Currently, the scene strip preview uses the existing dependency
graph built for that scene. This was not a big issue before
the sequencer scene was introduced, because the user would have
to create two main Blender windows to run into problems.
Now with the sequencer scene, it's possible to look at the scene
of a scene strip within the same window.
When the user is editing the scene (e.g. moving an animated object)
any open sequencer preview will cause the edits to be flushed.
This can e.g. result in visual jumping of animated objects, and more.
This PR attempts to fix the issue in a straightforward way: Use
a separate dependency graph for rendering the sequencer preview.
While this fixes the immediate issues, there are some consequences:
* The memory usage of the scene dependency graph _can_ roughly
double (since there are now likely two instances of the same
dependency graph). Because of implicit sharing, unmodified data
will not be copied. But for example modifiers on meshes would
currently create two copies of the evaluated data in the two
dependency graphs.
* Creating the dependency graph can be costly, which will cause the
first frame that the scene has to render to be slower.
Note: The current code changes some properties of the original scene
like the frame, subframe etc. before rendering and then restores
the original state. In theory, this part of the code can be removed,
but may be a bit too risky for just a fix. This should be improved
at a later stage.
Also resolves#146769, #139501.
Pull Request: https://projects.blender.org/blender/blender/pulls/147457
The final image produced by the compositor can have domain translation
on it (e.g. caused by a Translate or Transform node). Similar to how
the regular compositor viewer node remembers the output domain
translation, do the same in the compositor modifier.
Bubble back that translation up to VSE rendering code, where it is
then added to regular strip transform.
In order to make this "bubble up" part easier, refactored modifiers
so that instead of soup of parameters they all get a struct
ModifierApplyContext with all the relevant data.
Added a new VSE render test that covers various compositor
transformation nodes (translate, rotate, transform, corner pin).
Pull Request: https://projects.blender.org/blender/blender/pulls/147695
Currently when a strip has a transform that does not fill the whole
render area, first the image of the strip is transformed, and then
any modifiers are applied on that. This is mostly in the new
Compositor modifier, where procedural textures, gradients, image
coordinates "stick to the screen" instead of following the transformed
strip.
This changes the behavior so that first the modifiers are applied
to the strip image, and then the strip is transformed. This is
potentially a visually breaking change:
- This can alter visual look of existing strip, especially if they are
scaled. Previous behavior was first scale filtering, then modifier;
now it is first modifier, then scale filtering.
- Most obvious change is Compositor modifier (which is new in 5.0).
- Compositor modifier can actually expand the input image (e.g. Blur
node with "expand bounds" option set), and that works.
- Note that Masks continue to be applied in global/screen space. There
can be small look differences with rotated/scaled strips that use
masks, due to Mask application now needing to do filtered mask image
lookups.
- If anyone needs previous behavior (modifier is applied on the
"whole screen"), they can put transformed strip into a meta strip,
and apply the modifier on the meta strip itself.
Compositor modifier examples with images in the PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/146181
If a VSE compositor modifier is used to do keying or introduce an alpha
somehow to the image, the alpha is not used when blending. This is
because the VSE evaluator assumes all modifiers except Mask could not
introduce an alpha channel. So we need to extend the check to also make
an exception for the compositor modifier.
Pull Request: https://projects.blender.org/blender/blender/pulls/147456
When a strip modifier uses an adjustment layer that is above it
as the mask input, this leads to recursive rendering. Similar to the
fix in !146624, pass SeqRenderState to modifiers as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/147029
When effect of adjustment layer strip is moved below the adjustment
layer, this causes infinite loop in strip rendering. Same happens when
you use multicam strip and set source channel to one of its effects.
This is fixed by passing `SeqRenderState` to the effects. If any strip
renders "seqbase" pointer of strip is stored in set in the render state
struct. If the pointer exists in this set, function returns without
rendering anything. In other words, The strip must never render itself.
Pull Request: https://projects.blender.org/blender/blender/pulls/146624
- "Is this strip type an effect?" has been done by reserving 3rd bit
of the effect enum values to indicate that. That is a very strange
decision, so make that be done by strip_is_effect() function. The
enum values have to stay the same for backwards compat however.
- Both "strip type" and "strip blend mode" were sourced from the
same STRIP_TYPE_ enum, with only parts of the values meant for "type",
and other parts of values meant for "blend mode". That again is highly
confusing; split that off into two enums. Yes there's a handful of
values in them that overlap, but not the majority.
