Difference between revisions of "Extract DVD"
From Maze's wiki
(→Space not an Issue?) |
(→Subtitles) |
||
| (41 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
| + | ==Decryption== | ||
| + | For decrypting encrypted DVDs build and install libdvdcss from http://download.videolan.org/pub/libdvdcss/last | ||
| + | <pre> | ||
| + | ./configure --prefix=/usr | ||
| + | make | ||
| + | make install | ||
| + | </pre> | ||
==Space not an Issue?== | ==Space not an Issue?== | ||
| − | If hdd space is not an issue then grab the whole DVD | + | If hdd space is not an issue then grab the whole DVD<br/> |
| − | Read the DVD to a rescue image ignoring bad blocks | + | ===Installation=== |
| + | Install the following software | ||
| + | <pre> | ||
| + | apt-get install lsdvd gddrescue vobcopy genisoimage | ||
| + | </pre> | ||
| + | |||
| + | ===Extract=== | ||
| + | Run once to get the css key | ||
| + | <pre> | ||
| + | lsdvd | ||
| + | </pre> | ||
| + | Read the DVD to a rescue image ignoring bad blocks abd mount the image | ||
<pre> | <pre> | ||
| − | ddrescue /dev/<dvddevice> <title>. | + | ddrescue -n -b 2048 /dev/<dvddevice> <title>_ddrescue.iso |
| + | mkdir <title>.mnt | ||
| + | mount <title>_ddrescue.iso <title>.mnt | ||
</pre> | </pre> | ||
| + | Extract and unencrypt the video files. | ||
<pre> | <pre> | ||
| − | vobcopy | + | vobcopy <title>.mnt -m -t <title> |
</pre> | </pre> | ||
Then create an iso image from it | Then create an iso image from it | ||
| Line 21: | Line 42: | ||
Install the packages | Install the packages | ||
<pre> | <pre> | ||
| − | apt-get install libav-tools lsdvd mplayer oggz-tools | + | apt-get install libav-tools lsdvd mplayer oggz-tools dvdauthor tesseract-ocr |
</pre> | </pre> | ||
| − | + | ===Prepare=== | |
| + | Create an image of the disc to avoid disc read problems | ||
<pre> | <pre> | ||
| − | + | ddrescue -n -b 2048 /dev/<dvddevice> <title>_ddrescue.iso | |
| − | |||
| − | |||
</pre> | </pre> | ||
| − | |||
| − | |||
Use lsdvd to see what's on DVD. Determine the stream you would like to extract as well as the aid for audio and the sid for subtitles. | Use lsdvd to see what's on DVD. Determine the stream you would like to extract as well as the aid for audio and the sid for subtitles. | ||
<pre> | <pre> | ||
| − | lsdvd -x | + | lsdvd -x <title>_ddrescue.iso |
</pre> | </pre> | ||
Write the stream to the harddrive so the next steps will go faster. | Write the stream to the harddrive so the next steps will go faster. | ||
<pre> | <pre> | ||
| − | mplayer dvdnav://<stream> -dumpstream -dumpfile <title>.vob | + | mplayer dvdnav://<stream>/<title>_ddrescue.iso -dumpstream -dumpfile <title>.vob |
</pre> | </pre> | ||
Investigate the VOB file and note the numbers for the videostream and the audiostream | Investigate the VOB file and note the numbers for the videostream and the audiostream | ||
| Line 51: | Line 69: | ||
===Audio=== | ===Audio=== | ||
| + | Calculate the target bitrate | ||
| + | <pre> | ||
| + | bitrate=`avprobe -show_format <title>_audio.ac3 2>/dev/null | grep bit_rate | cut -d= -f 2`; samplerate=`avprobe -show_streams <title>_audio.ac3 2>/dev/null | grep sample_rate | cut -d= -f 2` ; echo "scale=10;${bitrate}/(${samplerate}/44100)/2/1000" | bc | cut -d. -f 1 | ||
| + | </pre> | ||
Now extract audio from the stream and convert to OGG. | Now extract audio from the stream and convert to OGG. | ||
<pre> | <pre> | ||
| − | avconv -y -i <title>_audio.ac3 -c:a libvorbis - | + | avconv -y -i <title>_audio.ac3 -c:a libvorbis -b:a <bitrate>k -ar 44100 -vn -sn -f ogg <title>.audio |
