blender/release/scripts/bpymodules/BPyRender.py
Campbell Barton dcc834f3fa updated render settings to support recent changes.
options like saveBuffers are available from Python and threads can be set from 1 to 8
usefull for python based renderfarms.
2007-01-30 03:02:58 +00:00

500 lines
14 KiB
Python

import Blender
from Blender import Scene, sys, Camera, Object, Image
from Blender.Scene import Render
Vector= Blender.Mathutils.Vector
def extFromFormat(format):
if format == Render.TARGA: return 'tga'
if format == Render.RAWTGA: return 'tga'
if format == Render.HDR: return 'hdr'
if format == Render.PNG: return 'png'
if format == Render.BMP: return 'bmp'
if format == Render.JPEG: return 'jpg'
if format == Render.HAMX: return 'ham'
if format == Render.TIFF: return 'tif'
if format == Render.CINEON: return 'cine'
if format == Render.DPX: return 'tif'
if format == Render.OPENEXR: return 'exr'
if format == Render.IRIS: return 'rgb'
return ''
def imageFromObjectsOrtho(objects, path, width, height, smooth, alpha= True, camera_matrix= None, format=Render.PNG):
'''
Takes any number of objects and renders them on the z axis, between x:y-0 and x:y-1
Usefull for making images from a mesh without per pixel operations
- objects must be alredy placed
- smooth, anti alias True/False
- path renders to a PNG image
- alpha weather to render background as alpha
returns the blender image
'''
ext = '.' + extFromFormat(format)
print ext
# remove an extension if its alredy there
if path.lower().endswith(ext):
path= path[:-4]
path_expand= sys.expandpath(path) + ext
print path_expand, 'path'
# Touch the path
try:
f= open(path_expand, 'w')
f.close()
except:
raise 'Error, could not write to path:' + path_expand
# RENDER THE FACES.
scn= Scene.GetCurrent()
render_scn= Scene.New()
render_scn.makeCurrent()
render_scn.Layers |= (1<<20)-1 # all layers enabled
# Add objects into the current scene
for ob in objects:
render_scn.link(ob)
render_context= render_scn.getRenderingContext()
render_context.setRenderPath('') # so we can ignore any existing path and save to the abs path.
render_context.imageSizeX(width)
render_context.imageSizeY(height)
if smooth:
render_context.enableOversampling(True)
render_context.setOversamplingLevel(16)
else:
render_context.enableOversampling(False)
render_context.setRenderWinSize(100)
render_context.setImageType(format)
render_context.enableExtensions(True)
#render_context.enableSky() # No alpha needed.
if alpha:
render_context.alphaMode= 1
render_context.enableRGBAColor()
else:
render_context.alphaMode= 0
render_context.enableRGBColor()
render_context.displayMode= 0 # fullscreen
# New camera and object
render_cam_data= Camera.New('ortho')
render_cam_ob= Object.New('Camera')
render_cam_ob.link(render_cam_data)
render_scn.link(render_cam_ob)
render_scn.setCurrentCamera(render_cam_ob)
render_cam_data.type= 'ortho'
# Position the camera
if camera_matrix:
render_cam_ob.setMatrix(camera_matrix)
# We need to take into account the matrix scaling when setting the size
# so we get the image bounds defined by the matrix
# first get the x and y factors from the matrix.
# To render the correct dimensions we must use the aspy and aspy to force the matrix scale to
# override the aspect enforced by the width and weight.
cent= Vector() * camera_matrix
xvec= Vector(1,0,0) * camera_matrix
yvec= Vector(0,1,0) * camera_matrix
# zvec= Vector(0,0,1) * camera_matrix
xlen = (cent-xvec).length # half height of the image
ylen = (cent-yvec).length # half width of the image
# zlen = (cent-zvec).length # dist to place the camera? - just use the loc for now.
# less then 1.0 portrate, 1.0 or more is portrate
asp_cam_mat= xlen/ylen # divide by zero? - possible but scripters fault.