Pull Request: https://projects.blender.org/blender/blender/pulls/145746
This PR updates the VSE strip modifiers interface.
It now uses the same design as the object modifiers.
Changes:
* Except for the "Mask Input" subpanel, the modifier UIs are unchanged.
* Modifiers can now be rearranged using drag & drop.
* Additionally, there is now an active strip modifier. This is exposed
though python via `strip.modifiers.active`.
This is in part for !139634 which needs the concept of an active modifier.
Notes:
* The `modifier.cc` file included all the implementation of all modifiers.
With the addition of a another new callback in this PR, this file was
getting quite big so I split everything out into individual files for all
modifiers. The modifiers are getting registered at launch.
* The modifier panels are getting added using a UI template
(`template_strip_modifiers`) very similar to the object modifiers.
Pull Request: https://projects.blender.org/blender/blender/pulls/145367
Currently, sequencer structs contain several pointers to data within
other structs. These pointers need to be remapped as the structs are
reallocated when reading from blend files. That has worked so far
because the pointers are exactly the values from the Blender session
that saved the file. With the implementation of #127706, the pointers
in the file aren't "real" anymore, and we can't offset them to get the
struct that contained the data. I'm working on the pointer stability
solution now to address a memfile undo performance regression
in 5.0 due to the `AttributeStorage` read/write design.
This commit replaces these 4 mid-struct pointers to point to the
containing strips instead, and uses some trivial logic to access the
fallback root sequence channels and strips. This makes the pointer
remapping on file load possible again.
This change is backward compatible but not forward compatible.
Second try after #144626
Pull Request: https://projects.blender.org/blender/blender/pulls/144878
Currently, sequencer structs contain several pointers to data within
other structs. These pointers need to be remapped as the structs are
reallocated when reading from blend files. That has worked so far
because the pointers are exactly the values from the Blender session
that saved the file. WIth the implementation of #127706, the pointers
in the file aren't "real" anymore, and we can't offset them to get the
struct that contained the data.
This commit replaces these 4 mid-struct pointers to point to the
containing strips instead, and uses some trivial logic to access the
fallback root sequence channels and strips. This makes the pointer
remapping on file load possible again.
The downside is that this isn't strictly backward or forward compatible,
but only on a UI-level. The active meta-strip information will be lost,
and the sequencer will reset to displaying the root sequence.
Depends on #144624
Pull Request: https://projects.blender.org/blender/blender/pulls/144626
With the aim of removing `seqbasep` to remove the complicated logic for
repairing pointers within the `Strip` struct when loading files and undo
steps, this commit just moves access of the variable behind a function.
In the future the function will retrieve the list from a Strip pointer,
for now it just returns the existing pointer.
Overall motivation is that blend file pointer manipulation is incompatible
with the changes required for #127706.
Pull Request: https://projects.blender.org/blender/blender/pulls/144624
`render_give_ibuf_direct()` just called `seq_render_strip()`, which
tried to get image from intra frame cache. This always succeeded,
because the intra frame cache does not accept timeline frame argument.
Using intra frame cache here is incorrect. Source cache must be used.
If source cache lookup fails, intra frame cache must be invalidated
before calling `seq_render_strip()`.
Pull Request: https://projects.blender.org/blender/blender/pulls/144396
- Some functions that took integer arguments for proxy size enums; make
them take said enums directly.
- Add get_render_scale_factor() that calculates effective render scale,
use that in places that did manual "use either scene render scale setting
or proxy size" calculation in 5 places.
- Replace previous double with float in proxy size scale factors; all
the factors are exactly representable as floats, and all the calling
places used them as floats too.
- In various VSE related DNA structs, indicate which enums they use
as comments,
- Add enum types for the enums that were untyped,
- Move Wipe effect enums to DNA header for clarity, since they are
serialized values,
- Reduce SeqRetimingKey size from 40 bytes to 32 bytes; there
was too much unnecessary padding.
- Rename some enums SEQUENCE_ -> STRIP_ where relevant.
- Rename wipe enums DO_MEOW_WIPE -> SEQ_WIPE_MEOW
Pull Request: https://projects.blender.org/blender/blender/pulls/143565
As mentioned in comments on #43646, a negative channel value in preview
view settings can be used to climb the metastack. But the functionality
did not work as expected in most cases.
