(Looks like big commit, but is mostly just change of API call!)
Particle emittors now can be parented to an armature Bone, and give the
correct path for each particle. Note that this doesn't work for deform!
And, for each particle the entire armature is evaluated, all actions and
NLA strips.
It used to work a little while BTW, but the code just called ALL armatures
and made ALL deforms again. Was quite slow... thats why the API call
change: do_all_actions() now accepts Object * to only do that one. With
a NULL it now does all still. Will disapppear in recode of course...
Envmap rendering was using Gauss filtering, causing dark outlines in edges
of cube pictures. Now it temporally disables gauss in envmap render, which
isn't really noticable since the pics are filtered + gaussed anyway in the
end result.
Underflow case in Blinn spec... resulting in large negative values to be
returned in certain cases, causing (again!) black dots.
Thanks randall for clear report with nice demo file. :)
With a sun lamp shining exact perpendicular to a face, the result was
undefined (black or full lit). Was caused by using acos() without checking
the input range. Float calculus can result in small under or overflows...
therefore: always use the arithb.c function saacos() !!!
Found more acos() usage in rendercore (in blinn and arealight too), removed
all of them. Might solve the white or black dots people were complaining
about?
- cancelled previous commit to add RE_findTFAce, instead just added
a MemArena to render struct... free'd at end of render, can be used
to store other data as well
- switch rendering to using DerivedMesh API... this is slightly more
inefficient now because it is doing some unnecessary copying. Can
be fixed by defining a DerivedMesh function to convert the object
into a render object (on todo list)
are owned by mesh (or displistmesh)... this causes problems for
adapting to systems that build tfaces on the fly. Added RE_findTFace
function to allow allocating tfaces inside renderer itself.
The gamma functions in gammaCorrectionTables.c cannot be used to correct
and correct back with identical results... causing banding in rendering
pictures with halos.
NOTE: BLI_winstuff.h was meant to be a wrapper around windows.h to handle
undefining various crap that windows.h defines. Platform specific headers
should only have to be included in a few places. This reduces the number
of inclusions of BLI_winstuff.h to 16 which is a much more reasonable
number (than the 144 or whatever it used to be)
- removed reference in render.h (really bad, shouldn't include a platform
specific header so widely unless really necessary)
- added M_PI, M_PI_2, M_SQRT, M_SQRT_2 defines to BLI_arithb.h... this is
a better place as it is more the "standard" blender math header. left
in winstuff.h as well for the moment for simplicity
- other changes are patches to code so everything works ok with this
shuffling.
- Made framework for 3d Transform Manipulators (widgets)
- The Manipulators act like '2d buttons', by default with LeftMouse and with
while-hold-move-release
- Implemented now: Translation Widget, which allows:
- four hotspots for axis grab or view-aligned grab
- center defined by 'around' setting
- SHIFT+LMB gives planar constraint on other 2 axes
- works in Object mode and Edit mode (not posemode yet)
Enable it with (temporal) icon in 3D header. All other 'normal' transforms
then keeps working btw.
On the todo for this widget:
- choice for Global, Local or Normal orientation
The way the widgets are going to work is in review still. Commit is also for
Matt for his proposal/paper on topic.
Some notes regarding this project;
- no life updates (on mouse over) like 3DS, I think that's neurotic
- on click, dominant axis changes to theme defined "Transform" color, other axes disappear, like maya
- manipulater size is fixed preset, independent zoom.
- manipulator follows selection, and is located based on 'around' mode
Unresolved;
- in maya, when you select 2 or more objects, the manipulator draws on the 'active' object, also interesting
- what to do with G,R,S hotkeys? It could switch the manipulator "mode"...
- header button/menu for manipulator mode?
Halo render had too narrow clipping for left/right or top/bottom part of
image. Was caused by fix for 2.36, incorrect clipping for distance...
distance was OK, but the old hack should be still there for X or Y clip!
Bumpmaps for skin on Env's dinosaurs appeared to be less nice in 2.36. This
was caused by the bugfix to make bumpmapping correct for rotations, which
should only work for flat/cube mapping. It also rotated it for sphere/tube
though, which gives less interesting bumps.
So; now the correction is skipped for tube/sphere mapping bumping.
Wood/marble now have three waveforms to choose from: Sine, Saw and Triangle.
The Saw wave allows for much more realistic wood, especially in combination
with a ColorBand. A blender3d.org release page is being constructed about it.
Added: commit in editmesh_add.c to remove circle warning in face-select mode.
The dynamical allocated arrays (rwenzlaff, april 03) for render vertices/
faces/halos were freed in a way that requires the last element of array to
be NULL.
This wasn't coded however, causing memory errors in exceptional cases;
like when amount of faces/vertices is exact between 261888 and 262144. :)
Environment maps type "Load" don't need to have an "Ob" object set to
render. Without object it renders applied envmaps like reflection maps.
Please note that envmaps then still "mirror" realistic, provided the
original object the envmap was rendered with was not rotated.
border selected.
