|
|
|
|
|
|
|
CODEC
Pulse-code modulation
PCM is a general expression for digitizing an analogue signal. A PCM
signal is in fact the numerical or digital representation of the analogue
signal. In most cases, the analogue signal is sampled by means of an
Analog-to-Digital Converter (ADC), which is represented by this symbol:
Likewise, the signal can be converted back from the digital to the analogue
domain with a Digital-to-Analog Converter (DAC). A DAC is usually
represented by this symbol:
The quality of the recorded data
is determined by the sample rate (T) and the bit-depth — i.e. the total number
of possible different values for a single sample. For example: when using
8 bits, a total of 28 = 256 different values are available.
A sample is always rounded off to the nearest bit value, which introduces an
error. This error is responsible for the quantisation noise. Furthermore,
a low sampling rate causes a less accurate representation of the original
signal, which introduces aliasing. Both effects are responsible for the amount
of distortion of the signal.
Generally speaking: the higher the bit-depth and the sampling rate, the smaller
the error and the lower the distortion. The diagram below shows an arbitrary
sound wave that is sampled at 'T' intervals. Move the mouse over the diagram
above to see the resulting digital approximation.
Sending PCM data typically requires (at least) twice the bandwidth of the
analogue original, but the quality is unsurpassed. It is suitable for speech
as well as high fidelity music.
PCM is very suitable for high-quality sound transmission, as long as
sufficient bandwidth is available.
➤ PCM on Wikipedia
|
In practice, PCM is often combined with a compression algorithm
to lower the amount of data and, hence, the bandwidth. This allows an
originally wideband signal to be send over a narrowband channel. At the
receiving end, the signal must be decompressed before it can be used.
Most compression techniques not only reduce the amount of data, but also
the quality of the sound. Such a technique is known as a lossy compression.
In contrast, a lossless compression technique preserves the full fidelity
of the orignal signal, but at the cost of a lower compression ratio.
In military communications lossy compression is commonly used,
such as LPC-10 or CELP.
If the transmission path is cryptographically secured, the data will usually
be encrypted after compression. At the receiving end, it must then be
decrypted before it can be decompressed.
|
In its simplest implementation, the quantization levels are linearly uniform,
which means that the possible values are evently spread over the Y-axis.
This is known as Linear PCM, or LPCM, although the term PCM is also used.
High-fidelity audio systems (i.e. systems with a high bit-depth, e.g. 16 or
24 bits) generally use
LPCM, as it covers the full dynamic range of the signal.
With low bit-rates the number of possible values is limited, as a result of
which the audio signal cannot be accurately described. For example: with
8 bits, only 256 discrete values are possible (128 for negative amplitudes
and 128 for positive amplitudes).
This is particularly the case with speech, of which the wide dynamic range does
not lend itself well to efficient linear digital encoding.
In such cases it is possible to distribute the values as a function of the
amplitude, like this:
Examples of non-linear distribution algorithms
are A-law and µ-law.
They effectively reduce the
dynamic range of the signal, but increase the coding efficiency, resulting
in a signal-to-distortion ratio that is superior to that obtained by
linear PCM [4].
|
µ-law, also written as µ-law or u-law or PCMU, is a standard companding
algorithm, used in American and Japanese 8-bit PCM systems.
It optimises (i.e. modifies) the dynamic range of an analog
signal for digitizing. Together with A-law
it is part of the G.711 standard of the ITU-T [3].
➤ µ-law on Wikipedia
|
A-law, also known as PCMA,
is a standard companding algorithm, used in European 8-bit PCM
systems. Like µ-law It optimises (i.e. modifies) the dynamic range of
an analog signal for digitizing. Together with µ-law it is part of
the G.711 standard of the ITU-T [4].
➤ A-law on Wikipedia
Of these two companding algorithms, µ-law provides a slightly larger
dynamic range than A-law, at the cost of worse proportional distorition
for low-amplitude signals. In international telephony, A-law is used by
convention for international connections, if one of the parties uses it.
|
|
|
|
Any links shown in red are currently unavailable.
If you like the information on this website, why not make a donation?
© Crypto Museum. Created: Wednesday 08 May 2024. Last changed: Wednesday, 08 May 2024 - 13:12 CET.
|
 |
|
|
|
