Summary
When using the MRQ in 5.8 with the ‘Basic’ settings, rendering to MP4, the resulting video produces an error when transcoding in Syncsketch with the error: “the specified clip start/length is outside the duration of this file”. Syncsketch takes and upload and does a conform transcoding by default. This transcoding can be bypassed by holding Shift when drag and dropping for upload.
The default however, will fail on processing when using the Basic settings workflow.
I believe this is a result of the audio muxing and it fails even when the sequence has no audio. I believe the audio track is getting the wrong duration or no duration metadata is being set for it.
I found that using the ‘Preset’ settings with minimal settings and H264 output (to mimic the ‘Basic’ workflow), the issues does not present and Syncsketch has no issue transcoding. Inspecting output log for both renders, I don’t see any clear indicators of a different pipeline. When inspecting the mp4 files with MediaInfo, the only difference I can see is the ‘Presets’ workflow sets a duration in the audio file and the ‘Basic’ does not, and the ‘Basic’ rounds up the duration. The ‘Presets’ workflow also includes a toggle to ‘include audio’. The toggle in either state has no issue transcoding. Then the toggle is off, no audio is in the mp4 (as expected). When on, correctly formatted audio is written into the mp4 (with no sound in my case).
What Type of Bug are you experiencing?
UI / Tools
Steps to Reproduce
Render locally a sequence with no audio using the ‘Basic’ settings and output MP4. Render a second using the H264 output. Include audio can be on but audio in the sequencer is not required. Report the the length of both with ffprobe:
ffprobe -v error -select_streams a:0 -show_entries stream=duration,nb_frames,start_time -of default=noprint_wrappers=1 .\basicMP4_LS_LightDemo.mp4
start_time=0.000000
duration=4.999958
nb_frames=N/A
ffprobe -v error -select_streams a:0 -show_entries stream=duration,nb_frames,start_time -of default=noprint_wrappers=1 .\Preset_LS_LightDemo.mp4
start_time=0.000000
duration=4.970646
nb_frames=233
Or this one which walks the file for samples:
ffprobe -v error -select_streams a:0 -count_packets -show_entries stream=nb_read_packets -of default=noprint_wrappers=1 .\basicMP4_LS_LightDemo.mp4
nb_read_packets=N/A
ffprobe -v error -select_streams a:0 -count_packets -show_entries stream=nb_read_packets -of default=noprint_wrappers=1 .\Preset_LS_LightDemo.mp4
nb_read_packets=233
Expected Result
The Basic workflow should handle audio in the same way as the Presets workflow with MP4 outputs and make an identical file
Observed Result
the Basic workflow shows no audio packets and incorrect duration for audio tracks, while the Preset workflow shows audio channels being formatted correctly with the correct duration
Affects Versions
5.8
Platform(s)
Windows
Upload an image
Additional Notes
I like that there is a ‘Basic’ workflow and it would be great if it produced the same output as the ‘Presets’ workflow when rendering from Windows directly to MP4
I suppose optionally a toggle to disable audio in the Basic workflow (like the Presets workflow) would also solve this in the case where there is no audio present in sequencer