv->actor == NULL)
return FALSE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
/* The paint box is the bounding box of the actor's paint volume in
* stage coordinates. This will give us the size for the framebuffer
* we need to redirect its rendering offscreen and its position will
* be used to setup an offset viewport */
if (clutter_actor_get_paint_box (priv->actor, &box))
clutter_actor_box_get_size (&box, &fbo_width, &fbo_height);
else
{
/* If we can't get a valid paint box then we fallback to
* creating a full stage size fbo. */
ClutterActor *stage = _clutter_actor_get_stage_internal (priv->actor);
clutter_actor_get_size (stage, &fbo_width, &fbo_height);
}
/* First assert that the framebuffer is the right size... */
if (!update_fbo (effect, fbo_width, fbo_height))
return FALSE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
/* get the current modelview matrix so that we can copy it
* to the framebuffer
*/
cogl_get_modelview_matrix (&modelview);
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
/* let's draw offscreen */
cogl_push_framebuffer (priv->offscreen);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
/* Set up the viewport so that it has the same size as the stage,
* but offset it so that the actor of interest lands on our
* framebuffer. */
clutter_actor_get_size (priv->stage, &width, &height);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
clutter_actor_box_get_origin (&box, &priv->x_offset, &priv->y_offset);
/* Expand the viewport if the actor is partially off-stage,
* otherwise the actor will end up clipped to the stage viewport
*/
xexpand = 0.f;
if (priv->x_offset < 0.f)
xexpand = -priv->x_offset;
if (priv->x_offset + fbo_width > width)
xexpand = MAX (xexpand, (priv->x_offset + fbo_width) - width);
yexpand = 0.f;
if (priv->y_offset < 0.f)
yexpand = -priv->y_offset;
if (priv->y_offset + fbo_height > height)
yexpand = MAX (yexpand, (priv->y_offset + fbo_height) - height);
/* Set the viewport */
cogl_set_viewport (-(priv->x_offset + xexpand), -(priv->y_offset + yexpand),
width + (2 * xexpand), height + (2 * yexpand));
/* Copy the stage's projection matrix across to the framebuffer */
_clutter_stage_get_projection_matrix (CLUTTER_STAGE (priv->stage),
&projection);
cogl_set_projection_matrix (&projection);
/* If we've expanded the viewport, make sure to scale the modelview
* matrix accordingly (as it's been initialised to work with the
* original viewport and not our expanded one).
*/
if (xexpand > 0.f || yexpand > 0.f)
{
CoglMatrix correction;
gfloat new_width, new_height;
new_width = width + (2 * xexpand);
new_height = height + (2 * yexpand);
cogl_matrix_init_identity (&correction);
cogl_matrix_scale (&correction,
width / new_width,
height / new_height,
1);
cogl_matrix_multiply (&correction, &correction, &modelview);
modelview = correction;
}
/* Copy the modelview that would have been used if rendering onscreen */
cogl_set_modelview_matrix (&modelview);
cogl_color_init_from_4ub (&transparent, 0, 0, 0, 0);
cogl_clear (&transparent,
COGL_BUFFER_BIT_COLOR |
COGL_BUFFER_BIT_DEPTH);
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
cogl_push_matrix ();
/* Override the actor's opacity to fully opaque - we paint the offscreen
* texture with the actor's paint opacity, so we need to do this to avoid
* multiplying the opacity twice.
*/
_clutter_actor_set_opacity_override (priv->actor, 0xff);
return TRUE;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
}
static void
clutter_offscreen_effect_real_paint_target (ClutterOffscreenEffect *effect)
{
ClutterOffscreenEffectPrivate *priv = effect->priv;
guint8 paint_opacity;
paint_opacity = clutter_actor_get_paint_opacity (priv->actor);
cogl_material_set_color4ub (priv->target,
paint_opacity,
paint_opacity,
paint_opacity,
paint_opacity);
cogl_set_source (priv->target);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 13:04:19 -04:00
/* At this point we are in stage coordinates translated so if
* we draw our texture using a textured quad the size of the paint
* box then we will overlay where the actor would have drawn if it
* hadn't been redirected offscreen.