This combines a one-line fix by Sergey with an updated description for the
property which documents the feature.
In the future we may want to change its operation to be less obscure,
since it has some use-cases in aligning strips within a metastrip with
those outside of it.
Regression caused by 9e4c26574a.
Add strip stack as a key to the final cache.
Use the pointer to the list of sequences, as it is the easiest one to
obtain in all places where it is needed. This is slightly different
from the code prior to the 9e4c26574a where strips.last() was used,
but it allows to use the same logic in the prefetch job.
Pull Request: https://projects.blender.org/blender/blender/pulls/141778
Blender expects float buffers to be in scene linear space, which was
violated bu the 1012bit movie reading code. While such image buffers
can be displayed correctly, performing operations in various areas
of Blender might lead to unexpected results.
The non-linear colorspace for ImBuf is expected to be "internal-only"
to a specific area, like VSE.
This is only done for Image and MovieClip data-blocks, sequencer still
reads movie files in their original colorspace as it helps performance and
sequencer can not be referenced from places where linear colorspace
for float buffer is really important.
Pull Request: https://projects.blender.org/blender/blender/pulls/141603
Make display channel part of the key for the final cache.
The prefetch uses display channel of 0, which is the default. It might
need to be adjusted, but it is unclear whether it was behaving different
prior to the 9e4c26574a.
Pull Request: https://projects.blender.org/blender/blender/pulls/141670
If render settings resolution scale had lowered resolution, cached
images from render image/animation session could "stay around"
and be incorrectly used in the VSE preview area. Two cases I found are
fixed here:
- Intra-frame cache was not flushed upon actual final resolution
change,
- "Source images" for effect/scene strips were not removed when
requested resolution no longer matches their rendered resolution.
Pull Request: https://projects.blender.org/blender/blender/pulls/141297
Text strip along with other effect strips draws into buffer, which is
defined by common render size. However this makes selection, snapping
and other features not work correctly from user perspective.
There are 2 ways to remedy this:
- Draw text into buffer which size is defined by text boundbox
- Just draw the correct size boundbox and make UI use it correctly
This PR implements second option, as there are multiple difficulties
asociated with first option. Main problem is, that one would need to
calculate boundboxes for all individual text effects(blur, shadow, ...),
but many of these effects need to know the size of buffer we are trying
to calculate.
Big question here was: What the boundox should look like?
For selection it is best, if it encapsulates the text itself along with its
effects like shadow, blur or box. However for snapping and pivot range
it seems, that encapsulating only the text without any effects applied is
the best solution. This was discussed with UI module and we agreed,
that encapsulating only the text would provide better usability.
The implementation:
Most of the features are on "preview side", where all dimensions are
related to scene render size. Temporary `TextVarsRuntime` is created to
calculate apparent text strip image size. This allows boundbox of
correct size to be drawn.
However to draw boundbox and do transformations correctly, it is
necessary to remap origin point to the subset of image defined by the
boundbox. Since recalculated origin is relative, this function is used
for image transformation during rendering as well.
Main drawback of this method is, that boundbox and origin point
calculation may be needed from multiple places, and `TextVarsRuntime`
may be recalculated about 10x on redraw. `text_effect_calc_runtime`
function can be executed from ~5us for single character to 0.5ms with
512 characters with default font(not sure if that matters).
Drawing origin only during transformation would help with general UI
responsiveness. During playback, these overlays are not drawn, so
there should be no change in performance.
Pull Request: https://projects.blender.org/blender/blender/pulls/140420
`seq_need_scale_to_render_size` returned true, when scaling was not
needed. The code that used this function worked correctly, so the logic
was inverted in both places.
Before this change prefetching would not work correctly with only RAW
caches. The RAW caches wouldn't previously be emptied either if
switching them off while prefetching was turned on.
This is part of the short term roadmap goal of simplifying the
compositor workflow
(see https://projects.blender.org/blender/blender/issues/134214).
The problem is that many users don't know how to get started with
compositing in Blender, even when they have used Blender for other
areas, e.g. modeling.
Note: although the solution makes compositor node trees reusable
accross blend files, this is a nice side effect and not the main goal
of the PR.
