SUMMARY adding the "Freeze" effect to freeze a frame in combination with adding a subtitle causes the video to become greyish - loss of contrast, for the duration of the freeze frame. STEPS TO REPRODUCE 1. Select a short clip and apply the "Freeze" effect. Works as it should. 2. Add a subtitle track, the project monitor window looses contrast. Performing the subtitle before adding the "freeze" effect also has the same issue. The loss of contrast also appears in the rendered output. 3. I even tried adding the "freeze" effect without the subtitle, then adding the edited sequence to a new sequence with just the subtitle, but the problem still exists. OBSERVED RESULT When the "freeze" effect is used in combination with a subtitle track the video for the duration of the freeze has lower contrast. EXPECTED RESULT Adding a subtitle track should not change the contrast of output. SOFTWARE/OS VERSIONS Linux: Ubuntu 24.04 LTS Windows: not tried macOS: not tried (available in the Info Center app, or by running `kinfo` in a terminal window) Linux/KDE Plasma: Kdenlive 24.08.03 appimage KDE Plasma Version: ??? KDE Frameworks Version: 6.7.0 Qt Version: 6.8.0 MLT 7.28.0 ADDITIONAL INFORMATION Workaround found. use (right click - Extract frame to project) from Project Monitor window and then add the extracted frame as a clip instead of using the "freeze" effect. Add subtitle as required - no loss of contrast.
Further investigation today (25th Nov 2024) I can confirm that this bug was also in Kdenlive 24.08.1 using Ubuntu 20.04, KDE 6.5.0 and Qt 6.7.2 Kdenlive 23.08.4 with Ubuntu 20.04 and KDE 5.112.0 and Qt 5.15.10 was working OK - this bug was not present.
Snap package of Kdenlive 24.08.3 with Ubuntu 24.04, KDE 6.6.0, and Qt 6.7.2, MLT 7.24.0 seems to be OK. It's possible that this bug is confined to the AppImage package mentioned in the bug report.
Thank you for reporting. I can confirm the issue. Some investigating on MLT is ongoing already: https://github.com/mltframework/mlt/commit/68319ffb2bcaa5dacb65bbb5639c2f02fa5d867e https://github.com/mltframework/mlt/pull/1054