Add Stereo Audio Tracks to MKV Files

Posted by admin on September 26, 2009 under Tech Tips | Be the First to Comment

If you have Matroska Video (MKV) files encoded with AC3 Dolby Digital 5.1 or DTS audio tracks, you may want to simply extract the audio, convert it to a 2-channel stereo format like WAV, MP3 OGG, etc, and then add it back into the MKV as a separate audio track. This is useful when your media player (e.g. Western Digital Media Player WDAVN00) will not downscale the audio from a digital format like AC3 or DTS to stereo when you don’t have a receiver or TV with a built in Dolby Digital decoder.  Now you’ll have the choice of either audio format depending on your technical requirements.

The great thing about the Matroska multimedia container is that you can easily manipulate these files without having to re-encode, saving lots of time. I’ll be using mkvextract to extract the AC3 audio, ffmpeg to convert ac3 to mp3, and finally mkvmerge to add and remux the new audio track to the MKV container. All of these are available to a number of platforms, but in my examples, I’m using Linux.  Check out the MKVToolnix and FFMpeg websites for more info on the software.

If using Ubuntu Linux, install the relevant mkvtoolnix,  mkvtoolnix-gui and ffmpeg packages.

sudo apt-get install mkvtoolnix mkvtoolnix-gui ffmpeg libavcodec-unstripped-52

To view the existing tracks of the MKV, use the mkvmerge -i option. In the following example, you see my “Cool.Video.mkv” file has an MPEG4 video in track 1, an AC3 Dolby Digital audio file in track 2, and subtitles in track 3.

mkvmerge -i Cool.Movie.mkv 
File 'Cool.Movie.mkv': container: Matroska
Track ID 1: video (V_MPEG4/ISO/AVC)
Track ID 2: audio (A_AC3)
Track ID 3: subtitles (S_TEXT/UTF8)

Using mkvextract, extract the AC3 Dolby Digital audio from track 2, saving it to a file called audio.ac3.

mkvextract tracks Cool.Movie.mkv 2:audio.ac3
Extracting track 2 with the CodecID 'A_AC3' to the file 'audio.ac3'. Container format: Dolby Digital (AC3)
Progress: 100%

ls -lh audio.ac3
-rw-r--r-- 1 gmendoza gmendoza 432M 2009-09-26 11:58 audio.ac3

Convert the 6-channel ac3 file to a 2-channel stereo MP3 using ffmpeg. If you prefer a higher audio bitrate, adjust the -ab value as desired. e.g. 256, 384, etc, and adjust the audio rate to your liking as well.

ffmpeg -i audio.ac3 -acodec libmp3lame -ab 160k -ac 2 audio.mp3
[output omitted for brevity]

ls -lh audio.*
-rw-r--r-- 1 gmendoza gmendoza 432M 2009-09-26 11:58 audio.ac3
-rw-r--r-- 1 gmendoza gmendoza 87M 2009-09-26 12:08 audio.mp3

To simplify things, you could actually skip the digital format extraction process by running ffmpeg against the MKV file directly.

ffmpeg -i Cool.Movie.mkv -acodec libmp3lame -ab 160k -ac 2 audio.mp3

If you prefer encoding with more advanced options, you could extract the audio as a 2-channel WAV file instead, and then process it with LAME, Oggenc, or some other encoder of your choosing. The following shows the extraction to WAV, and then conversion to various formats for fun, e.g. MP3, OGG, and FLAC.

ffmpeg -i Cool.Movie.mkv -acodec pcm_s16le -ac 2 audio.wav
lame -V0 -q0 --vbr-new audio.wav audio.mp3
oggenc -q6 audio.wav
flac audio.wav

Use mkvmerge to combine the original MKV with the MP3 audio track to create a new file called Cool.Movie.New.mkv. Make sure you have enough disk space for both the original and new MKV file.

mkvmerge -o Cool.Movie.New.mkv Cool.Movie.mkv audio.mp3
mkvmerge v2.4.1 ('Use Me') built on Dec 13 2008 21:03:46
'Cool.Movie.mkv': Using the Matroska demultiplexer.
'audio.mp3': Using the MP2/MP3 demultiplexer.
Warning: 'audio.mp3': Skipping 32 bytes at the beginning (no valid MP3 header found).
'Cool.Movie.mkv' track 1: Using the MPEG-4 part 10 (AVC) video output module.
'Cool.Movie.mkv' track 2: Using the AC3 output module.
'Cool.Movie.mkv' track 3: Using the text subtitle output module.
'audio.mp3' track 0: Using the MPEG audio output module.
The file 'Cool.Movie.New.mkv' has been opened for writing.
Progress: 100%
The cue entries (the index) are being written...
Muxing took 270 seconds.

