Source code for direct.filter.FilterManager


The FilterManager is a convenience class that helps with the creation
of render-to-texture buffers for image postprocessing applications.

See :ref:`generalized-image-filters` for information on how to use this class.

Still need to implement:

* Make sure sort-order of buffers is correct.
* Matching buffer size to original region instead of original window.
* Intermediate layer creation.
* Handling of window clears.
* Resizing of windows.
* Do something about window-size roundoff problems.


from panda3d.core import NodePath
from panda3d.core import Texture
from panda3d.core import CardMaker
from panda3d.core import GraphicsPipe, GraphicsOutput
from panda3d.core import WindowProperties, FrameBufferProperties
from panda3d.core import Camera
from panda3d.core import OrthographicLens
from panda3d.core import AuxBitplaneAttrib
from panda3d.core import LightRampAttrib
from direct.directnotify.DirectNotifyGlobal import *
from direct.showbase.DirectObject import DirectObject

__all__ = ["FilterManager"]

[docs]class FilterManager(DirectObject): notify = None
[docs] def __init__(self, win, cam, forcex=0, forcey=0): """ The FilterManager constructor requires you to provide a window which is rendering a scene, and the camera which is used by that window to render the scene. These are henceforth called the 'original window' and the 'original camera.' """ # Create the notify category if FilterManager.notify is None: FilterManager.notify = directNotify.newCategory("FilterManager") # Find the appropriate display region. region = None for dr in win.getDisplayRegions(): drcam = dr.getCamera() if drcam == cam: region = dr if region is None: self.notify.error('Could not find appropriate DisplayRegion to filter') return # Instance Variables. = win self.forcex = forcex self.forcey = forcey self.engine = win.getGsg().getEngine() self.region = region self.wclears = self.getClears( self.rclears = self.getClears(self.region) = cam self.caminit = cam.node().getInitialState() self.camstate = self.caminit self.buffers = [] self.sizes = [] self.nextsort = - 9 self.basex = 0 self.basey = 0 self.accept("window-event", self.windowEvent)
[docs] def getClears(self, region): clears = [] for i in range(GraphicsOutput.RTPCOUNT): clears.append((region.getClearActive(i), region.getClearValue(i))) return clears
[docs] def setClears(self, region, clears): for i in range(GraphicsOutput.RTPCOUNT): (active, value) = clears[i] region.setClearActive(i, active) region.setClearValue(i, value)
[docs] def setStackedClears(self, region, clears0, clears1): clears = [] for i in range(GraphicsOutput.RTPCOUNT): (active, value) = clears0[i] if not active: (active, value) = clears1[i] region.setClearActive(i, active) region.setClearValue(i, value) return clears
[docs] def isFullscreen(self): return ((self.region.getLeft() == 0.0) and (self.region.getRight() == 1.0) and (self.region.getBottom() == 0.0) and (self.region.getTop() == 1.0))
[docs] def getScaledSize(self, mul, div, align): """ Calculate the size of the desired window. Not public. """ winx = self.forcex winy = self.forcey if winx == 0: winx = if winy == 0: winy = if div != 1: winx = ((winx+align-1) // align) * align winy = ((winy+align-1) // align) * align winx = winx // div winy = winy // div if mul != 1: winx = int(round(winx * mul)) winy = int(round(winy * mul)) return winx,winy