asp_image_res= float(width)/height
#print 'asp quad', asp_cam_mat, 'asp_image', asp_image_res
#print 'xylen', xlen, ylen, 'w/h', width, height
# Setup the aspect
if asp_cam_mat > asp_image_res:
# camera is wider then image res.
# to make the image wider, reduce the aspy
asp_diff= asp_image_res/asp_cam_mat
min_asp= int(round(asp_diff * 200))
#print 'X', min_asp
elif asp_cam_mat < asp_image_res: # asp_cam_mat < asp_image_res
# camera is narrower then image res
# to make the image narrower, reduce the aspx
asp_diff= asp_cam_mat/asp_image_res
min_asp= int(round(asp_diff * 200))
#print 'Y', min_asp
else:
min_asp= 200
# set the camera size
if xlen > ylen:
if asp_cam_mat > asp_image_res:
render_context.aspectX= 200 # get the greatest range possible
render_context.aspectY= min_asp # get the greatest range possible
else:
render_context.aspectY= 200 # get the greatest range possible
render_context.aspectX= min_asp # get the greatest range possible
#print "xlen bigger"
render_cam_data.scale= xlen * 2
elif xlen < ylen:# ylen is bigger
if asp_cam_mat > asp_image_res:
render_context.aspectX= 200 # get the greatest range possible
render_context.aspectY= min_asp # get the greatest range possible
else:
render_context.aspectY= 200 # get the greatest range possible
render_context.aspectX= min_asp # get the greatest range possible
#print "ylen bigger"
render_cam_data.scale= ylen *2
else:
# asppect 1:1
#print 'NOLEN Bigger'
render_cam_data.scale= xlen * 2
#print xlen, ylen, 'xlen, ylen'
else:
if width > height:
min_asp = int((float(height) / width) * 200)
render_context.aspectX= min_asp
render_context.aspectY= 200
else:
min_asp = int((float(width) / height) * 200)
render_context.aspectX= 200
render_context.aspectY= min_asp
render_cam_data.scale= 1.0
render_cam_ob.LocZ= 1.0
render_cam_ob.LocX= 0.5
render_cam_ob.LocY= 0.5
Blender.Window.RedrawAll()
render_context.threads= 2 # good for dual core cpu's
render_context.render()
render_context.saveRenderedImage(path)
Render.CloseRenderWindow()
#if not B.sys.exists(PREF_IMAGE_PATH_EXPAND):
# raise 'Error!!!'
scn.makeCurrent()
Scene.Unlink(render_scn)
# NOW APPLY THE SAVED IMAGE TO THE FACES!
#print PREF_IMAGE_PATH_EXPAND
try:
target_image= Image.Load(path_expand)
return target_image
except:
raise 'Error: Could not render or load the image at path "%s"' % path_expand
return
#-----------------------------------------------------------------------------#
# UV Baking functions, make a picture from mesh(es) uvs #
#-----------------------------------------------------------------------------#
def mesh2uv(me_s, PREF_SEL_FACES_ONLY=False):
'''
Converts a uv mapped mesh into a 2D Mesh from UV coords.
returns a triple -
(mesh2d, face_list, col_list)
"mesh" is the new mesh and...
