PAL/NTSC aware v210/MOV to FFV1/MKV recipe

transcode.py

#!/usr/bin/env/python

import os
import argparse
import subprocess
import multiprocessing


### For extra utitily when running this script remotely
# nohup transcode.py -d directory/of/mov -e mov -o directory/of/future/mkv &
# The & puts the job in the background
# The nohup autologs ffmpeg output to a nohup.out file and keeps the process alive even when the ssh session stops

# Turn into command line script
def _make_parser():
    parser = argparse.ArgumentParser()
    parser.description = "survey designated folder and child folders for files by extensions"

    parser.add_argument("-d", "--directory",
    help = "path of directory to survey",
        required = True
    )
    parser.add_argument("-e", "--extension",
        help = "type of extension to find",
        required = True
    )
    parser.add_argument("-o", "--output",
        help = "directory to store stuff",
        required = True
    )
    return parser

  
# Create ffmpeg recipe based on height of video
def get_transcode_video_cmd(fn, height, output):
    ffv1_command = [
        'ffmpeg',
        '-vsync', '0',
        '-i', fn,
        '-map', '0',
        '-dn',
        '-c:v', 'ffv1',
        '-level', '3',
        '-coder', '1',
        '-context', '1',
        '-g', '1',
        '-slicecrc', '1',
        '-slices', '24',
        '-c:a', 'flac',
        ]
    if height == '486':
        ffv1_command += [
        '-field_order', 'bt',
        '-vf', 'setfield=bff,setdar=4/3',
        '-color_primaries', 'smpte170m',
        '-color_range', 'tv',
        '-color_trc', 'bt709',
        '-colorspace', 'smpte170m',
        ]
    if height == '576':
        ffv1_command += [
        '-field_order', 'tb',
        '-vf', 'setfield=tff,setdar=4/3',
        '-color_primaries', 'bt470bg',
        '-color_range', 'tv',
        '-color_trc', 'bt709',
        '-colorspace', 'bt470bg',
        ]
    ffv1_command += [os.path.join(output, os.path.splitext(os.path.basename(fn))[0] + '.mkv',)]
    return ffv1_command


# function to run the ffmpeg command
def work(cmd):
    subprocess.check_call(cmd)
    return


# This is where the script actually starts
def main():
    # Get the arguments entered on the command line
    args = _make_parser().parse_args()

    # Create a list to hold a whole bunch of transcode commands
    queue = []
    
    # Generate a command for every file in a directory that matches the extension.
    for dirpath, dirnames, filenames in os.walk(args.directory):
        for filename in [f for f in filenames if f.endswith(args.extension)]:
            # Make sure to use full path so we don't deal with working directory hijinks
            full_path = os.path.join(dirpath, filename)
            # Run MediaInfo to check height of video (and so NTSCness or PALness)
            # Might be good to use pymediainfo in the future, but .rstrip().decode('UTF-8') works well enough to get the data we want
            height = subprocess.check_output(
                ['mediainfo', '--Language=raw', '--Full',
                "--Inform=Video;%Height%", full_path]).rstrip().decode('UTF-8')
            # Use the input file name, height, and output path to generate an ffmpeg command, and then append that line to the list
            queue.append(get_transcode_video_cmd(full_path, height, args.output))

    # Quick sanity check for fun
    print(queue)
    
    # Use multiprocess to manage the queue of transcode work.
    # It will take one command from the list at-a-time, run the command, and on completion take the next one
    # ffmpeg already know how to parallize work well, so it's probably not useful to use processes= >1
    pool = multiprocessing.Pool(processes=1)
    r = pool.map_async(work, queue)
    r.wait()


if __name__ == '__main__':
    main()