[docs] def renderSceneInto(self, depthtex=None, colortex=None, auxtex=None, auxbits=0, textures=None, fbprops=None, clamping=None): """ Causes the scene to be rendered into the supplied textures instead of into the original window. Puts a fullscreen quad into the original window to show the render-to-texture results. Returns the quad. Normally, the caller would then apply a shader to the quad. To elaborate on how this all works: * An offscreen buffer is created. It is set up to mimic the original display region - it is the same size, uses the same clear colors, and contains a DisplayRegion that uses the original camera. * A fullscreen quad and an orthographic camera to render that quad are both created. The original camera is removed from the original window, and in its place, the orthographic quad-camera is installed. * The fullscreen quad is textured with the data from the offscreen buffer. A shader is applied that tints the results pink. * Automatic shader generation NOT enabled. If you have a filter that depends on a render target from the auto-shader, you either need to set an auto-shader attrib on the main camera or scene, or, you need to provide these outputs in your own shader. * All clears are disabled on the original display region. If the display region fills the whole window, then clears are disabled on the original window as well. It is assumed that rendering the full-screen quad eliminates the need to do clears. Hence, the original window which used to contain the actual scene, now contains a pink-tinted quad with a texture of the scene. It is assumed that the user will replace the shader on the quad with a more interesting filter. """ if textures: colortex = textures.get("color", None) depthtex = textures.get("depth", None) auxtex = textures.get("aux", None) auxtex0 = textures.get("aux0", auxtex) auxtex1 = textures.get("aux1", None) else: auxtex0 = auxtex auxtex1 = None if colortex is None: colortex = Texture("filter-base-color") colortex.setWrapU(Texture.WMClamp) colortex.setWrapV(Texture.WMClamp) texgroup = (depthtex, colortex, auxtex0, auxtex1) # Choose the size of the offscreen buffer. (winx, winy) = self.getScaledSize(1,1,1) if fbprops is not None: buffer = self.createBuffer("filter-base", winx, winy, texgroup, fbprops=fbprops) else: buffer = self.createBuffer("filter-base", winx, winy, texgroup) if buffer is None: return None cm = CardMaker("filter-base-quad") cm.setFrameFullscreenQuad() quad = NodePath(cm.generate()) quad.setDepthTest(0) quad.setDepthWrite(0) quad.setTexture(colortex) quad.setColor(1, 0.5, 0.5, 1) cs = NodePath("dummy") cs.setState(self.camstate) # Do we really need to turn on the Shader Generator? #cs.setShaderAuto() if auxbits: cs.setAttrib(AuxBitplaneAttrib.make(auxbits)) if clamping is False: # Disables clamping in the shader generator. cs.setAttrib(LightRampAttrib.make_identity()) quadcamnode = Camera("filter-quad-cam") lens = OrthographicLens() lens.setFilmSize(2, 2) lens.setFilmOffset(0, 0) lens.setNearFar(-1000, 1000) quadcamnode.setLens(lens) quadcam = quad.attachNewNode(quadcamnode) self.region.setCamera(quadcam) self.setStackedClears(buffer, self.rclears, self.wclears) if auxtex0: buffer.setClearActive(GraphicsOutput.RTPAuxRgba0, 1) buffer.setClearValue(GraphicsOutput.RTPAuxRgba0, (0.5, 0.5, 1.0, 0.0)) if auxtex1: buffer.setClearActive(GraphicsOutput.RTPAuxRgba1, 1) self.region.disableClears() if self.isFullscreen(): dr = buffer.makeDisplayRegion() dr.disableClears() dr.setCamera( dr.setActive(1) self.buffers.append(buffer) self.sizes.append((1, 1, 1)) return quad
