You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Joined copy from an extension keeps the render audio without retiming it #39
spool/joins.py:332. With a 25 fps render in a 24 fps sequence, the plain take's audio is retimed (5.201 s) but the joined copy copies the audio stream (5.0 s). It drifts by up to 0.2 s and ignores the measured shift.
Decide whether the joined copy should retime its audio the same way as the plain take.
Source: PR #35, report section "N-06" (report). Line numbers refer to main at 7caf5c9, before the PR.
Implementation status (2026-11-07)
PR #35 trims/delays joined-file audio for the measured shift, retimes it by sequence/native frame rate and pads/trims to video duration. Real FFmpeg tests verify beep alignment for shifts -1/0/+1, duration and the no-audio case. Studio playback/export still reads the plain take; no UI audio-routing change is claimed. Review/merge remains open. Current branch head: cf37734368c3de9a6d576a7222f26aa345adc107.
🤖 GPT-6.1 Sol responding on behalf of doubletwisted
Joined-file audio retiming is in PR #35. Real FFmpeg fixtures verify positive/zero/negative measured shifts, sequence/native-rate conversion, exact joined duration and no-audio inputs; six join tests pass. Studio linked-audio playback/export still deliberately reads the plain take, so this does not claim that broader routing change.
Activity
doubletwisted commentedon Nov 7, 2026
Note
🤖 GPT-6.1 Sol responding on behalf of doubletwisted
Joined-file audio retiming is in PR #35. Real FFmpeg fixtures verify positive/zero/negative measured shifts, sequence/native-rate conversion, exact joined duration and no-audio inputs; six join tests pass. Studio linked-audio playback/export still deliberately reads the plain take, so this does not claim that broader routing change.