Audio Primer
|
|
Sound in General |
Sound is pressure waves that propagate through the air (or
other medium - but not a vacuum!!) by slightly moving molecules in a towards-away
motion. This air motion causes the same type of motion in our eardrums,
which we perceive as sound. To accurately reproduce sound, the
pressure waves must be accurately received, converted to something a bit
easier to record than pressure waves, and put on a medium that has excellent
retention properties so the recording can be played back at a later time
with full fidelity. Playing sound from the recording involves reading
the medium used in the original recording and converting the information back
to pressure waves directed at our ears. If done perfectly, we won't
be able to tell the difference between the original signal and the recorded
signal. Since we have only two ears, it would seem logical that we only need to record two channels of sound - the signal that would originally have reached our two ears - to provide a perfect reproduction. The field of psycho-acoustics revolves around explaining why it isn't that simple, and even the best recording can still be identified as a recording. This field also attempts to explain the other end of the spectrum - why an absolutely horrid recording can still be accurately interpreted. |
Analog Recording |
Since audio is an analog signal (continuous - no discrete counting),
our first attempts at recording were analog. The phonograph, reel-to-reel
tape, cassette and 8-track are all analog devices. A variable
amount of surface shape (for the phonograph) or magnetic flux (for tape)
is a representation of the original signal. Unfortunately, all of
these wear out with age, and the signal loses some of its fidelity for
each generation of recording. |
Digital Recording |
We have very nearly mastered the art of putting a representation
of an audio signal onto a medium that will remain accurate for many years.
By converting the analog signal into a series of counts that represent
the magnitude of the analog signal, we can use technology developed for
computers to store the audio signal. The CD has virtually replaced
all other mediums for storing audio, though there are those that argue that
the fidelity is not as good as a clean analog recording. |
Compression |
Once we have a digital representation of an audio signal, we
can make use of another computer- inspired technique - compression. Compression
simply reduces the amount of data needed to store an audio signal with
some loss in fidelity. There is a trade between compression and fidelity.
Two popular compression techniques, especially when bandwith is limited,
are MP3 and Real Audio. |
MP3 |
MP3 compression is popular since it is tunable (frequency,
bitrate and number of channels are selectable) and provides a good compression
ratio. The most popular compression attributes for web applications
is 44kHz bandwidth, 128 kilobits/second, two channels. This
provides the full audio spectrum with near-CD quality in stereo with approximately
a 10:1 compression ratio. |
Real Audio |
Real audio compression has been the answer to streaming
audio (playing in real time, like a radio station) for quite a while. The
sound quality hass improved considerably, with 56K modem speeds now able
to play very respectable audio. The biggest advantage of Real Audio
is that the beginning of a large file will start playing while the rest of
the file is loading. The appearance is that the music is playing in
real time - just don't try to jump ahead - the file will have to load up
the that point before it will begin playing again. |