The new encodedock.cpp code is designed to support 2-pass encoding based on the parameters given in the SVT-AV1 documentation. However, 2-pass encoding is not wired up between SVT-AV1 and FFmpeg internally yet. This means Shotcut will signal for 2-pass but it won't actually happen. Given that the SVT-AV1 team specifically changed their multi-pass API/ABI to be compatible with FFmpeg in v2.0.0, I assume this will be fixed sooner than later, hence why I didn't disable 2-pass encoding altogether.
Known quirks:
When using average bitrate mode (VBR), set B-frames to 1 on the Shotcut export panel. VBR mode requires "random-access mode", which means P/B-frames, and is signaled by pred-struct=2.
When using constant bitrate mode (CBR), set B-frames to 0 on the Shotcut export panel. CBR mode requires "low delay mode", which means P-frames only, and is signaled by pred-struct=1.
Known limitation:
encodedock.cpp
code is designed to support 2-pass encoding based on the parameters given in the SVT-AV1 documentation. However, 2-pass encoding is not wired up between SVT-AV1 and FFmpeg internally yet. This means Shotcut will signal for 2-pass but it won't actually happen. Given that the SVT-AV1 team specifically changed their multi-pass API/ABI to be compatible with FFmpeg in v2.0.0, I assume this will be fixed sooner than later, hence why I didn't disable 2-pass encoding altogether.Known quirks:
pred-struct=2
.pred-struct=1
.