Dear all,
Please find the audio of the talk at http://nikegravel.com/seminar
or use this direct link http://nikegravel.com/seminar/audio.html
I have attached a readme for your reference.
Kudos to Senthil Kumar of NMSworks for recording.
Some sessions are missing. We were unable to process the files.
Sorry about that!
Should you have any questions please mail
koushik.narayanan@xxxxxxxxx or girish1729@xxxxxxxxx
Best,
Girish
This is a document that explains what we did to get the audio to you.
As always there is a great level of technical detail for those who
might want to do such a thing in future.
The original sound files recorded by Senthil Kumar of NMSWorks,
IIT using his mobile were in AMR (Adaptive MultiRate) format.
It is easier to convert them to uncompressed PCM (WAV) and then do
the processing or further conversion.
To convert AMR to WAV there are 3 possible ways (AFAIK)
1.) Use ffmpeg with amr_nb support
2.) Use mplayer with amr support
3.) Use the decoder provided by 3gpp.
ffmpeg segfaults when decoding files with greater than a minute of
audio. The decoder provided by 3gpp is difficult to use. So the choice
is mplayer. The command will be:
mplayer -ao pcm:fast:file=filename.wav filename.amr
To convert all the files at once, the shell can be used:
#!/bin/sh
# Convert all amr files to wav
mkdir wav
for i in *.amr
do
mplayer -ao pcm:fast:file=`basename $i .amr`.wav $i
done
Now that we have the wav files sophisticated post processing can be done.
Here increasing volume is of prime importance as the audibility of
most files are extremely low. Increasing volume
can be done with sox (-V) or with normalize. A batch normalization of
all files is done using the following command:
normalize -a 0.5 wav/*.wav
Then individual files are tested for audiblity. If required further
boosting is done using
normalize -a 1 wav/cracking-tools.wav
Once the processed audio is obtained, encoding to compressing formats
can be done. AAC,Ogg Vorbis and MP3 are the formats of choice. The
following script does the encoding with tagging.
To get an appropriate title tag for each file, each file was named
appropriately so that it can be extracted and used.
encode.sh:
#!/bin/bash
mkdir mp4-cover-art
mkdir mp4
mkdir ogg
mkdir mp3
album="Seminar on TCP/IP Networking,Cryptography and Spam Control"
artist="Girish Venkatachalam"
year=2007
comment="http://nikegravel.com/seminar/";
for i in *.wav
do
name=`basename $i .wav`
# mp4 without cover art
faac -o mp4/$name.mp4 --title "$name" --year "$year" \
--artist "$artist" --comment "$comment" --album "$album" $i
# mp4 with cover art
faac -o mp4-cover-art/$name.mp4 --title "$name" --year "$year"\
--artist "$artist" --comment "$comment" --album "$album" \
--cover-art seminarlogo.png $i
# ogg
oggenc -o ogg/$name.ogg -t "$name" -a "$artist" -c "$comment" \
-d "$year" -l "$album" $i
# mp3
lame --tt "$name" --ta "$artist" --ty "$year" --tc "$comment" \
--tl "$album" $i mp3/$name.mp3
done
Note that this also takes care of adding the appropriate ID tags.