Ripping A Video From DVD Using FFMPEG

I have a collection of DVDs I need to put in storage to clear up some space. This page lists what I have learned about how to take DVD video and create a video file as a result.

There are many other articles on the subject, such as this one.

A DVD usually has two directories on it, AUDIO_TS and VIDEO_TS. For DVD videos we only care about the VIDEO_TS directory.

Inside the VIDEO_TS directory there are a collection of files, e.g.:

     16384 Aug 23  2005 VIDEO_TS.BUP
     16384 Aug 23  2005 VIDEO_TS.IFO
    157696 Aug 23  2005 VIDEO_TS.VOB
     30720 Aug 23  2005 VTS_01_0.BUP
     30720 Aug 23  2005 VTS_01_0.IFO
 176521216 Aug 23  2005 VTS_01_0.VOB
  22235136 Aug 23  2005 VTS_01_1.VOB
     43008 Aug 23  2005 VTS_02_0.BUP
     43008 Aug 23  2005 VTS_02_0.IFO
    157696 Aug 23  2005 VTS_02_0.VOB
1073565696 Aug 23  2005 VTS_02_1.VOB
1073565696 Aug 23  2005 VTS_02_2.VOB
 839757824 Aug 23  2005 VTS_02_3.VOB

We’re only interested in *.VOB files, and generally anything after VTS_02_* as VTS_01_* is usually the contents page for the DVD.

So, we want to make a video out of the VTS_02_*.VOB files. Before we can proceed any further it is necessary for us to make a single, continuous .mpeg file. Be warned, however, this file will be large (3 gigabytes for 45 minutes).

Thus we want to have an input file that is a concatenation of the VOB files, in order. We use the input file type of concat:file_1:file_2:…

Again, before we proceed further, we need to find out what streams are in the source video:

~$ ffmpeg -i "concat:VTS_02_0.VOB|VTS_02_1.VOB|VTS_02_2.VOB|VTS_02_3.VOB" -c copy /dev/null
Input #0, mpeg, from 'concat:VTS_02_0.VOB|VTS_02_1.VOB|VTS_02_2.VOB|VTS_02_3.VOB':
  Duration: 00:45:08.97, start: 0.280000, bitrate: 8821 kb/s
    Stream #0:0[0x1bf]: Data: dvd_nav_packet
    Stream #0:1[0x1e0]: Video: mpeg2video (Main), yuv420p(tv), 720x576 [SAR 16:15 DAR 4:3], 8000 kb/s, 25 fps, 25 tbr, 90k tbn, 50 tbc
    Stream #0:2[0x20]: Subtitle: dvd_subtitle
    Stream #0:3[0x81]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
    Stream #0:4[0x80]: Audio: ac3, 48000 Hz, stereo, fltp, 224 kb/s

Now, when we create our intermediate (temporary) .mpeg file, we need to specify which streams we want to extract. By default ffmpeg only extracts the best video and best audio stream it can find. If we want to keep all our audio streams then we have to explicitly tell ffmpeg this using the -map directive, one per stream we want to keep, in order.

In this case I want to keep streams 0:1 (video), 0:3 (5.1 audio), and 0:4 (audio stereo). But actually I want the stereo track to come before the 5.1 audio, so I’ll use the following command to make my intermediate file:

~$ ffmpeg -i "concat:VTS_02_0.VOB|VTS_02_1.VOB|VTS_02_2.VOB|VTS_02_3.VOB" -map 0:1 -map 0:4 -map 0:3 -f mpeg -c copy /tmp/intermediate.mpeg
Input #0, mpeg, from 'concat:VTS_02_0.VOB|VTS_02_1.VOB|VTS_02_2.VOB|VTS_02_3.VOB':
  Duration: 00:45:08.97, start: 0.280000, bitrate: 8821 kb/s
    Stream #0:0[0x1bf]: Data: dvd_nav_packet
    Stream #0:1[0x1e0]: Video: mpeg2video (Main), yuv420p(tv), 720x576 [SAR 16:15 DAR 4:3], 8000 kb/s, 25 fps, 25 tbr, 90k tbn, 50 tbc
    Stream #0:2[0x20]: Subtitle: dvd_subtitle
    Stream #0:3[0x81]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
    Stream #0:4[0x80]: Audio: ac3, 48000 Hz, stereo, fltp, 224 kb/s
Output #0, mpeg, to '/tmp/intermediate.mpeg':
  Metadata:
    encoder         : Lavf57.15.100
    Stream #0:0: Video: mpeg2video, yuv420p, 720x576 [SAR 16:15 DAR 4:3], q=2-31, 8000 kb/s, 25 fps, 25 tbr, 90k tbn, 25 tbc
    Stream #0:1: Audio: ac3, 48000 Hz, stereo, 224 kb/s
    Stream #0:2: Audio: ac3, 48000 Hz, 5.1(side), 448 kb/s
Stream mapping:
  Stream #0:1 -> #0:0 (copy)
  Stream #0:4 -> #0:1 (copy)
  Stream #0:3 -> #0:2 (copy)
video:2645654kB audio:222221kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.389088%