I changed the test to this:
if(R.r.border.xmax <= R.r.border.xmin ||
R.r.border.ymax <= R.r.border.ymin) {
Original patch tested xmin=0 which isn't quite right.
Kent
#2187: Append to active layer now uses the real active layer
#2191: Stars didnt show up correctly in envmap render
#2194: Startfame > Endframe in anim render crashed (now it gives error)
It used to be a simple hack, scaling lens with 100, and moving the camera
to the back with an equivalent amount.
Because of the hack, making it 100% compatible with older files I could not
achieve (yet?). To help reminding users, I've added a print when reading
old files with Ortho cameras.
Full description of how it works can be found here;
http://www.blender3d.com/cms/Render_changes.515.0.html
Removed it in OSA from special thread loop now, making it unthreaded but
render OK. Main reason; subpixel-based adding with gauss doent work
for halos yet.
Unified render does that though, but result is just filter, no real
subpixel.
that seems not to be solved on intel platforms.
Previously, a positive signed integer overflow addition was checked by
assuming it to become negative. Now it checks 'safe' with this version:
if(vergz > 0x7FFFFFF0 - dz) vergz= 0x7FFFFFF0;
else vergz+= dz;
Reason why using signed ints is cumbersome... overflows for unsigned are
much easer to find.
Had moved normalizing the view vector just a couple of lines to much...
causing OSA vectors to be calculated wrong, and image texture to show
much to filtered (nearly invisible)
Object "time" ipo was evaluated before it applied offset for field and
motion blur. Should be after! (ancient one)
in buttons_editing.c: removed sound include file
- removed ugly pointerhack from OSA render (negative indices denoted
pointers). this should solve memory errors when using >1.5 gig mem
- cleaned up usage of zbuffer values. These are signed in Blender, and
treated as unsigned all over, giving confusing code
- fixed incorrect gamma-adding for halos (caused in after xmas commit)
And bugfix #2101; wire render didn't give correct rendering for mist.
This caused by fact wires are 2D pixel lines, and not correctly filled
in faces. Retrieving the 3d coordinate while render cannot use a face-
equation then. Solved by retrieving 3D coordinate based on zbuffer value.
Still todo here: calculating correct texture coordinates for wire-edges
that are no faces.
When loading a file with linked external scene, and that external scene
didnt exist anymore, and that scene was active in the file -> crash!
Render code; changed to use local 'puno' flag for threaded render.
Alpha for halos was always added, even when "Add" slider was zero. Fixed.
(also; made BROW define BROW16, to distinguish from int version in
BKE_utildefines.h)
For normal maps only:
Slider for "nor" mapping in "Map to" panel allowed values for over 1.0 in
preview render (not for render itself, causing confusement).
- zbuf; error caused by using unsigned int as counter (while >= 0) in radio
- backbuffer didnt load in initrender (was removed from sky render)
- unitialized variable in envmap.c
Render:
- New; support for dual CPU render (SDL thread)
Currently only works with alternating scanlines, but gives excellent
performance. For both normal render as unified implemented.
Note the "mutex" locks on z-transp buffer render and imbuf loads.
- This has been made possible by major cleanups in render code, especially
getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct
OSA or using Materials or Texture data to write to.
- Made normal render fully 4x32 floats too, and removed all old optimizes
with chars or shorts.
- Made normal render and unified render use same code for sky and halo
render, giving equal (and better) results for halo render. Old render
now also uses PostProcess options (brightness, mul, gamma)
- Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer
after render. Using PostProcess menu you will note an immediate re-
display of image too (32 bits RGBA)
- Added "Hue" and "Saturation" sliders to PostProcess options
- Render module is still not having a "nice" API, but amount of dependencies
went down a lot. Next todo: remove abusive "previewrender" code.
The last main global in Render (struct Render) now can be re-used for fully
controlling a render, to allow multiple "instances" of render to open.
- Renderwindow now displays a smal bar on top with the stats, and keeps the
stats after render too. Including "spare" page support.
Not only easier visible that way, but also to remove the awkward code that
was drawing stats in the Info header (extreme slow on some ATIs too)
- Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping
defines.
- I might have forgotten stuff... and will write a nice doc on the architecture!
It appeared that the method as used in unified render for "render all sub-
pixels" isn't very well resistant to having large amounts of faces in a
single pixel. The bug file had about 16x70x2 faces per pixel... causing
tremendous slowdown and even wrong render.
I've disabled the option (was coded by Nzc in NaN days) and made it
sampling in the main render loop. Goes much faster, error free. Only loss
is in that it doesnt use superiour subsamples for gauss anymore. Here the
normal render performs slightly better. Not a real issue though. Hard to
notice.
Combination of:
- render movie
- with border set
- and "crop" set
Crashed.
Was due to feeding movie initialize code wrong frame sizes. Another oldie!
Also found that crop+border+parts doesnt work. This now is prevented with
warning and return.