*/
cogl_rectangle_with_texture_coords (0, 0,
priv->target_width,
priv->target_height,
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
0.0, 0.0,
1.0, 1.0);
}
static void
clutter_offscreen_effect_post_paint (ClutterEffect *effect)
{
ClutterOffscreenEffect *self = CLUTTER_OFFSCREEN_EFFECT (effect);
ClutterOffscreenEffectPrivate *priv = self->priv;
CoglMatrix modelview;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
if (priv->offscreen == COGL_INVALID_HANDLE ||
priv->target == COGL_INVALID_HANDLE ||
priv->actor == NULL)
return;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
cogl_pop_matrix ();
cogl_pop_framebuffer ();
cogl_push_matrix ();
/* Now reset the modelview to put us in stage coordinates so
* we can drawn the result of our offscreen render as a textured
* quad... */
cogl_matrix_init_identity (&modelview);
_clutter_actor_apply_modelview_transform (priv->stage, &modelview);
cogl_matrix_translate (&modelview, priv->x_offset, priv->y_offset, 0.0f);
cogl_set_modelview_matrix (&modelview);
/* Remove the opacity override */
_clutter_actor_set_opacity_override (priv->actor, -1);
/* paint the target material; this is virtualized for
* sub-classes that require special hand-holding
*/
clutter_offscreen_effect_paint_target (self);
cogl_pop_matrix ();
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
}
static void
clutter_offscreen_effect_finalize (GObject *gobject)
{
ClutterOffscreenEffect *self = CLUTTER_OFFSCREEN_EFFECT (gobject);
ClutterOffscreenEffectPrivate *priv = self->priv;
if (priv->offscreen)
cogl_handle_unref (priv->offscreen);
if (priv->target)
cogl_handle_unref (priv->target);
G_OBJECT_CLASS (clutter_offscreen_effect_parent_class)->finalize (gobject);
}
static void
clutter_offscreen_effect_class_init (ClutterOffscreenEffectClass *klass)
{
ClutterActorMetaClass *meta_class = CLUTTER_ACTOR_META_CLASS (klass);
ClutterEffectClass *effect_class = CLUTTER_EFFECT_CLASS (klass);
GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
g_type_class_add_private (klass, sizeof (ClutterOffscreenEffectPrivate));
klass->create_texture = clutter_offscreen_effect_real_create_texture;
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
klass->paint_target = clutter_offscreen_effect_real_paint_target;
meta_class->set_actor = clutter_offscreen_effect_set_actor;
effect_class->pre_paint = clutter_offscreen_effect_pre_paint;
effect_class->post_paint = clutter_offscreen_effect_post_paint;
gobject_class->finalize = clutter_offscreen_effect_finalize;
}
static void
clutter_offscreen_effect_init (ClutterOffscreenEffect *self)
{
self->priv = G_TYPE_INSTANCE_GET_PRIVATE (self,
CLUTTER_TYPE_OFFSCREEN_EFFECT,
ClutterOffscreenEffectPrivate);
}
/**
* clutter_offscreen_effect_get_target:
* @effect: a #ClutterOffscreenEffect
*
* Retrieves the material used as a render target for the offscreen
* buffer created by @effect
*
* You should only use the returned #CoglMaterial when painting. The
* returned material might change between different frames.
*
* Return value: (transfer none): a #CoglMaterial or %NULL. The
* returned material is owned by Clutter and it should not be
* modified or freed
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
*
* Since: 1.4
*/
CoglMaterial *
effect: Add OffscreenEffect The OffscreenEffect class is meant to be used to implement Effect sub-classes that create an offscreen framebuffer and redirect the actor's paint sequence there. The OffscreenEffect is useful for effects using fragment shaders. Any shader-based effect being applied to an actor through an offscreen buffer should be used before painting the resulting target material and not for every actor. This means that doing: pre_paint: cogl_program_use(program) set up offscreen buffer paint: [ actors ] → offscreen buffer → target material post_paint: paint target material cogl_program_use(null) Is not correct. Unfortunately, we cannot really do: post_paint: cogl_program_use(program) paint target material cogl_program_use(null) Because the OffscreenEffect::post_paint() implementation also pops the offscreen buffer and re-instates the previous framebuffer: post_paint: cogl_program_use(program) change frame buffer ← ouch! paint target material cogl_program_use(null) One way to fix it is to allow using the shader right before painting the target material - which means adding a new virtual inside the OffscreenEffect class vtable in additions to the ones defined by the parent Effect class. The newly-added paint_target() virtual allows the correct sequence of actions by adding an entry point for sub-classes to wrap the "paint target material" operation with custom code, in order to implement the case above correctly as: post_paint: change frame buffer cogl_program_use(program) paint target material cogl_program_use(null) The added upside is that sub-classes of OffscreenEffect involving shaders really just need to override the prepare() and paint_target() virtuals, since the pre_paint() and post_paint() do all that's needed.
2010-05-17 06:39:27 -04:00
clutter_offscreen_effect_get_target (ClutterOffscreenEffect *effect)
{
g_return_val_if_fail (CLUTTER_IS_OFFSCREEN_EFFECT (effect),
COGL_INVALID_HANDLE);
return effect->priv->target;
}
/**
* clutter_offscreen_effect_paint_target:
* @effect: a #ClutterOffscreenEffect
*
* Calls the paint_target() virtual function of the @effect
*
* Since: 1.4
*/
void
clutter_offscreen_effect_paint_target (ClutterOffscreenEffect *effect)
{
g_return_if_fail (CLUTTER_IS_OFFSCREEN_EFFECT (effect));
CLUTTER_OFFSCREEN_EFFECT_GET_CLASS (effect)->paint_target (effect);
}
/**
* clutter_offscreen_effect_create_texture:
* @effect: a #ClutterOffscreenEffect
* @width: the minimum width of the target texture
* @height: the minimum height of the target texture
*
* Calls the create_texture() virtual function of the @effect
*
* Return value: (transfer full): a handle to a Cogl texture, or
* %COGL_INVALID_HANDLE. The returned handle has its reference
* count increased.
*
* Since: 1.4
*/
CoglHandle
clutter_offscreen_effect_create_texture (ClutterOffscreenEffect *effect,
gfloat width,
gfloat height)
{
g_return_val_if_fail (CLUTTER_IS_OFFSCREEN_EFFECT (effect),
COGL_INVALID_HANDLE);
return CLUTTER_OFFSCREEN_EFFECT_GET_CLASS (effect)->create_texture (effect,
width,
height);
}
gboolean
_clutter_offscreen_effect_get_target_size (ClutterOffscreenEffect *effect,
gfloat *width,
gfloat *height)
{
ClutterOffscreenEffectPrivate *priv = effect->priv;
if (priv->target == NULL)
return FALSE;
if (width)
*width = priv->target_width;
if (height)
*height = priv->target_height;
return TRUE;
}