Verify that the audio track has been added. You can see Track ID 4 has been successfully added.

mkvmerge -i New.Cool.Movie.mkv 
File 'New.Cool.Movie.mkv': container: Matroska
Track ID 1: video (V_MPEG4/ISO/AVC)
Track ID 2: audio (A_AC3)
Track ID 3: subtitles (S_TEXT/UTF8)
Track ID 4: audio (A_MPEG/L3)

That’s really all there is to it. There are quite a few options available when editing MKV container files. For example, I wanted nice descriptions for my tracks since various media players will read and display them for you during menu navigation. I recommend using the mkvmerge gui application as shown in this screenshot.

It’s really just a front-end application to mkvmerge, and the following text shows the commands that were used to specify the language for each tag, re-order the audio tracks, disable subtitles by default, and give useful descriptions to each Track ID.

mkvmerge -o "Cool.Movie.New.mkv" \
--language 1:eng \
--track-name "1:Cool Movie (MPEG4)" \
--default-track 1:yes \
--display-dimensions 1:40x17 \
--language 2:eng \
--track-name "2:Dolby Digital 5.1 (AC3)" \
--default-track 2:yes \
--language 3:eng \
--track-name "3:English Subtitles" \
--default-track 3:no \
-a 2 -d 1 -s 3 Cool.Movie.mkv \
--language 0:eng \
--track-name "0:2-Channel Stereo (MP3)" \
--default-track 0:no \
-a 0 -D -S audio.mp3 \
--track-order 0:1,0:2,1:0,0:3

mkvmerge -i Cool.Movie.New.mkv
File 'Cool.Movie.New.mkv': container: Matroska
Track ID 1: video (V_MPEG4/ISO/AVC)
Track ID 2: audio (A_AC3)
Track ID 3: audio (A_MPEG/L3)
Track ID 4: subtitles (S_TEXT/UTF8)

Batch MP3 Encoding with Linux and LAME

Posted by admin on December 2, 2007 under Tech Tips | Be the First to Comment

If you have a number of audio files that you would like to convert to mp3, but don’t want to hassle with graphical applications, there is a simple way to accomplish the task using Linux, LAME and a little shell scripting. By performing a basic “for loop” to invoke LAME, you can easily convert any group of audio files using one line of shell code.

Here is an example of a “for loop” that runs the command “lame” against a set of files in your current working directory with the .wav file extension.

$ for f in *.wav ; do lame $f ; done

If your audio file names have spaces in them, then you will need to use quotation marks around “$f” variable.

$ for f in *.wav ; do lame "$f" ; done

I typically create my original audio files without the file name extension of .wav or .au. This is because when you run lame against a file name, and do not omit the extension in the output option, the resulting file will have two extensions in the file name. e.g. filename.wav.mp3. Yes, I can use sed or basename in the for loop to prevent this, but to keep it simple, I just choose to not use the file extension to begin with.

If you are working with a group of files that have all been named using the convention of “Artist – Album – ## – Track Title”, (notice the spaces in the name), the following will work.

$ for f in Artist - Album* ; do lame "$f" ; done

Once the job is finished, you will be left with a directory full of your original audio files, and your newly created mp3’s.

Extra Credit

Okay, since we’re on the topic of shell scripting. If you want to delete all the original audio files (the ones without any file name extensions), and without first moving the new mp3’s to a different directory, one overly complicated example would be the following.

for f in *.mp3 ; do AUFILE=`basename "$f" .mp3` ; rm "$AUFILE"; done

This would have been easier if the original files could have been identified with .wav extensions, (rm *.wav), but since they had no file extensions to begin with, a wild card alone would not work. Now folks, this is just an example, and there’s a million other ways you can go about this. But in any case, I hope it helps you start exploring on your own!