<<<Listening A1 //---------\\ Recording Basics>>>
Due to the equipment we are using in this class, our sounds will be recorded
in a digital format. In fact, while the sound is within the computer and
MBox
it is represented
digitally.
 |
Digital data is nothing more than a stream of 1s and 0s. In terms of circuits: high voltages and low voltages. For now, take my word that sound requires a lot of bits…thousands, in fact, for just a second of sound. |
In an analog system—like a tape deck—sound is represented and stored
as a “drawing” of the sound’s amplitude plotted over time.
That way, the speaker can reproduce the patterns of sound waves based of this
continuous record of each sound. The screenshot below is a representation of
what you would see if you looked very closely at analog tape. On the right
is a tape deck with reels. Click here for more pictures of analog technology.
 |
 |
In digital systems, the smooth waves of sound pressure must be broken down into numbers; they must be quantized. Amplitudes are recorded at regular intervals. These regular intervals are very small (on the order of thousands of times per second), but more on the reasons for this later. Since the memory in a computer is finite, we can only record a finite amount of samples from the sound. Below, we see first a sawtooth wave represented as bars. Each bar represents an amplitude value that can be written to a memory location in the computer. On the right we see a sine wave quantized into higher and higher resolutions. If the concern is quality, higher resolution is better (to a point).
 |
 |
Nyquist Limit: A very smart mathematician named Harry
Nyquist came up with
a theory to tell what frequencies can be represented by a certain sampling
rate.
It
turns
out
that
the sampling rate must be 2 times the highest frequency we wish to represent.
As we talked about when we discussed acoustics, the average person can hear
frequencies up to about 20,000 Hz. Therefore, audio engineers and programmers
settled on
44,100 Hz as an appropriate rate to sample digital audio. Other sampling rates
have been used. 8kHz (8,000 Hz) is the sampling rate for basic digital telephone
service. This is because human speech is mainly in a range below 4,000 Hz.
Many digital recording systems and software now can be set at 48,000 Hz sampling
rate,
and this is the accepted rate for high-quality audio applications. This increases
the limit on frequencies. Some systems sample at 4 times this rate, 192,000
Hz. These systems are mainly used to justify charging loads of money for a
HD DVD
system.
There is one thing that digital completely eliminates: tape hiss. We probably do not appreciate this aspect of our technology. In the olden times, all sound was recorded on large tape decks. The tape was drawn from one spool to another. In a perfect world, the tape would be drawn across the magnetic heads and rollers with out creating any friction. In the real world, there is friction between all the moving parts, and this friction creates noise in the recording. We call this noise tape hiss. Hiss is heard in all cheap consumer-grade tape decks, and to a lesser extent in professional decks. Even with the best, most expensive gear, a tiny amount of hiss exists. By mixing two or more taped sounds together (or by amplifying a taped sound) we combine (or amplify) the hiss from each recording. This can build up over the process of putting together a large piece. In the early years, this was unavoidable, but as the recording industry grew and the tapes became more valuable as commodities, the engineers found ways to clean up the recordings.
<<<Listening A1 //---------\\ Recording Basics>>>