"face_list" is the faces that were used to make the mesh,
"material_list" is a list of materials used by each face
These are in alligned with the meshes faces, so you can easerly copy data between them
'''
render_me= Blender.Mesh.New()
render_me.verts.extend( [Vector(0,0,0),] ) # 0 vert uv bugm dummy vert
face_list= []
material_list= []
for me in me_s:
me_materials= me.materials
FACE_SEL= Blender.Mesh.FaceFlags.SELECT
if PREF_SEL_FACES_ONLY:
me_faces= [f for f in me.faces if f.flag & FACE_SEL]
else:
me_faces= me.faces
face_list.extend(me_faces)
# Dittro
if me_materials:
material_list.extend([me_materials[f.mat] for f in me_faces])
else:
material_list.extend([None]*len(me_faces))
# Now add the verts
render_me.verts.extend( [ Vector(uv.x, uv.y, 0) for f in face_list for uv in f.uv ] )
# Now add the faces
tmp_faces= []
vert_offset= 1
for f in face_list:
tmp_faces.append( [ii+vert_offset for ii in xrange(len(f))] )
vert_offset+= len(f)
render_me.faces.extend(tmp_faces)
render_me.faceUV=1
return render_me, face_list, material_list
def uvmesh_apply_normals(render_me, face_list):
'''Worldspace normals to vertex colors'''
for i, f in enumerate(render_me.faces):
face_orig= face_list[i]
f_col= f.col
for j, v in enumerate(face_orig):
c= f_col[j]
nx, ny, nz= v.no
c.r= int((nx+1)*128)-1
c.g= int((ny+1)*128)-1
c.b= int((nz+1)*128)-1
def uvmesh_apply_image(render_me, face_list):
'''Copy the image and uvs from the original faces'''
for i, f in enumerate(render_me.faces):
f.uv= face_list[i].uv
f.image= face_list[i].image
def uvmesh_apply_vcol(render_me, face_list):
'''Copy the vertex colors from the original faces'''
for i, f in enumerate(render_me.faces):
face_orig= face_list[i]
f_col= f.col
for j, c_orig in enumerate(face_orig.col):
c= f_col[j]
c.r= c_orig.r
c.g= c_orig.g
c.b= c_orig.b
def uvmesh_apply_matcol(render_me, material_list):
'''Get the vertex colors from the original materials'''
for i, f in enumerate(render_me.faces):
mat_orig= material_list[i]
f_col= f.col
if mat_orig:
for c in f_col:
c.r= int(mat_orig.R*255)
c.g= int(mat_orig.G*255)
c.b= int(mat_orig.B*255)
else:
for c in f_col:
c.r= 255
c.g= 255
c.b= 255
def uvmesh_apply_col(render_me, color):
'''Get the vertex colors from the original materials'''
r,g,b= color
for i, f in enumerate(render_me.faces):
f_col= f.col
for c in f_col:
c.r= r
c.g= g
c.b= b
def vcol2image(me_s,\
PREF_IMAGE_PATH,\
PREF_IMAGE_SIZE,\
PREF_IMAGE_BLEED,\
PREF_IMAGE_SMOOTH,\
PREF_IMAGE_WIRE,\
PREF_IMAGE_WIRE_INVERT,\
PREF_IMAGE_WIRE_UNDERLAY,\
PREF_USE_IMAGE,\
PREF_USE_VCOL,\
PREF_USE_MATCOL,\
PREF_USE_NORMAL,\
PREF_USE_TEXTURE,\
PREF_SEL_FACES_ONLY):
def rnd_mat():
render_mat= Blender.Material.New()
mode= render_mat.mode
# Dont use lights ever
mode |= Blender.Material.Modes.SHADELESS
if PREF_IMAGE_WIRE:
# Set the wire color
if PREF_IMAGE_WIRE_INVERT:
render_mat.rgbCol= (1,1,1)
else:
render_mat.rgbCol= (0,0,0)
mode |= Blender.Material.Modes.WIRE
if PREF_USE_VCOL or PREF_USE_MATCOL or PREF_USE_NORMAL: # both vcol and material color use vertex cols to avoid the 16 max limit in materials
mode |= Blender.Material.Modes.VCOL_PAINT
if PREF_USE_IMAGE:
mode |= Blender.Material.Modes.TEXFACE
# Copy back the mode
render_mat.mode |= mode
return render_mat
render_me, face_list, material_list= mesh2uv(me_s, PREF_SEL_FACES_ONLY)
# Normals exclude all others
if PREF_USE_NORMAL:
uvmesh_apply_normals(render_me, face_list)
else:
if PREF_USE_IMAGE:
uvmesh_apply_image(render_me, face_list)
uvmesh_apply_vcol(render_me, face_list)
elif PREF_USE_VCOL:
uvmesh_apply_vcol(render_me, face_list)
