Fixes: https://trello.com/c/Bdl2DSnt
Solution:
Binds playback speed in editor to deltatime instead of assuming that the
FPS is 60.
Optimised rendering by allowing it to go above 60fps since we are not
FPS bound anymore.
Removed smaaPass because it was making the editor blury and not as crisp
as it could be.
Tests:
Checked creating movie - p.s Also is faster now with no FPS cap
Checked playback speed with stop watch.
Checked different scenes.
Checked Transform & Animation engine.
- waveform now has a loading state
- audio library now has loading new items state
- fix so that the "New" state works on both inference created audio
items and uploaded audio items
- fix so that uploaded audio items takes the file name as its title
Co-authored-by: wil <wil@storyteller.ai>
- Finished UI and functionality for Pagination
- Pagination implemented in audio library
- Finished UI for uploading audio
- Connected the UI to the upload media endpoint
- Audio Library auto-refreshes upon finishing upload
- Fixed up audio upload bugs that already existed in V2V workflow
Co-authored-by: wil <wil@storyteller.ai>
Fixed timeline update with generating of preview as well as fixed memory
leak.
Call ffmpeg.exit() to fix memory leak with generation.
if the editor is in preview mode it gens frames.
Tests:
tested multi spam click on previews.
tested editor timeline.
tested normal click on preview.
No crashes and task manager does not show memory leak.
Fixed resolution.
The bokehPass post processing was messed up with the FOV and was not
needed so was removed.
The resolution now updates with the UI aswell.
Updated camera layer on line 499 to fix P1 Loading in a scene bugs out
the camera view in trello.
Fixed loading because load_glb does not exist anymore and was replaced
with loadGlbWithPlaceholderso that is now updated
fixed points creation when loading and added rotation and scale.
also updated the parimitive loading because it returns an object and we
need a UUID
guaranteed audioItems to have length either from the server, or retrieve
it from the loading of the file,
the unit of mediaItem's length is FPS, adjusted the math for supplying
length in frames
Co-authored-by: wil <wil@storyteller.ai>
Updates the points position on update didn't need to update the object
because `this.transform_engine.clips[object_uuid].setTransform` does it
already.
1. use the new tab title component
2. fix all memory persistency problems,
3. furthuer minimized state change calls
4. "new*" tag on newly generated media
Co-authored-by: wil <wil@storyteller.ai>
simplified the ui code
waveform on upload module works
uploading file works!!!!
clean/optimized code for next step
v2v endpoint called and successfully processed
useSignals is essential
audio type pill done
Co-authored-by: wil <wil@storyteller.ai>
Quicksticks-oss
Miles Ryan
Victor OandVictor
wilandwil
ArEnScandQuicksticks-oss
Brandon Thomas