Faster and better looking VSE scopes & "show overexposed". Waveform &
RGB Parade now can also show HDR color intensities. (Note: this is
only about VSE scopes; Image Space scopes are to be improved separately)
- Waveform, RGB Parade, Vectorscope scopes are done on the GPU now, by
drawing points for each input pixel, and placing them according to
scope logic. The point drawing is implemented in a compute shader,
with a fragment shader resolve pass; this is because drawing lots of
points in the same location is very slow on some GPUs (e.g. Apple).
The compute shader rasterizer is several times faster on regular
desktop GPU as well.
- If a non-default color management is needed (e.g. VSE colorspace is
not the same as display colorspace, or a custom look transform is used
etc. etc.), then transform the VSE preview texture into display space
RGBA 16F texture using OCIO GPU machinery, and calculate scopes
from that.
- The "show overexposed" (zebra) preview option is also done on the
GPU now.
- Waveform/Parade scopes unlock zoom X/Y aspect for viewing HDR scope,
similar to how it was done for HDR histograms recently.
- Added SEQ_preview_cache.hh that holds GPU textures of VSE preview,
this is so that when you have a preview and several scopes, each of
them does not have to create/upload their own GPU texture (that would
both waste memory, and be slow).
Screenshots and performance details in the PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/144867
Most of these crashes happen because it is assumed that the scene will
always be present even if we have an uninitialized `Editing`, which is
no longer the case with #140271.
- Fix crash when clicking and dragging in the scrub area by checking for
valid `sequencer_scene` in `change_frame_poll`
- Fix crashes when selecting menu items by disabling them in the UI
until a `sequencer_scene` is present
- Fix crashes running operators from the F3 menu by changing to more
restrictive polls that check for `sequencer_scene`
- For good measure, check before dereferencing in
`channels_displayed_get`, `active_seqbase_get`, and `editing_get`
Pull Request: https://projects.blender.org/blender/blender/pulls/145145
Implements the proposed design (with some modifications) in #135058.
## Sequencer Scene
This adds a new property called `sequencer_scene` to workspaces. This scene is used
by the video sequence editors in the current workspace for their context.
This is a first step towards "detaching" the VSE from the active scene in the window.
Each sequencer timeline editor shows the sequencer scene that is being used.
By default, when no sequencer scene is selected, the timeline and preview are empty.
Pressing the "new" button will add a new scene and assign it to the sequencer
scene for the current workspace.
## Contextual Playback
Pressing `Space` (by default) for starting the animation playback is now contextual:
depending on the context (where your mouse cursor is), the scene that is played back
might be different. E.g. with a 3D Viewport and a Sequencer open, pressing "play"
in the 3D Viewport will play the _active scene_ of the window, while pressing "play"
in the sequencer will play the _sequencer scene_.
## Time & Scene Synchronization
Additionally, this adds a toggle called "Sync Active Scene".
With the property turned on, the active scene & scene time in the window will be
synced with the time & scene of the current scene strip in the sequencer.
Note that this is _not_ bi-directional. The sequencer can change the active scene
and map time, but it's not possible the other way around since it one can have
multiple strips using the same scene (+camera, and even time!).
Currently this setting is exposed in the footer of the sequencer timeline as well
as in the workspace settings.
This allows for one of the core concepts that the story tools projects aims at: Working
in a scene (e.g. in the 3D viewport) while also working with the edit
(in the sequencer timeline).
## Some technical notes
* Undoing while playback is running will now cancel playback. This is to avoid the timer,
that points to the scene and viewlayer that are playing, to get de-synced after loading
the memfile undo step.
* When the sequencer scene is not the same as the active scene, we ensure it has
a depsgraph.
* Normally, when a `NC_SCENE` notifier points to a specific scene, the notifier is dropped
if that scene doesn't match the active one in the window. We now also check that it
doesn't match the sequencer scene in the active workspace.
* When loading older files, we need to make sure that the active workspace in a window
uses the active scene as the sequencer scene. This is to make sure that the file opens with
the same sequences open.
* Tool settings are stored per scene. To make sure the sequencer uses the tool settings for
the sequencer scene, the "context.tool_settings" and `CTX_data_tool_settings` members
are overridden in the sequence editors.
Pull Request: https://projects.blender.org/blender/blender/pulls/140271
Caused by casting float value to int in source cache eviction logic. The cache
key was changed from int to float for timeline frame, but this was not changed
in `source_image_cache_evict()`. Compiler did not catch this unfortunately.
Pull Request: https://projects.blender.org/blender/blender/pulls/145019
* PROP_COLOR_GAMMA is sRGB, not display space
* Hex colors are always sRGB
* Image byte buffers are in byte_buffer.colorspace
Fixes for sequencer text, image painting, render stamp and tooltips.
The default display space is sRGB, so this change will not be noticed
in most files.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
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
Before, trying to add multiple strips would always create an image
sequence. This patch allows images to be added in bulk, just like movie
& sound strips.
