Hi,
Is it possible to get the command line used by DGDemux to combine Dolby Vision tracks ?
Thank you.
DoVi Tool
DoVi Tool
Thank you and sorry, I didn't see this topic...
Just a question : any reason to use --eos-before-el option ?
I use the latest dovi tool and quietvoid say "Write the EOS/EOB NALUs before the EL. Defaults to false.
This flag enables the same behaviour yusesope's mux script.
Enabling this therefore results in identical output using dovi_tool"
Thank you.
Just a question : any reason to use --eos-before-el option ?
I use the latest dovi tool and quietvoid say "Write the EOS/EOB NALUs before the EL. Defaults to false.
This flag enables the same behaviour yusesope's mux script.
Enabling this therefore results in identical output using dovi_tool"
Thank you.
DoVi Tool
The --eos-before-el options makes the output equal to the output from MakeMKV.Nico83500 wrote: ↑Wed Feb 28, 2024 9:37 amThank you and sorry, I didn't see this topic...
Just a question : any reason to use --eos-before-el option ?
I use the latest dovi tool and quietvoid say "Write the EOS/EOB NALUs before the EL. Defaults to false.
This flag enables the same behaviour yusesope's mux script.
Enabling this therefore results in identical output using dovi_tool"
Thank you.
Not using it, according to the dovi_tool help, follows the HEVC standard though. No idea if it makes any difference wether one uses it or not. The end result is not the same, but the differences doing an filecompare are not huge.
renols
DoVi Tool
Thank you. I don't know if MakeMKV still use --eos-before-el option and why quietvoid set it to false. But mkvtoolnix output is the same than dovi_tool without this option.
DoVi Tool
I asked to quietvoid, dovi tool creator, and "(enable this option) is not really relevant today"
DoVi Tool
Ok, thank you for the update. Seems no reason to change anything, though, yes?
DoVi Tool
It doesn't seem to impact the output (full quote from quietvoid's Github) :
EOS/EOB NALUs usually signal the end of a bitstream OR segment (in the case of seamless branching).
So when muxing the EL stream into a BL containing EOS/EOB, the EL is inserted either before or after depending on the setting.
By default the EL is inserted before EOS/EOB from the BL, to avoid issues with parsers stopping early when encountering one and ignoring the EL.
It should have no impact since they are not commonly in the middle of the bitstream.
The option is only there to reproduce bitstreams the same as yusesope's script.
In the HEVC specification, EOS/EOB type should always be the last NALUs in the access unit.
This is how inject-rpu behaves as well, inserting the UNSPEC62 (RPU) NALU before EOS/EOB.
FFmpeg is the only known parser that may ignore the RPU before EOS/EOB, because it always expects that the AU ends with UNSPEC62.