Use of simple vad method or combined with filters?

[libero08]

According with this :
-l tr -m large-v2 --vad_method pyannote_v3 -o source --ff_vocal_extract mb-roformer
wich are the difference between using one method only vad as example : -vad_method=pyannote_v3 or --ff_vocal_extract=mb-roformer
wich is better : --standard or sentece and which are the difference .
whre where i can find a wiki of this commands ...?
can you replay here with a professional comand wixch include more params like number of character in single line or where to add break pluse use of vad and audio filter. Thank you. ?
i use GPU system.

[Purfview]

whre where i can find a wiki of this commands ...?

Use --help

wich are the difference between using one method only vad as example : -vad_method=pyannote_v3 or --ff_vocal_extract=mb-roformer

Not sure I understand your question because the second argument is not VAD.

wich is better : --standard or sentece and which are the difference .

There is no "better", these are two different things, read the help.

can you replay here with a professional comand wixch include more params like number of character in single line or where to add break

--max_line_width [but then don't use --standard because it will override it]

[libero08]

Escuse my bad english at this hour, ff vocal is a filter ... vad is a metod of detection ... is better use both or only vad method /filter.

--max_line_width i dont knew exactly, subedit use default 43 , normaly 48 is almost 3 seconds of time on screen, bud this depend also for the person who can read more chars ( ... i am able to read 23-25 ) standard is above 17,5 but this parameter need to be combined with another one for breacking at X .
i will use help and i will return if i considering that.
Thank you for that quickly answer.

[Purfview]

is better use both or only vad method /filter.

Depends from audio, if there is speech without the larger gaps then no need to use VAD, if there is no non-speech "noise" then no need to use voice extraction.

--max_line_width i dont knew exactly, subedit use default 43 , normaly 48 is almost 3 seconds of time on screen, bud this depend also for the person who can read more chars ( ... i am able to read 23-25 ) standard is above 17,5 but this parameter need to be combined with another one for breacking at X .

It's not related to "time on screen". [there is no option at all related to that]

[libero08]

Thank you for replay.