feat(studio): render geometry-sized video filmstrips - #4234
Open
miguel-heygen wants to merge 3 commits into
Open
miguel-heygen wants to merge 3 commits into
miguel-heygen wants to merge 3 commits into
Conversation
miguel-heygen
force-pushed
the
hfoss34/timeline-filmstrip
branch
from
September 21, 2026 04:06
d4e484a to
ed17c41
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Video clips request a geometry-sized filmstrip after measurement and keep a one-frame poster tier available during fast scroll. Geometry requests use stable quantized cache identities and one 4K geometry ceiling.
Before
Origin/main, timeline region at Fit with video clips and the Voiceover audio row at rest:
Origin/main, mid-scroll timeline region with video clips and the Voiceover audio row:
Origin/main, the same timeline at 4x zoom and the same scroll position as After:
After
After the change, timeline region at Fit with video clips and the Voiceover audio row:
After the change, timeline region at 4x zoom with video clips and the Voiceover audio row:
The Footage Frame 1 and Footage Frame 2
no previewlabels are not an intended 4x behavior. The non-rich one-frame poster lease remains active at every zoom and is the fallback while rich work is unavailable; those rows are failing both poster and rich decoding in this capture and therefore remain an open preview defect, not an expected filmstrip state.Decode cost
Devbox decode-cost proxy on the 1080x1920 fixture, evenly spaced timestamps across the first 6 seconds, 3 runs: 6 frames averaged 1.67 s; 20 frames averaged 5.47 s, 3.26x. The 500px example needs 5 visible frames but quantizes the rich request to 8, trading a bounded decode step for stable cache identities. Every video clip now enqueues TWO jobs where main enqueued one, against
concurrentVideoDecodes: 2; the 1-frame poster is what paints during scroll, and the filmstrip follows once scrolling settles.Scheduler safety
120 mounted clips do not decode 20 frames each at once. The scheduler limits video work to the concurrency bucket at
packages/studio/src/player/lib/thumbnailScheduler.ts:345-348, pauses only rich work while scrolling atpackages/studio/src/player/lib/thumbnailScheduler.ts:207-210and:258-261, and retains results only within byte and entry budgets atpackages/studio/src/player/lib/thumbnailScheduler.ts:327-334and:384-392, configured bypackages/studio/src/player/lib/timelineViewportBudgets.ts:70-82. The one-frame poster request is non-rich, so it remains eligible during scrolling. Quantised request counts and the cache byte budget keep a 120-clip timeline from holding 120 x 20 decoded frames.Validation
Math.ceil(3840 / (66 * (16 / 9))), named once with its geometry reasoning beside the owner budget.