</pre> | </pre> | ||
| Line 61: | Line 83: | ||
avconv -y -i <title>_video.vob -t 600 -vf cropdetect -an -sn -f rawvideo /dev/null 2>&1 | tail | head -n 1 | sed 's/^.*crop=//' | avconv -y -i <title>_video.vob -t 600 -vf cropdetect -an -sn -f rawvideo /dev/null 2>&1 | tail | head -n 1 | sed 's/^.*crop=//' | ||
</pre> | </pre> | ||
| − | + | Calculate the target bitrate | |
<pre> | <pre> | ||
| − | avconv -y -i <title>_video.vob -r 25 -g | + | filesize=`find ./ -name <title>_video.vob -printf '%s\n'`; duration=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep duration | cut -d= -f 2 | sed 's/:/\//'`; sar=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep sample_aspect_ratio | cut -d= -f 2 | sed 's/:/\//'`; framerate=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep r_frame_rate | cut -d= -f 2 | sed 's/:/\//'`;echo "scale=10;((${filesize}/3)*8/${duration}/(${sar}))*(25/${framerate})/1000" | bc | cut -d. -f 1 |
| + | </pre> | ||
| + | Convert the video stream to ogv in 2 passes | ||
| + | <pre> | ||
| + | avconv -y -pass 1 -i <title>_video.vob -r 25 -g 100 -bf 16 -filter:v yadif,crop=<cropvalues>,scale=in_w:in_h/sar -b:v <bitrate>k -c:v libtheora -an -sn -f ogg <title>.video | ||
| + | avconv -y -pass 2 -i <title>_video.vob -r 25 -g 100 -bf 16 -filter:v yadif,crop=<cropvalues>,scale=in_w:in_h/sar -b:v <bitrate>k -c:v libtheora -an -sn -f ogg <title>.video | ||
</pre> | </pre> | ||
===Subtitles=== | ===Subtitles=== | ||
| + | Force a RGB palette file with the name <title>.rgb and the following contents | ||
| + | <pre> | ||
| + | ff0000 | ||
| + | ffff00 | ||
| + | ff00ff | ||
| + | 00ff00 | ||
| + | ff0000 | ||
| + | ffff00 | ||
| + | ff00ff | ||
| + | 00ff00 | ||
| + | ff0000 | ||
| + | ffff00 | ||
| + | ff00ff | ||
| + | 00ff00 | ||
| + | ff0000 | ||
| + | ffff00 | ||
| + | ff00ff | ||
| + | 00ff00 | ||
| + | </pre> | ||
| + | The 1st color is red, the 2nd color is yellow, the 3rd color is purple, the 4th color is green<br /> | ||
Extract the subtitles from the stream. | Extract the subtitles from the stream. | ||
<pre> | <pre> | ||
| − | + | spuunmux -s <sid> -p <title>.rgb <title>.vob | |
| + | </pre> | ||
| + | Using an image viewer determine the center color (no outlines) of the subtitle in the image and make this black text on a white background using: | ||
| + | <pre> | ||
| + | for file in *.png; do convert $file -fill '#000000' -opaque '#<palette>' -threshold 1% $file.pnm; tesseract -l <3 letter code language> $file.pnm $file; done | ||
| + | </pre> | ||
| + | Now combine the txt files and the sub.xml file and convert it to the SRT format | ||
| + | <pre> | ||
| + | linenr=1; cat sub.xml | sed '/subpictures/d' | sed '/stream/d' | cut -d\" -f 2-6 | tr ' ' ',' | tr -d '"' | sed 's/,[^=]*=/,/g' | sed 's/,\([^,]*\)$/ --> \1/' | while read line; do file=`echo $line | cut -d, -f 1`; times=`echo $line | cut -d, -f 2`; echo $linenr; echo $times | tr '.' ',' ; cat $file.txt; echo; linenr=$(($linenr+1)); done > <title>.srt | ||
</pre> | </pre> | ||
| Line 75: | Line 130: | ||
Combine everything in a ogg container | Combine everything in a ogg container | ||
<pre> | <pre> | ||
| − | oggz-merge -o <title>.ogv <title>.video <title>.audio | + | oggz-merge -o <title>.ogv <title>.video <title>.audio <title>.srt |
</pre> | </pre> | ||
Latest revision as of 13:51, 19 December 2012
Contents
Decryption
For decrypting encrypted DVDs build and install libdvdcss from http://download.videolan.org/pub/libdvdcss/last
./configure --prefix=/usr make make install
Space not an Issue?
