search

nstanger / process_podcast

Python script for processing and assembling a podcast from various input sources
0 stars 0 forks source link

Differently sized video inputs #24

Open nstanger opened 8 years ago

nstanger commented 8 years ago

Issue by nstanger Tuesday Sep 20, 2016 at 12:04 GMT Originally opened as https://github.com/Otago-InfoSci-Database/Tools/issues/24

Not very likely to happen, but if there are video inputs of different sizes, we should resize the larger ones to match the input with the smallest dimensions. This already happens for frame inputs (see #23).

The way to do this is to prepend a -filter:v scale=w:h to the output options whenever we generate a video file, where w and h are the scaled width and height, respectively. (Or maybe just do it on the final render regardless of whether it’s needed? It has to re-encode everything at that point anyway, so a resize to the same dimensions should have no appreciable effect on output quality, and it’s eventually going to get horribly mangled by Otago Capture in any case.)

nstanger commented 8 years ago

Confirmed: a resize to the same dimensions as the input while also re-encoding produces identical output as a straight re-encode. :thumbsup:

nstanger commented 7 years ago

Can’t use both -filter:v and -filter_complex in the same command line 😦. Need to append to the complex filter instead.

nstanger commented 7 years ago

Probably need a --resize option. Simply resizing to smallest is fine until one of your video sources is much smaller than the others. Perhaps --resize [WxH]: if the dimensions are omitted, just resize everything the same as the smallest video input.