So now we have an intermediate file. It is from this we can create the desired output.

You might have an encrypted region-specific DVD. In this case the ffmpeg tool will not work by itself. A tool that will read encrypted DVDs is a port of ffmpeg called vgtmpeg. There are Windows and Linux builds available. This uses the same command line except for the input you specify an input of something like:

~$ vgtmpeg -i dvd:///media/user/DVD_VIDEO/?title=3
C:\> vgtmpeg.exe -i dvd://F:\title=3

Note that region-coded DVDs will not work in a DVD drive that has no region code set (some new drives have no code initially configured). In Linux you can set a drive’s region code (a maximum of 5 times) using the command-line tool regionset. Once the drive’s region has been set, however, the vgtmpeg tool should work with a drive from any region even if it is different from the drive’s specified region.

Now you can compress the intermediate file.

You will probably want to tinker with the settings to suit your preferences.

First you have to specify which streams you want to encode, again. You may be forced to specify the display aspect ratio. And you may wish to add metadata (descriptions) to the audio channels (and subtitles if you use them).

Here are some settings for an example:

~$ ffmpeg -i /tmp/intermediate.mpeg \
  -map 0:0 -map 0:1 -map 0:2 \
  -c:v libx264 -preset medium -b:v 600k -aspect 16:9 -strict -2 -g 50 -vf yadif \
  -c:a:0 aac -b:a:0 160k -metadata:s:a:0 title=\"2 channel\" \
  -c:a:1 ac3 -b:a:1 320k -metadata:s:a:1 title=\"5.1 channel\" \
  /tmp/final.mkv

The -map directives select what streams we want from our input file.

The -c:v libx264 directive selects the video codec. The -preset option chooses an H.264 encoding speed to compression ratio. The -b:v option specifies a constant output bitrate (you can use two-pass encoding if you wish to utilise a different strategy). The -aspect 16:9 may not be necessary but forces the default display by the client reading the file later. The -strict -2 options are required for H.264 encoding. The -g 50 option makes the encoder store a whole image frame every 50 frames (useful for being able to forward or reverse into the video). The -vf yadif video filter de-interlaces the source video.

The -c:a:0 option sets the audio codec on the first audio stream (which is different from the stream number assigned). The -b:a:0 option sets the bit rate for the first audio stream. And the -metadata:s:a:0 option sets a title for the stream, which can make it easier for those watching your video to choose an appropriate named audio track rather than guessing what the difference is.

Finally the output file name is specified.

Audio

A guide to high quality audio with ffmpeg is documented on the ffmpeg wiki.

A 5.1 stream is perhaps better encoded using the “ac3” (Dolby Digital) codec.

If you want to convert a 5.1 stream to stereo you can use the following to select a particular audio stream and apply a filter to it:

-af:a:0 'pan=stereo|FL=FC+0.30*FL+0.30*BL|FR=FC+0.30*FR+0.30*BR'

Subtitles

DVD subtitles may not appear in the first VOB file in a set; it may be necessary to tell ffmpeg to scan ahead deep into the file so that it can find the subtitles in the second VOB file. The directives to do this are -probesize and -analyzeduration – the first being the number of bytes to read ahead into the file, and the second being the number of microseconds (of which there are 1,000,000 per second). These directives must come before the input file directive.

So, to scan ahead 4GB and 1 hour, the an example command may be:

~$ ffmpeg -i -probesize 4G -analyzeduration 3600M "concat:VTS_02_0.VOB|VTS_02_1.VOB|VTS_02_2.VOB|VTS_02_3.VOB" -c copy /dev/null

You can specify a title to go along with a subtitle track, such as:

-metadata:s:s:0 title="English"