fix(desktop): preserve quality during sparse delivery jitter - #84
Merged
Merged
Conversation
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Small desktop updates with continuing delivery could repeatedly trip the timing-only receipt signal and drive the software encoder toward its 100 kbps floor. That degraded Full HD and increased codec skips even though reducing the size of a small update cannot fix retransmission or path jitter.
The sender now records newly delayed payload bytes. This soft signal requires at least 16 KiB in each of two adjacent pacing intervals; outstanding stalled receipts and hard failures retain their independent response. Private health and reduction logs expose payload totals alongside delayed-frame counts. No wire, identity, admission or timeout change.
Validation: the old quality fixture failed; the corrected feedback fixture preserves the encoder target across 240 successful small-update observations. All 45 session tests and 112 desktop unit/integration tests passed on macOS; strict workspace/all-target desktop clippy and fmt passed. Substantial sustained delay, hard failure, blocked-keyframe bounds and cancellation regressions remain covered. Fresh Linux/macOS CI and installed WAN acceptance are recorded separately in the qualification report.
The threshold is an application heuristic, not a capacity measurement or a completed native latency/stability gate.