A new "Image Sequence Detection" property has been added which works
like the image editor. Toggling it off will always import individual
images and never create image sequences.
If it's on, you can also optionally "Use Placeholders" to fill in gaps,
which exhibits the same behavior as before (placeholders will show as
pink in the preview until you have the files on your system in the right
folders).
You can even add both image sequences and regular images at the same
time. Only caveat is that this doesn't work if the regular images are
interspersed within sequence frames in the file browser view when you
add them. I.e., all frames of a sequence should be listed without any
regular frames in between (but the sequence frame numbers don't have to
be in order). This can usually be ensured on all sort types (like date,
name).
NOTE: This patch also replaces the "end_frame" property on image /
effect strips with "length," to support behavior when multiple strips
are involved. This is a breaking change for the API, but is a fairly
small change and will be included in the release notes.
The "length" property allows one to adjust all image strips' lengths at
once. It does nothing for image sequences, however, since their lengths
are defined by their image count.
Pull Request: https://projects.blender.org/blender/blender/pulls/143974
- Deduplicate `scale_fit_methods` enum
- Reorder functions, polls, and code to locations that make sense
- Rename some functions to remove `seq` terminology in favor of `strip`
- Renames `len` in `LoadData` to `count` for images
- Add clarifying comments and improve descriptions
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
Not a common scenario, but if you had very few source images (e.g.
long image strips with one frame) and many final frames cached,
the full cache eviction policy was not handling that well, since it
was always removing at least one source image for each final frame.
Resulting in effectively the cache only being full of final frames.
Change that so:
1) evict_caches_if_full has a loop that tries to evict until the caches
are not full (i.e. not necessarily just one image), and
2) only evict source image per final frame, if there are more source
images cached.
Pull Request: https://projects.blender.org/blender/blender/pulls/144612
`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
Duplicating a strip that references an ID like the scene strip would not
duplicate the scene. This is wanted in some workflows.
To align with the rest of Blender, this changes the behavior for how
strips are duplicated:
* `Shift + D` ("duplicate"): Duplicate the strip and also duplicate the
IDs referenced by the strip. Currently this only affects `Scene`,
`MovieClip`, and `Mask` strips.
* `Alt + D` ("duplicate linked"): Duplicate the strip, but reference the
same IDs. This is the current behavior in `main`.
Part of #144063.
Pull Request: https://projects.blender.org/blender/blender/pulls/144138
Broken with new 4.5 cache implementation; the loop for building
image sequence proxies was not flushing the intra-frame cache
between frame changes. This affects image proxies since they go
through the whole VSE render stack; does not affect video proxies
since they just manually decode the file frames.
Pull Request: https://projects.blender.org/blender/blender/pulls/144269
This removes the `STRIP_DUPE_UNIQUE_NAME` and `STRIP_DUPE_ALL` macros
and replaces them with an `enum StripDuplicate`.
A flag with a value of 0 was implicitly used as "duplicate selected strips" so the
enum uses `Selected = 0` to document this behavior better.
Pull Request: https://projects.blender.org/blender/blender/pulls/144125
VSE cache processing was handling thumbnail generation requests in
parallel, however often that was resulting in CPU core over-subscription
that negatively affects user interface smoothness. Process requests
serially; some parts of image/video loading are internally multi-threaded
already. This makes thumbnails load somewhat slower, but while that
is happening the rest of blender feels much smoother, and the CPU
fans don't spin up as much.
Additionally, make the video thumbnail loading code use smallest
available proxy file, if present. This was overlooked in the original
new thumb cache implementation (!126405).
SpriteFright timeline edit v135 (mostly JPG images, some movies),
loading thumbnails for the whole timeline on Ryzen 5950X (Windows):
- 4.2.1: 35 seconds, CPU usage ~35%, UI smooth
- Main: 13 seconds, CPU usage ~90%, UI choppy
- This PR: 19 seconds, CPU usage ~35%, UI smooth
Manually made timeline with 10 video files:
- 4.2.1: 12 seconds, CPU usage ~50%, UI smooth
- Main: 6 seconds, CPU usage ~100%, UI choppy
- This PR: 9 seconds (no proxies) / 4 seconds (with proxies), CPU usage
~50%, UI smooth
Candidate for 4.5.x backport
Pull Request: https://projects.blender.org/blender/blender/pulls/144120
Replaces the `FPS` macro with `scene->frames_per_second()`.
The macro has two major issues:
* It hides that it depends on a `Scene *` variable named `scene`.
* It makes debugging harder.
This is now replaced with a member function on the scene.
Pull Request: https://projects.blender.org/blender/blender/pulls/144127
- BLF font flags were an untyped enum; change to actual enum and
update usages from int to that.
- Move all BLF API enums to their own header, so that if something
needs just the enums they don't have to include whole BLF_api.hh
Pull Request: https://projects.blender.org/blender/blender/pulls/143692
Use the form "Processing something..." (gerund, sentence case,
ellipsis) for running jobs progress label, according to Blender's
Human Interface Guidelines.
