The issue was caused by intermediate DerivedMesh being created with
scene's Simplify settings taken into account. This is what happens
when one area makes implicit decisions based on whether passed Scene
pointer is not NULL.
Made it so ignoring simplification serttings is an explicit flag,
which makes it easier to follow what's going on.
This is actually a workaround for the crash in OpenSubdiv.
Topology refiner will have a crash when special conditions
are met:
- Refiner is configured to use infinitely sharp patches.
- Refinement happens for the level 1 (which we call Quality 1 on
Blender side).
- Mesh has non-quad faces.
The workaround is to force refinement to happen to level 2 (or
quality 2 on Blender side) when those conditions are met.
Later on with the next OpenSubdiv update we can remove this
workaround, since there was work done on OpenSubdiv side to
deal better with such configurations.
The modifier will now be somewhat slower, but this will be
compensated with upcoming topology cache enabled by default.
The workaround is done when initializing settings, so the
comparison of topology refiner settings is happening without
any extra workarounds there.
Not only were those often making doublons with already existing
BLI_math's stuff, but they were also used to hide implicit type
conversions...
As usual this adds some more exotic inlined vector functions (one of
the rare cases where I really miss C++ and its templates... ;) ).
Before that only normal component was averaged, which is not
really correct.
Unfortunately, the new code is somewhat slower due to more
involved math to deal properly with non-quad faces, but the
plan is to move averaging from runtime to edit time, This
means, that mdisps will always be continuous around the edges
and no averaging on every frame change of animated character
will be needed.
The idea is to run reshaping for every boundary vertex
of a grid rather than trying to copy boundary grid
elements.
While this is somewhat slower, this avoids all this
tangent flipping magic, which tempts to be rather tricky
and fragile.
The boundary copy code was not dealing correct with flipping
tangent vectors, hence causing discontinuity in the final
positions.
Now we only copy boundaries for quads, where we know how to
deal with tangent vectors and where we know that this is
needed.
More clear solution could be to change the code in a way that
handles handles displacement grids of quads in the same way
as it's done for non-quad faces.
Some of Eevee's Bloom defaults are not very good for physically based rendering. This patches addresses this issue.
This picture shows one of the problems with current default. Bloom looks very foggy:
{F6280495}
Even worse, light emitters much dimmer than the Sun can make everything equally hazy if Clamp is set to 1.0 and intensity to 0.8 (current default). Artists often forget to adjust Clamp value and do not know what value to use for realistic intensity. Also, currently both Clamp and Intensity do not have good UI ranges. This is why often Eevee renders end up very hazy and bloom often does not look right.
Bloom effect plays important role to help to distinguish between bright and relatively dim light sources. With current defaults this is broken because Clamp set to 1.0. Also, it cannot be disabled if set to 0 like expected. This patch fixes this and sets it to 0 by default. If users need to clamp, they can do so easily with UI range up to 1000. This range is good enough for most cases and provides enough precision to control lower values, and the highest value helps to limit bloom from the Sun if necessary and will leave untouched most other light emitters. If needed, much higher values for Clamp can be entered manually up to 100000. 10000 is still affects the Sun, but up to 100000 highest limit allows to clamp anything that is much brighter than the Sun if user needs to limit bloom in such cases (for example, bright explosion in the sky or anything else very bright).
I propose new default for bloom Intensity - 0.05 and UI range to suggests realistic values. Bloom Intensity > 0.1 is not realistic for clean lens but the user can enter manually much larger values if needed.
For comparison, here is a my own photo with and without bloom caused by the Sun (on second photo the Sun was occluded with an object).
{F6280500}
{F6280492}
In real life bloom is much more subtle and does not look hazy. If Clamp is disabled, then out of 0.1, 0.05 and 0.025 values I have tried, 0.05 looks most similar to the photo. Here is test render with and without bloom with the Sun in similar position like on the photo:
{F6280496}
{F6280494}
Using color probe 27x27 I compared lightness below the horizon under the Sun. In rendered by Eevee images lightness difference was 17. In case of the photos lightness difference in similar place was 11. I then compared leftmost spot (also below the horizon) and lightness difference was approximately 2 between two photos and 1 between rendered images. In other words, with these settings bloom effect is not too strong and is not too weak. Visually it may seem like decreasing bloom intensity may increase photorealism, but then bloom effect would be too localized even for the Sun.
Besides this single test, I tested in many other scenes as well, with and without the Sun, with different HDRIs, and as far as I can tell 0.05 intensity turned out to be good default - it produces bloom strong enough to be noticeable and not too hazy.
In Cycles shutter default value is 0.50, so for consistency set to 0.5 by default in Eevee too. Besides, 0.5 is typical standard for real cameras, and values higher than 0.5 usually are needed only if very strong motion blur is desired.