elif PREF_USE_MATCOL:
uvmesh_apply_matcol(render_me, material_list)
elif PREF_USE_TEXTURE:
# if we have more then 16 materials across all the mesh objects were stuffed :/
# get unique materials
tex_unique_materials= dict([(mat.name, mat) for mat in material_list]).values()[:16] # just incase we have more then 16
tex_me= Blender.Mesh.New()
# Backup the original shadless setting
tex_unique_materials_shadeless= [ mat.mode & Blender.Material.Modes.SHADELESS for mat in tex_unique_materials ]
# Turn shadeless on
for mat in tex_unique_materials:
mat.mode |= Blender.Material.Modes.SHADELESS
# Assign materials
render_me.materials= tex_unique_materials
tex_material_indicies= dict([(mat.name, i) for i, mat in enumerate(tex_unique_materials)])
tex_me.verts.extend([Vector(0,0,0),]) # dummy
tex_me.verts.extend( [ Vector(v.co) for f in face_list for v in f ] )
# Now add the faces
tmp_faces= []
vert_offset= 1
for f in face_list:
tmp_faces.append( [ii+vert_offset for ii in xrange(len(f))] )
vert_offset+= len(f)
tex_me.faces.extend(tmp_faces)
# Now we have the faces, put materials and normal, uvs into the mesh
if len(tex_me.faces) != len(face_list):
# Should never happen
raise "Error face length mismatch"
# Copy data to the mesh that could be used as texture coords
for i, tex_face in enumerate(tex_me.faces):
orig_face= face_list[i]
# Set the material index
try:
render_face.mat= tex_material_indicies[ material_list[i].name ]
except:
# more then 16 materials
pass
# set the uvs on the texmesh mesh
tex_face.uv= orig_face.uv
orig_face_v= orig_face.v
# Set the normals
for j, v in enumerate(tex_face):
v.no= orig_face_v[j].no
# Set the texmesh
render_me.texMesh= tex_me
# END TEXMESH
# Handel adding objects
render_ob= Blender.Object.New('Mesh')
render_ob.link(render_me)
if not PREF_USE_TEXTURE: # textures use the original materials
render_me.materials= [rnd_mat()]
obs= [render_ob]
if PREF_IMAGE_WIRE_UNDERLAY:
# Make another mesh with the material colors
render_me_under, face_list, material_list= mesh2uv(me_s, PREF_SEL_FACES_ONLY)
uvmesh_apply_matcol(render_me_under, material_list)
# Handel adding objects
render_ob= Blender.Object.New('Mesh')
render_ob.link(render_me_under)
render_ob.LocZ= -0.01
# Add material and disable wire
mat= rnd_mat()
mat.rgbCol= 1,1,1
mat.alpha= 0.5
mat.mode &= ~Blender.Material.Modes.WIRE
mat.mode |= Blender.Material.Modes.VCOL_PAINT
render_me_under.materials= [mat]
obs.append(render_ob)
elif PREF_IMAGE_BLEED and not PREF_IMAGE_WIRE:
# EVIL BLEEDING CODE!! - Just do copys of the mesh and place behind. Crufty but better then many other methods I have seen. - Cam
BLEED_PIXEL= 1.0/PREF_IMAGE_SIZE
z_offset= 0.0
for i in xrange(PREF_IMAGE_BLEED):
for diag1, diag2 in ((-1,-1),(-1,1),(1,-1),(1,1), (1,0), (0,1), (-1,0), (0, -1)): # This line extends the object in 8 different directions, top avoid bleeding.
render_ob= Blender.Object.New('Mesh')
render_ob.link(render_me)
render_ob.LocX= (i+1)*diag1*BLEED_PIXEL
render_ob.LocY= (i+1)*diag2*BLEED_PIXEL
render_ob.LocZ= -z_offset
obs.append(render_ob)
z_offset += 0.01
image= imageFromObjectsOrtho(obs, PREF_IMAGE_PATH, PREF_IMAGE_SIZE, PREF_IMAGE_SIZE, PREF_IMAGE_SMOOTH)
# Clear from memory as best as we can
render_me.verts= None
if PREF_IMAGE_WIRE_UNDERLAY:
render_me_under.verts= None
if PREF_USE_TEXTURE:
tex_me.verts= None
# Restire Shadeless setting
for i, mat in enumerate(tex_unique_materials):
# we know there all on so turn it off of its not set
if not tex_unique_materials_shadeless[i]:
mat.mode &= ~Blender.Material.Modes.SHADELESS
return image