Pull Request: https://projects.blender.org/blender/blender/pulls/141573
Gaussian blur effect size is expressed in pixels (not ideal, but that is
what it is). It was not taking render output scale or preview proxy
size into account, resulting in much larger blur at lower render scales.
Pull Request: https://projects.blender.org/blender/blender/pulls/143627
- 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
Use `BLI_strncpy_utf8` & `BLI_snprintf_utf8` for fixed size buffers in
DNA and screen data structures such as panels, menus & operators.
This could be considered a fix as copying a UTF8 string into a smaller
buffer without proper truncation can create an invalid UTF8 sequence.
However identifying which of these users are likely to run into would
be time consuming and not especially useful.
VSE final image cache invalidation for the newly added effect
strip was invoked before the strip time range values were set,
effectively not invalidating anything.
Pull Request: https://projects.blender.org/blender/blender/pulls/142854
Quite a few flags in `Strip.flag` were used only at runtime; they were
temporarily set and then cleared thereafter. Move them to
`StripRuntime.flag` instead.
Most of the moves are trivial renames, with some old versioning code for
`SEQ_FLAG_DELETE` deleted. That code fixed a very uncommon edge-case bug
in 2.x which doesn't seem worthwhile to keep deprecated code around for.
Pull Request: https://projects.blender.org/blender/blender/pulls/142776
Previously, newly added Gaussian Blur effect did not blurring at all,
since blur sizes were set to zero. Make them default to 9 pixel size
blur, which is similar to the amount of blur the Glow effect does by
default.
Pull Request: https://projects.blender.org/blender/blender/pulls/142851
Ultimately, the issue was caused by updating sound sequence handle in
`AUD_SequenceEntry_setSound` on each new frame regardless if the
modifier data has changed.
This commit fixes the issue by implementing a means for modifiers to
check, if their parameters or inputs are changed.
This is done by storing these parameters in `StripModifierDataRuntime`
struct, that is shared between all modifier types. This is not ideal,
but it significantly simplifies dependency graph runtime store/restore
code.
Function `strip_update_sound_modifiers` passes boolean `needs_update`
to strip stack update functions. If any needs to be updated, it sets
value of `needs_update` to true allowing following update functions to
skip parameter checking to speed up the process.
Original code updated sound sequence handle twice. Once by function
`BKE_sound_update_scene_sound` then by `strip_update_sound_modifiers`.
If sound modifier is used, only `strip_update_sound_modifiers` needs to
be called, so there is condition to decide which one of these functions
is called.
Also fixes#139605
Pull Request: https://projects.blender.org/blender/blender/pulls/141595
This PR switches out the internal scene context function
used from the sequencer. This is done in preparation for
#140271.
Using a separate context function allows us to keep the
existing `context.scene` to refer to the active scene
and use a separate context member to refer to the
sequencer scene in the workspace (which is added in #140271).
Previous attempts simply overrode the `context.scene`
to refer to a different scene than the active one in the window.
This has two issues:
1) Any operator that wants to use the active scene in the window
can't do it in the context of the sequencer. This is a problem for
example for poll functions of operators that don't have anything to
do with the sequencer.
2) For better or for worse, Blender expects the `context.scene` to
always exist. For the sequencer, we'd like to possibly have no
sequence selected.
Using a different context member for the sequencer has some
advantages:
1) Although we have to change quite a few places, it's limited to the
sequencer. We don't have to change other parts of Blender to make
things work.
2) It allows us to prepare for the future when we might want to
separate the VSE from the scene. This is a step in that direction.
Having a different context function makes it easy to find the places
that would need to be refactored.
Pull Request: https://projects.blender.org/blender/blender/pulls/141271
This initial commit properly clamps handles for video/audio strips, and
provides functionality to enable/disable the behavior for all strip types
(addresses #90280).
Toggling handle clamping is done with "C",
just like with the redesigned slip operator (#137072).
If a strip is not already clamped when you start moving its handles,
then clamping behavior is disabled starting out. This means no abrupt
clamp until you explicitly ask for it.
Transform logic was altered, fixing a few bugs:
- When initializing a transform, `createTransSeqData` would already
create some clamping data for channels. This patch replaces it with
`offset_clamp` (for unconditional clamping which cannot be disabled)
and `handle_xmin/xmax` (for hold offset clamping, which is optional).
- Collecting this data ahead of time is necessary for the double
handle tweak case -- `flushTransSeq` only works one strip at a
time, so we can't clamp post-hoc.
- In `applySeqSlideValue`, we apply `transform_convert_sequencer_clamp`
before values are printed to the header, but let the unclamped values
get flushed to the strips themselves. This is so that we can have the
data later at the individual strip level to recalculate clamps.
Otherwise, if transform values are clamped preemptively, then we have
no idea whether strips are clamped vs. merely resting at their
boundaries.
Note that currently, handle clamping is drawn identically to overlaps.
More information in PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/134319
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
