Originally printed in the March 1983 issue of Gramophone.

With an apology to those readers who have been assiduously reading all about the Compact Disc already, in these pages and elsewhere, I propose to start right at the beginning, with a physical description. The main features are shown in Table 1, with the equivalent values for an LP record listed alongside for comparison. It will be seen that the word 'compact' really does apply to the new disc (apart from its clever allusion to that other Philips-invented music-carrier, the Compact Cassette). The Compact Disc is only 120mm (4.7 inches) in diameter, thus having only about one-sixth the area of an LP, and is a mere 1.2mm thick. This compactness has obvious advantages in reducing storage space for the discs and permitting the design of small player units (possibly for portable and in-car use in the future).

The sound quality potential of CD is markedly greater than the LP, particularly in such parameters as signal-to-noise, channel separation and wow-and-flutter. Such improvements can be attributed to the use of digital techniques and the replacement of a mechanical scanner (stylus) with a laser optical beam—which introduces the further inestimable benefit of eliminating record wear and resisting the effects of dust, scratches and static charges.

While the digital process is of mind-boggling complexity, an idea of its basic sequence can be gathered. First we have the waveform of a typical music signal - the electrical waveform is a direct imitation or 'analogue' of the alternating swings in air pressure as the original soundwaves reached the microphone. For the purposes of digital recording, this analogue signal has first to be converted to digital form. This involves two processes: sampling and quantization. Sampling of the waveform amplitude takes place at frequent fixed time intervals as dictated by a quartz oscillator. Quantization is the process of encoding and storing as a binary number (consisting of a stream of '0' or '1' digits) each of the sampled values on a fixed amplitude scale. So sampled values in volts are then expressed in 3-bit binary numbers (000=0, 001 = I, etc.) and stored as the on/off pulse train signal. To replay a digital recording, this signal is passed through a digital-to-analogue converter which produces a stepped waveform requiring only a low-pass filter to smooth out the steps and recreate the original analogue waveform.

Even in this simplified account it will be obvious that a number of errors and approximations take place. Sampling a waveform at intervals can never be quite the same as tracing it continuously. However, the sampling frequency used for the Compact Disc is 44.1kHz which has been found adequate for resolution of audio waveforms up to 20kHz. Again, quantization errors will occur each time that a sample does not have precisely one of the stepped values encodable on the chosen binary scale—giving rise to 'quantization noise'. In the CD system, 16-bit encoding is used which gives a scale of 65,536 different values corresponding t…o a signal-to-noise ratio of 96dB, and the remaining relatively small errors are made less audible by mixing in a white noise signal called 'dither'.

The PCM (Pulse Code Modulation) train is etched on the CD disc surface as a series of pits (or bumps, depending on the way you look at it) along a track which begins at the inner radius and spirals out towards the periphery. Individual pits are only about 0.5µm wide and 0.2µm deep, with a pitch (track width) of 1.6µm —about one thirtieth of the thickness of a human hair. The scanning velocity is fixed at about I.2m/s (covering an amazing 4.3 million bits per second) and so the rotational speed (unlike an LP player) must begin fast at the centre and progressively slow down towards the outside edge (ranging from approximately 500rpm down to 200rpm). Sophisticated control signals interpolated within the PCM stream instruct the drive motor as to the correct speed at each point on the disc.

Player scanning mechanism 

The Compact Disc is read by a semiconductor laser light-beam source. (LASER stands for Light Amplification by Stimulated Emission of Radiation.) This type of light-beam has the advantages of being monochromatic (at a single frequency), phase-coherent and strongly directional (due to its very short wavelength, around 0.78 µm)—and, at the low power level used here, perfectly safe. The beam passes upwards through a semi-reflecting prism and on to a lens system which focuses it sharply—with an incident beam width of only 1.87µm —on to the underside of the disc. As the recorded pulse-train of alternating pits and flat portions is scanned, the laser beam will switch between conditions of being scattered and strongly reflected back down its original path, whence it is redirected by the prism on to a light-sensitive diode. The latter therefore generates an electrical signal recreating the original bit-stream of '0’ and '1' digits.

Apart from the servo-motor already mentioned for maintaining the correct scanning velocity at all points on the disc, there is a focus servo to control the lens system for accurate focusing, a tracking servo which checks for centre-scanning or equal spill of the laser beam on to the tracks on either side of the track being scanned, and a traversing servo for moving the optical system between the inside and outside of the disc. Yet the technical cleverness of CD does not end here.

Cueing and displays

Different sections of the disc: the 'program area' encompasses up to 20,000 tracks and the 'lead in' area contains what is called the 'Table of Contents'. The latter carries in code all such details as the total playing time of the disc, the number of music items or movements, the start time of each item, etc. There is also spare capacity for a much wider range of visual display information not yet implemented. Even so, the data included on every frame of the recorded tracks can give a continuous read-out of track elapsed time (lap time), track number, total time remaining, versatile fore-and-aft cueing, varispeed search (with sound monitor) and pause. At the same time, complex error correction techniques can eliminate the effects of dust or scratches—whose importance is already substantially diminished by the sharp focusing of the laser beam through the clear plastic base on to the recorded surface. Error bursts of up to 3,500 bits, or a drop-out over 2mm long, can be played without audible effect.

To sum up, the Compact Disc format offers high-quality sound which is 'robust' in its ability to withstand mishandling and has unprecedented potential for cueing, programming and displaying data relating to the musical contents.

In terms of playing time, Compact Discs are initially appearing with durations of up to 60 minutes maximum. Potentially, however, the industry sees the CD format as being flexible enough for eventual production as 'singles', with durations down to, say, 5 minutes, and for an extension up to a maximum of 80 minutes to accommodate longer musical works.

How Compact Discs are made

There are clear parallels between the manufacturing procedures for Compact Disc and the conventional LP record. Both involve the cutting or etching of a master disc which is first silvered, to make it electrically conducting, then electroplated, the metal deposit producing a matrix which—after one or more further stages of plating—produces a mould or stamper to press out the discs. There are differences, of course, mainly due to the very tiny dimensions of the CD tracks. These necessitate a greater degree of accuracy at each stage, a clean-air working environment and the application of a protective layer over the recorded tracks on the finished disc.

The stages in CD manufacture are as follows:

1. A 240mm diameter glass plate, optically ground, cleaned and polished is given a coating of photoresist about 0.1 µm thick (somewhat like the light-sensitive coating on a photographic film).

2. A high-power laser, modulated by the signals from the PCM digital tape master, inscribes the desired spiralling pattern of pits and spaces in the photoresist.

3. In a ‘development’ stage, the exposed pit areas are etched away to leave the final recorded surface structure.

4. The disc is silvered, by an evaporation process and can then enter the manufacturing area.

5. An electroplating process deposits nickel on the disc, which can be stripped off to form a negative or 'master'.

6. By a secondary plating process, up to ten intermediary metal positives or ‘masters’ can be produced.

 7. Each master can be used, in yet a further plating stage, to produce up to 100 negative ‘stampers’. 

8. Using the stamper as a mould, compression or injection-moulding can be used to mass-produce discs of clear PVC.

9. The surface carrying the pattern of pits and spaces is given a micro-thin reflecting layer of aluminium. 

10. This mirror-like layer is then protected by a further layer of lacquer, on top of which the label can be applied or printed directly. 

11. The disc is precision-centred and the central hole is punched out.

John Borwick