Remove foreign language audio and subtitle tracks from mkv files in bulk

removeNonEnglish.py

#!/usr/bin/python

# Removes non-LANG audio tracks and subtitles from mkv files in a directory.
# Original script by greenbender at https://forum.videohelp.com/threads/343271-BULK-remove-non-English-tracks-from-MKV-container
# Modified by Joseph Milazzo for updated MkvMerge commands.
# 12/3/2021: Updated to Python 3.9

import os
import re
import sys
import subprocess
import json

# change this for other languages (3 character code)
LANG = "eng"

# set this to the path for mkvmerge
MKVMERGE = "mkvmerge.exe"

AUDIO_RE = re.compile(
    r"Track ID (\d+): audio \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})")
SUBTITLE_RE = re.compile(
    r"Track ID (\d+): subtitles \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})(?: track_name:\w*)? default_track:[01]{1} forced_track:([01]{1})")

if len(sys.argv) < 2:
    print("Please supply an input directory")
    sys.exit()

in_dir = sys.argv[1]

for root, dirs, files in os.walk(in_dir):
    for f in files:

        # filter out non mkv files
        if not f.endswith(".mkv"):
            continue

        # path to file
        path = os.path.join(root, f)

        # build command line
        cmd = [MKVMERGE, "-J", path]

        # get mkv info
        mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
        stdout, stderr = mkvmerge.communicate()
        if mkvmerge.returncode != 0:
            print >> sys.stderr, "mkvmerge failed to identify", path
            continue

        # find audio and subtitle tracks
        audio = []
        subtitle = []
        info_json = json.loads(stdout)
        tracks = info_json['tracks']
        for track in tracks:
            track['properties']['id'] = track['id']
            if track['type'] == 'audio':
                audio.append(track)
            elif track['type'] == 'subtitles':
                subtitle.append(track)

        # filter out files that don't need processing
        if len(audio) < 2 and len(subtitle) < 2:
            print >> sys.stderr, "nothing to do for", path
            continue

        # filter out tracks that don't match the language
        audio_lang = filter(lambda a: a['properties']['language'] == LANG, audio)
        subtitle_lang = filter(lambda a: a['properties']['language'] == LANG, subtitle)

        # filter out files that don't need processing
        if len(audio_lang) == 0 and len(subtitle_lang) == 0:
            print >> sys.stderr, "no tracks with that language in", path
            continue

        # build command line
        cmd = [MKVMERGE, "-o", path + ".temp"]
        if len(audio_lang):
            cmd += ["--audio-tracks", ",".join([str(a['id']) for a in audio_lang])]
            for i in range(len(audio_lang)):
                cmd += ["--default-track", ":".join([str(audio_lang[i]['id']), "0" if i else "1"])]
        if len(subtitle_lang):
            cmd += ["--subtitle-tracks", ",".join([str(s['id']) for s in subtitle_lang])]
            for i in range(len(subtitle_lang)):
                cmd += ["--default-track", ":".join([str(subtitle_lang[i]['id']), "0"])]
        cmd += [path]

        # process file
        print >> sys.stderr, "Processing", path, "...",
        mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
        stdout, stderr = mkvmerge.communicate()
        if mkvmerge.returncode != 0:
            print >> sys.stderr, "Failed"
            print(stdout)
            continue

        print >> sys.stderr, "Succeeded"

        # overwrite file
        os.remove(path)  # Don't overwrite
        os.rename(path + ".temp", path)

free to use for whatever. no terms.

Licensing is free to use as well. There was no license on the code i found floating around.

thank you for the fast responses. according to https://stackoverflow.com/a/71817350 you're supposed to ask the authors if none is provided

i get this error when i run it how can i fix it?

moe@moeserver:/hdd/TEST$ python3 removeNonEnglish.py /hdd/TEST
Traceback (most recent call last):
File "/hdd/TEST/removeNonEnglish.py", line 90, in
print >> sys.stderr, "Processing", path, "...",
TypeError: unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.TextIOWrapper'. Did you mean "print(, file=<output_stream>)"?
moe@moeserver:/hdd/TEST$

@i0moe that error means the code still has some python 2 and the print statement needs to be rewritten to use python 3 syntax print().

@i0moe It's been a while, but that's probably fixed in my modified script which worked for me on python 3.9?

@majora2007 would you please update this script to keep multiple audio/subtitles tracks in a specified language?

@majora2007 would you please update this script to keep multiple audio/subtitles tracks in a specified language?

Check this line to reconfigure the languages to keep: LANG = ["eng","nor"]

If you need to differentiate between subtitle and audio languages you can check my fork of the file at https://github.com/seantcanavan/apex-plex/blob/main/main.py

The lines in question are:

# change this for other languages (3 character code)
AUDIO_LANG = ["eng", "jpn"]
SUBTITLE_LANG = ["eng"]

Unfortunately it does not work on python 3.11

  File "script.py", line 73, in <module>
    if len(audio_lang) == 0 and len(subtitle_lang) == 0:
       ^^^^^^^^^^^^^^^
TypeError: object of type 'filter' has no len()

Hey,
It doesn't seem to work, os.walk(in_dir): seems to return nothing, all my mkv files are in 'Z:\Season 1' folder
But the program doesn't find any files

It seems it is a problem if the files are on a network drive

Hey, It doesn't seem to work, os.walk(in_dir): seems to return nothing, all my mkv files are in 'Z:\Season 1' folder But the program doesn't find any files

It seems it is a problem if the files are on a network drive

There are a lot of issues in general when working across operating systems and file systems over the network.

I spent hours one time trying to debug something and I was running Python on Windows but accessing files on a Linux shared drive. Windows is not case sensitive and linux is and it was a huge learning moment for me.

I would encourage you to try running the script on the same platform as your network drive or dig further into the exact file systems you're running on both devices and why you might be seeing an issue that you are.