This PR implements a "New" button that creates a new compositing node
tree, and manages trees as IDs. This has following advantages:
- Consistent with other node editors and other parts of Blender,
therefore making it easier to getting started with compositing if users
are familiar with shading or geometry nodes
- Give users the ability to reuse the compositing node tree by linking
or appending it.
Note: The parameter "Use Nodes" is still present in this PR, but will
be removed (in a backward compatible way) in a follow up PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/135223
- Previously, effects could be changed into types that take different
amount of inputs, e.g. from "add" to "glow", which led to undefined
and glitchy behavior (unreported bug). Fix by requiring same input count
for new effect.
- Previously, strip names were not updated when changing effect type,
which tended to be confusing. Now, if the default name was not changed,
update to the new effect type name.
- Remove "effect strip" menu from color and text strips. Although they
are effects internally, to a user, this is not clear, and providing the
option to modify effect inputs in these cases doesn't really make sense.
Pull Request: https://projects.blender.org/blender/blender/pulls/139514
Previously there was no mechanism to invalidate scene strip cache, when
changes were made to scene that the strip renders. Because of this, the
cache had to be cleared manually, even for other strips.
Strip invalidation is performed from function
`ED_render_id_flush_update`.
Recalc flags like selection or frame change are ignored as they are
unlikely to influence the rendered image.
When some ID is changed, all scene strips using
`DEGEditorUpdateContext::scene` will be invalidated.
Limitations:
- Because frame change is ignored, changes to physics caches would not
invalidate VSE strips
- Only active scene will be invalidated this way. So if an object is
linked to 2 scenes, inactive scene cache would be invalid
- If animated object is moved, and there is VSE preview to be updated
Scene re-rendering in VSE would cause object to snap to its evaluated
position. This is because scene frame is changed by VSE. Reported in
#139501.
To refresh VSE preview after cache is invalidated, relatively big chunk
of code was copied from `view3d_main_region_listener` to VSE preview
listener. Cases that would not result in visual changes like selection
were excluded.
Fixes#113137.
Pull Request: https://projects.blender.org/blender/blender/pulls/138231
This should make VSE code more readable and easier to understand from an
outside perspective.
The name was chosen to be `channel` rather than `channel_index` to keep
things short and concise -- it should be clear based on the context
whether we are talking about the strip's channel index (singular case,
`Strip::channel` or `SeqTimelineChannel::index`) vs. the channel list
(plural case, e.g. `Editing::channels`).
Pull Request: https://projects.blender.org/blender/blender/pulls/138919
This patch renames `seq1` and `seq2` to `input1` and `input2` in `Strip`
and `StripSelection` structs to consist with recent refactors in
#132179. It also renames other instances of `seq1` and `seq2` (e.g.
local variables) to `strip1` and `strip2` instead.
There is only one small breaking change to the Python API with
`strips.new_effect()` taking in the new names now. There should be
no other functional changes involved.
Pull Request: https://projects.blender.org/blender/blender/pulls/138296
Rework internals of how VSE caching is done. Primarily to make all the
caching logic more understandable from development point of view, but
also has several user visible implications (more details in the PR):
- Simpler and fewer caching UI options,
- Disk cache is gone (primary reason: proxies are kinda the same thing),
- VSE cache size set in preferences is actual size used for VSE caches
now (previously caching stopped as soon as whole Blender used that
much memory, even if some memory usage was not about VSE at all),
- Certain scenarios of cache invalidation are faster now.
Pull Request: https://projects.blender.org/blender/blender/pulls/137926
* Remove `DEG_get_evaluated_object` in favor of `DEG_get_evaluated`.
* Remove `DEG_is_original_object` in favor of `DEG_is_original`.
* Remove `DEG_is_evaluated_object` in favor of `DEG_is_evaluated`.
Pull Request: https://projects.blender.org/blender/blender/pulls/138317
Ref: #132179
Renames:
- `Editing.act_seq` -> `Editing.act_strip`
- `SequenceModifierData` -> `StripModifierData`
- Its member `mask_sequence` is now `mask_strip`.
- `MetaStack.parseq` -> `MetaStack.parent_strip`
- Remaining function names/parameters that were not dealt with in #132748
- Various references to `seq` or `sequence` throughout code and docs when
referring to a strip
Also moves `_get` to the end of the renamed function names where
applicable for standardization (unless "by" or "from" are used).
There should be no changes to current behavior.
Pull Request: https://projects.blender.org/blender/blender/pulls/138077