Here is summary of all changes:
Bloom Intensity: 0.8 > 0.05
Bloom Intensity UI range: 0-10 > 0-0.1
Bloom Clamp: 1.0 > 0.0 (disabled by default)
Bloom Clamp manual range: 0-1000 > 0-100000
Bloom Clamp UI range: 0-10 > 0-1000
Shutter: 1.0 > 0.5
This patch is related to the discussion in this thread, there are more examples of what bloom will look like with 0.05 intensity by me and others:
https://devtalk.blender.org/t/eevee-needs-to-have-physically-based-defaults/4700
Reviewers: fclem
Reviewed By: fclem
Subscribers: pablovazquez, billreynish, rboxman
Tags: #eevee
Differential Revision: https://developer.blender.org/D4212
This uses the same command as regular hierarchy delete, and is only
activated when debug value is set to 666 for now.
Here on file from T60419, it gives about 20% speed-up (from 5.5s to 4.4s).
There was no documentation at all, some very bad practices (like using
G.debug_value > 0 as some sort of global debug print switch), and even
an overlapping use of '1' value...
Also, python setter did not check for valid range (since this is a
short, not an int).
Main idea is to remove IDs to be deleted from Main, to avoid looping on
them to remove other deleted IDs usage (this is the most expensive
process in ID deletion, by far).
Speed improvements when deleting a large amount of IDs from a Main
containing a lot of them is quite significant, some examples for Objects:
* Removing 1k from 10k: 32% quicker (2.5s to 1.7s).
* Removing 10k from 20k: 60% quicker (59s to 23s).
* Removing 20k from 20k: 99.5% quicker (82s to 0.4s)!
Note however that this process is more risky/touchy, since we by-pass
some safety checks from regular ID removal here.
So will only give access to that code from python API for now (in
separate commit), so that it gets really tested. Also still need to
think about how to hook it up in UI (probably mostly for Outliner),
since we often do higher-level operations there...
Seems to be caused by cae3750 which changed free() function used
by bmain free to the one which does dependency graph tag. We do
no want to do any tags here.
This commit makes it so OpenSubdiv's topology refiner is kept
in memory and reused for until topology changes. There are the
following modifications which causes topology refiner to become
invalid:
- Change in a mesh topology (for example, vertices, edges, and
faces connectivity).
- Change in UV islands (adding new islands, merging them and
so on),
- Change in UV smoothing options.
- Change in creases.
- Change in Catmull-Clark / Simple subdivisions.
The following limitations are known:
- CPU evaluator is not yet cached.
- UV islands topology is not checked.
The UV limitation is currently a stopper for making this cache
enabled by default.
Add 'missing' bpy code from BKE_libblock_free_ex(), now both functions
do exactly the same thing, only the later is less flexible (fewer
'exotic' behaviors supported, like handling IDs outside of bmain etc.).
Next step: nuke usages of BKE_libblock_free functions, makes no sense to
have twice the same code around!
The existing Add and Multiply blending modes have limited usability,
because the appropriate operation for meaningfully combining values
depends on the channel. This adds a new mode that chooses the operation
automatically based on property settings:
- Axis+Angle channels are summed, effectively averaging the
axis, but adding up the angle. Default is forced to 0.
- Quaternion channels use quaternion multiplication:
result = prev * value ^ influence
- Scale-like multiplicative channels use multiplication:
result = prev * (value / default) ^ influence
- Other channels use addition:
result = prev + (value - default) * influence
Inclusion of default in the computation ensures that combining
keyframed default values of properties keeps the default state,
even if the default isn't 0 or 1.
Strips with this mode can be keyframed normally in Tweak mode,
except that for quaternion rotation keyframing always inserts
all 4 channels, and the channel value sliders on the left side
of Graph/Action editors won't insert keys without Auto Key.
Quaternion keys are also automatically normalized.
Differential Revision: https://developer.blender.org/D4190
Supporting a strip blending type that treats quaternions as a unit
also means being able to adjust all sub-channels as a unit when
inserting keyframes. This requires refactoring keyframe insertion
code to retrieve array property values for all channels at once,
before iterating over the indices being inserted.
Allows users to select a font for text strips in the video sequence editor.
Related: 3610f1fc43 Sequencer: refactor clipboard copy to no longer increase user count.
Reviewed by: Brecht
Differential Revision: https://developer.blender.org/D3621
When editing an action without a strip, or tweaking a strip without
time mapping enabled or supported, the extents of the virtual strip
can't be controlled and are purely derived from keys in the action.
Thus, cutting off evaluation of the action at these arbitrary points
gets in the way of observing the natural extrapolation of the F-Curves
and thus appears to be a mis-feature.
With this change non-mapped actions are evaluated with infinite
range, exactly like they are handled without NLA, unless extend
mode is set to Nothing.
The viewport stereoscopy support helpers are finally ported to 2.80.
We now can scale the camera and the "stereo cameras" will scale
in the viewport as well (unlike 2.7x).
At the moment I disabled the drawing of the camera frame when
stereo is selected and you are looking through the camera.
It is to be fixed later, but for now it draws the border wrong.
In 2.79 this was not a problem because the camera frame was drawn
afterwards as a hack.
Viewport > Stereoscopy:
* Cameras
* Convergence plane
* Convergence plane alpha
* Stereoscopy volume
* Stereoscopy volume alpha