If hdd space is not an issue then grab the whole DVD
Installation
Install the following software
apt-get install lsdvd gddrescue vobcopy genisoimage
Extract
Run once to get the css key
lsdvd
Read the DVD to a rescue image ignoring bad blocks abd mount the image
ddrescue -n -b 2048 /dev/<dvddevice> <title>_ddrescue.iso mkdir <title>.mnt mount <title>_ddrescue.iso <title>.mnt
Extract and unencrypt the video files.
vobcopy <title>.mnt -m -t <title>
Then create an iso image from it
genisoimage -dvd-video -o <title>.iso <title>
Make smaller files
This describes how to convert a DVD to OGG video using:
- libtheora for video
- libvorbis for audio
- srt subtitles
Installation
Install the packages
apt-get install libav-tools lsdvd mplayer oggz-tools dvdauthor tesseract-ocr
Prepare
Create an image of the disc to avoid disc read problems
ddrescue -n -b 2048 /dev/<dvddevice> <title>_ddrescue.iso
Use lsdvd to see what's on DVD. Determine the stream you would like to extract as well as the aid for audio and the sid for subtitles.
lsdvd -x <title>_ddrescue.iso
Write the stream to the harddrive so the next steps will go faster.
mplayer dvdnav://<stream>/<title>_ddrescue.iso -dumpstream -dumpfile <title>.vob
Investigate the VOB file and note the numbers for the videostream and the audiostream
avprobe <title>.vob
Some VOB files report the wrong duration so just in case rebuild the vob file but only with the required videostream and audiostream.
avconv -y -i <title>.vob -map 0:<videostream> -c:v copy -an -sn -f vob <title>_video.vob avconv -y -i <title>.vob -map 0:<audiostream> -c:a copy -vn -sn -f ac3 <title>_audio.ac3
Audio
Calculate the target bitrate
bitrate=`avprobe -show_format <title>_audio.ac3 2>/dev/null | grep bit_rate | cut -d= -f 2`; samplerate=`avprobe -show_streams <title>_audio.ac3 2>/dev/null | grep sample_rate | cut -d= -f 2` ; echo "scale=10;${bitrate}/(${samplerate}/44100)/2/1000" | bc | cut -d. -f 1
Now extract audio from the stream and convert to OGG.
avconv -y -i <title>_audio.ac3 -c:a libvorbis -b:a <bitrate>k -ar 44100 -vn -sn -f ogg <title>.audio
Video
Detect the amount to crop.
avconv -y -i <title>_video.vob -t 600 -vf cropdetect -an -sn -f rawvideo /dev/null 2>&1 | tail | head -n 1 | sed 's/^.*crop=//'
Calculate the target bitrate
filesize=`find ./ -name <title>_video.vob -printf '%s\n'`; duration=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep duration | cut -d= -f 2 | sed 's/:/\//'`; sar=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep sample_aspect_ratio | cut -d= -f 2 | sed 's/:/\//'`; framerate=`avprobe -show_streams <title>_video.vob 2>/dev/null | grep r_frame_rate | cut -d= -f 2 | sed 's/:/\//'`;echo "scale=10;((${filesize}/3)*8/${duration}/(${sar}))*(25/${framerate})/1000" | bc | cut -d. -f 1
Convert the video stream to ogv in 2 passes
avconv -y -pass 1 -i <title>_video.vob -r 25 -g 100 -bf 16 -filter:v yadif,crop=<cropvalues>,scale=in_w:in_h/sar -b:v <bitrate>k -c:v libtheora -an -sn -f ogg <title>.video avconv -y -pass 2 -i <title>_video.vob -r 25 -g 100 -bf 16 -filter:v yadif,crop=<cropvalues>,scale=in_w:in_h/sar -b:v <bitrate>k -c:v libtheora -an -sn -f ogg <title>.video
Subtitles
Force a RGB palette file with the name <title>.rgb and the following contents
ff0000 ffff00 ff00ff 00ff00 ff0000 ffff00 ff00ff 00ff00 ff0000 ffff00 ff00ff 00ff00 ff0000 ffff00 ff00ff 00ff00
The 1st color is red, the 2nd color is yellow, the 3rd color is purple, the 4th color is green
Extract the subtitles from the stream.
spuunmux -s <sid> -p <title>.rgb <title>.vob
Using an image viewer determine the center color (no outlines) of the subtitle in the image and make this black text on a white background using:
for file in *.png; do convert $file -fill '#000000' -opaque '#<palette>' -threshold 1% $file.pnm; tesseract -l <3 letter code language> $file.pnm $file; done
Now combine the txt files and the sub.xml file and convert it to the SRT format
linenr=1; cat sub.xml | sed '/subpictures/d' | sed '/stream/d' | cut -d\" -f 2-6 | tr ' ' ',' | tr -d '"' | sed 's/,[^=]*=/,/g' | sed 's/,\([^,]*\)$/ --> \1/' | while read line; do file=`echo $line | cut -d, -f 1`; times=`echo $line | cut -d, -f 2`; echo $linenr; echo $times | tr '.' ',' ; cat $file.txt; echo; linenr=$(($linenr+1)); done > <title>.srt
Merging
Combine everything in a ogg container
oggz-merge -o <title>.ogv <title>.video <title>.audio <title>.srt
