1140 Chapter 30
depending on where the pickup is located on the disc
surface—faster on inner diameters, and slower on outer
diameters. This is accomplished with yet another servo
loop; information from the data stream recovered by the
laser pickup is used to determine correct rotating speed,
and the spindle motor is regulated accordingly.
30.3.2 Data Decoding
The photodiode array and its processing circuits pro-
duce a signal resembling a series of high-frequency
sinusoids called the EFM signal. A collection of EFM
waveforms (called an eye pattern) is shown in Fig.
30-9. The digital data can be recovered from the EFM
signal if it can be determined when the signal crosses
the zero axis, relative to the timing constraints created
by the EFM encoding rules.
CD data decoding follows a procedure that essen-
tially duplicates, in reverse order, the encoding process.
The first data to be extracted from the signal is synchro-
nization words. This information is used to synchronize
the thirty three symbols of channel information in each
frame, and a synchronization pulse is generated to aid in
locating the zero crossing of the EFM pattern and to
generate a transition at those points to produce a binary
signal.
The EFM signal is demodulated so that every 17-bit
EFM word is reconverted to 8-bits. Demodulation can
be accomplished by logic circuitry or a look-up table. A
buffer is used to remove the effect of disc rotational
irregularities; data input to the buffer may be irregular
in time but clocking ensures that the buffer output is
precise. To guarantee that the buffer neither overflows
nor underflows, a correction signal is generated and
used to control the disc rotating speed.
Following demodulation, data is sent to a CIRC
decoder for deinterleaving, error detection, and correc-
tion. The CIRC decoding process reverses the
processing steps accomplished during encoding. The
CIRC decoder accepts one frame of thirty-two 8-bit
symbols; twenty four are audio symbols, and eight are
parity symbols. One frame of twenty-four 8-bit symbols
is output. The decoder utilizes parity from two
Reed-Solomon decoders and deinterleaving. The first
error correction decoder is designed to correct random
errors, and to detect burst errors. It flags all burst errors,
to alert the second error correction decoder.
Error concealment algorithms, employing interpola-
tion and muting circuits, follow the CIRC decoder.
Uncorrected words are detected through flags and dealt
with, while valid data passes through unprocessed.
Using error flags, the player’s signal-processing circuits
determine whether to output the data directly, to inter-
polate it, or to mute the sound.
For continuous errors, muting is employed as a last
resort; invalid data passed on to the D/A converter
could result in an audible click. Muting is accom-
plished by beginning attenuation many samples before
the invalid data, smoothly muting the invalid data, and
then smoothly restoring the signal level. This method of
muting is often largely inaudible.30.3.3 Signal ReconstructionAt the output stage, the digital data is converted to a ste-
reo analog audio signal. This reconstruction requires
low-pass filtering to suppress high-frequency image
components and D/A conversion. An oversampling dig-
ital filter uses samples from the disc as input and then
computes interpolation samples, digitally implementing
the response of a low-pass filter. A transversal filter can
be used to oversample (perhaps at an eight-times rate);
image components appear at multiples of the new sam-
pling rate. Because the separation between the baseband
and sidebands is greater, a low-order analog filter can be
used to remove the images. The type of oversampling
filter found in CD players is an example of a wider class
of FIR (finite impulse response) digital filters used in
many applications. These kinds of filters use addition,
multiplication, and delay elements to perform their
tasks, and fall under a wider category of technology
known as DSP (digital signal processing). The transver-
sal filter used in CD players resamples and filters
through interpolation. Resampling acts to increase the
sampling rate; for example, in an eight-times oversam-
pling filter, seven zero values are inserted for every data
value output from the disc. This increases the sampling
rate from 44.1 kHz to 352.8 kHz.
Interpolation is used to generate the values of inter-
mediate sample points—for example, seven intermediate
samples for each original sample. These samples areFigure 30-9. EFM eye pattern.
3T 4T 5T 6T 7T 8T 9T 10T 11T