[docs] def renderQuadInto(self, name="filter-stage", mul=1, div=1, align=1, depthtex=None, colortex=None, auxtex0=None, auxtex1=None, fbprops=None): """ Creates an offscreen buffer for an intermediate computation. Installs a quad into the buffer. Returns the fullscreen quad. The size of the buffer is initially equal to the size of the main window. The parameters 'mul', 'div', and 'align' can be used to adjust that size. """ texgroup = (depthtex, colortex, auxtex0, auxtex1) winx, winy = self.getScaledSize(mul, div, align) depthbits = int(depthtex is not None) if fbprops is not None: buffer = self.createBuffer(name, winx, winy, texgroup, depthbits, fbprops=fbprops) else: buffer = self.createBuffer(name, winx, winy, texgroup, depthbits) if buffer is None: return None cm = CardMaker("filter-stage-quad") cm.setFrameFullscreenQuad() quad = NodePath(cm.generate()) quad.setDepthTest(0) quad.setDepthWrite(0) quad.setColor(1, 0.5, 0.5, 1) quadcamnode = Camera("filter-quad-cam") lens = OrthographicLens() lens.setFilmSize(2, 2) lens.setFilmOffset(0, 0) lens.setNearFar(-1000, 1000) quadcamnode.setLens(lens) quadcam = quad.attachNewNode(quadcamnode) dr = buffer.makeDisplayRegion((0, 1, 0, 1)) dr.disableClears() dr.setCamera(quadcam) dr.setActive(True) dr.setScissorEnabled(False) # This clear stage is important if the buffer is padded, so that # any pixels accidentally sampled in the padded region won't # be reading from unititialised memory. buffer.setClearColor((0, 0, 0, 1)) buffer.setClearColorActive(True) self.buffers.append(buffer) self.sizes.append((mul, div, align)) return quad
[docs] def createBuffer(self, name, xsize, ysize, texgroup, depthbits=True, fbprops=None): """ Low-level buffer creation. Not intended for public use. """ winprops = WindowProperties() winprops.setSize(xsize, ysize) props = FrameBufferProperties(FrameBufferProperties.getDefault()) props.setBackBuffers(0) props.setRgbColor(1) if depthbits is True: # Respect depth-bits from Config.prc if props.getDepthBits() == 0: props.setDepthBits(1) else: props.setDepthBits(depthbits) props.setStereo( if fbprops is not None: props.addProperties(fbprops) depthtex, colortex, auxtex0, auxtex1 = texgroup if auxtex0 is not None: props.setAuxRgba(1) if auxtex1 is not None: props.setAuxRgba(2) buffer=self.engine.makeOutput(, name, -1, props, winprops, GraphicsPipe.BFRefuseWindow | GraphicsPipe.BFResizeable,, if buffer is None: return buffer if depthtex: buffer.addRenderTexture(depthtex, GraphicsOutput.RTMBindOrCopy, GraphicsOutput.RTPDepth) if colortex: buffer.addRenderTexture(colortex, GraphicsOutput.RTMBindOrCopy, GraphicsOutput.RTPColor) if auxtex0: buffer.addRenderTexture(auxtex0, GraphicsOutput.RTMBindOrCopy, GraphicsOutput.RTPAuxRgba0) if auxtex1: buffer.addRenderTexture(auxtex1, GraphicsOutput.RTMBindOrCopy, GraphicsOutput.RTPAuxRgba1) buffer.setSort(self.nextsort) buffer.disableClears() self.nextsort += 1 return buffer
[docs] def windowEvent(self, win): """ When the window changes size, automatically resize all buffers """ self.resizeBuffers()
[docs] def resizeBuffers(self): """ Resize all buffers to match the size of the window. """ for i in range(len(self.buffers)): (mul, div, align) = self.sizes[i] (xsize, ysize) = self.getScaledSize(mul, div, align) self.buffers[i].setSize(xsize, ysize)
[docs] def cleanup(self): """ Restore everything to its original state, deleting any new buffers in the process. """ for buffer in self.buffers: buffer.clearRenderTextures() self.engine.removeWindow(buffer) self.buffers = [] self.sizes = [] self.setClears(, self.wclears) self.setClears(self.region, self.rclears) self.camstate = self.caminit self.region.setCamera( if hasattr(self.region, 'clearCullResult'): self.region.clearCullResult() self.nextsort = - 9 self.basex = 0 self.basey = 0
#snake_case alias: is_fullscreen = isFullscreen resize_buffers = resizeBuffers set_stacked_clears = setStackedClears render_scene_into = renderSceneInto get_scaled_size = getScaledSize render_quad_into = renderQuadInto get_clears = getClears set_clears = setClears create_buffer = createBuffer window_event = windowEvent