[clug] Audio file formats

[Paul Wayper]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Kristy A. Bennett wrote:
| Hi all!
|
| I am sorry I haven't yet transitioned to the LinuxSA crowd!
|
| I have some 'work' to do that will involve audio recordings and need to
| buy a clue or ten as this is an area where I have no experience whatsoever.
|
| With the aid of Sound Recorder I can record in four formats: ogg, wav,
| spx and flac.  I understand the difference in quality but have no idea
| of cross-platform, cross-media-player-device implications of these
| formats.  Can someone clue me in?

There are very few portable media players that will play ogg, flac or spx: the
iRiver and the Rio Karma are the only two that I know of.  You can get plugins
for iTunes and Windows Media to play ogg, flac and spx but if you're wanting
'out of the box' playback then you may want to stick to WAV.

However, taking a wider view may mean you have more options.  Your best bet is
to record in raw WAV format at the highest possible sampling rate (44100 or
48000 Hz are typical values but as Mike says 96000Hz is used for audio
preservation) and bit rate (16 bits or 24 bits for audio preservation) that
the device supports.  Then import the files and use a tool such as Audacity to
do whatever processing you need to make the recording most listenable - remove
extraneous noise, compress (in the analogue audio sense - this means compress
the dynamic range so that everything sounds about the same loudness),
normalise, and so on.  I can give you much more information on this if you
want it.

If your objective is to have a small sound that any player can play, you may
then be able to resample the sound down to a lower sample rate and number of
bits.  For instance, you speech find that speech is still perfectly
intelligible at 8KHz sample rate and 8 bits monophonic - this is, after all,
what our phone system uses.  You can also use such encodings as U-law, A-law,
and ADPCM, which use different ways of expressing the numbers to give more
dynamic range in less bits.  Most of these are commonly understood by anything
that can play a 'WAV' file, since the WAV format does allow for specifying
these types of encoding.  Careful compression can reduce a WAV file to easily
an eighth of its 'CD-quality' uncompressed size without losing too much
intelligibility.  The 'sndfile-convert' program is what you want here.

Alternately, most devices support MP3, and if you're willing to stick your
tongue out at Thompson Multimedia and the other owners of the MP3 patents,
then LAME will give you easily twelve to twenty times compression for voice
depending on what settings you choose and will be playable on most devices.

So can you tell us what you're doing this for and what your end user will have
to play the file back?  The more I understand what you're trying to do the
better my advice will be :-)

Have fun,

Paul
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFIMii3u7W0U8VsXYIRAhI1AJ9pgZsleWYCKVDVenUuYDYhqdM46ACgpPCM
9A+mBX3S7Tez+f2Nb05ybCA=
=nhq3
-----END PGP SIGNATURE-----