Description: Python wrapper around ffmpeg and ffprobe to perform basic
conversions, alterations and muxing.
Author: Josh Buchbinder
ffmpeg-cnvt is a python script that uses ffmpeg and ffprobe to
perform some basic media file manipulation. The only required arguments
are input and output specifiers. The input may be a single file, multiple
files using a wildcard or a printf style numeric identifier such as
TLPS%04d.JPG when using the -sequence option. The output may be a single
file or a directory where output files will be created.
By default ffmpeg-cnvt will attempt to copy all streams of all types from
the input to the output without processing them, which can be used to change
the media container using -container, or by simply specifying a single
output file, the extension of that file will be used. By default,
ffmpeg-cnvt will use the input file type. Note that some stream types or
codecs may not be supported with all containers.
To copy only a single stream of a type from the primary input, for instance
the second audio stream, the -singlestream option can be used such as
-singlestream a 1 where a indicates audio streams and 1 is the zero
based index of audio streams in the input file. To prevent streams of a type
from being copied from the primary input, -nocopy can be used.
To add additional streams to the output other than those from the primary
input, -addfile is used to specify the stream type and path to the media
file to import streams from. As with the primary input, all streams of the
selected type will be copied unless -addstream is used to specify a
specific stream.
By default, streams added using -addfile will be ordered after the streams
of the same type from the primary input unless -addfirst is used. The
first of the added streams can be marked as the default stream of a type using
-adddefault, all other streams of that type from the primary or secondary
inputs will be marked as not default.
When combining streams of different lengths, ffmpeg by default will continue
encoding until the end of the last stream is reached. The length can be
limited by specifying a time period using -duration or limited to the
length of the shortest stream using -shortest. This can be combined with
-addlooped to have the streams of a shorter added input looped.
To process streams rather than simply copying them, a codec must be specified
using -codec or -addcodec. To display the codecs currently supported by
ffmpeg-cnvt use -codecs. There are additional arguments (-tune,
-preset, -bitrate, -mono, -stereo etc) that can be used to control
the encoding codec parameters.
By default, ffmpeg-cnvt will output video at the same resolution it was
input. If a resolution is specified using one of the resolution arguments
(-hd, -fhd, -qhd, -uhd) or by specifying -width and/or -height
the output will be scaled to fit within the supplied dimensions while
maintaining the aspect ratio. If -crop is used, the output video will be
cropped from the input video using the position supplised to -crop.
The supported containers and codecs are simply the ones I use most often, more can be added easily enough. More options and defaults for tuning encoding are likely needed for some codecs.
You can currently only specify encoding settings once so it is not possible
to encode streams from the primary and secondary inputs with different
settings such as -preset, -stereo or -bitrate.
As of Python 3.9, CPU affinity calls only work on "some Unix systems".