Digital Recording

Recording, particularly of music and video images, consumes massive amounts of digital memory. High-density discs are ideal for these types of recording. Direct digital recordings can play back recorded or modulated sound. Tape recording, though convenient and easy, could not store digital data until the development of the DAT tape. Carrier signals in digital recording are always pulse waves that alternate between voltages (analog signals). Consequently, the modulation method in most digital systems for music is pulse code modulation (PCM). On CDs and other disc formats read by laser, the physical structure uses islands or raised points and pits or low points as the zeros and ones; in pulse modulation, the high ends of the pulses are easily represented as the islands, and the low pulses are the pits. Pulse code modulation is actually an old development in the history of recordings. It was developed in 1939 by A. H. Reeves, but it took technology many years to find practical uses for Reeves' invention.

The conversion of sampling frequency from analog to digital is critical to sound recording. The frequency is measured or sampled many times per second and then averaged to produce the single piece of data for the digital input. Increasing the sampling frequency improves the sound quality but decreases the storage economy especially on tapes. Resolution is another important specification and describes the number of bits used to represent the amplitude of an instant on the recording. Each bit doubles the possibility for representing instantaneous amplitude levels. Typically, 14-bit resolution is used to give a range of 16,384 possibilities for representing instantaneous amplitude values. Quantization is the process that converts the collection of values into the amplitude the listener hears.

Recording media are all imperfect, thanks to specks of dust or other contamination that prevents equipment from imprinting the data on the medium. Data on CD and DAT tape are even more tightly compressed than those on analog tapes, so loss of data is effectively magnified. To fix this, special error correction codes are built into the data stream to weave the sample values throughout the data. Some of these error correction codes can be very complex, and, of course, they also consume valuable storage space on the CD or tape. The recording engineer must compromise the number of error codes to make enough storage for the sound data.

Analog systems also have the disadvantage that, when a recording is played back and rerecorded, distortion is increased by about 0.5 %. Each subsequent copy will be worse. Analog discs and tape are also nonlinear and do not record all sounds equally, leading to inaccurate reproduction. In a digital recording system, this distortion does not occur. The master recording may have minimal quantization errors, but these do not compound when copies are made. In this case, the absolute zero-orone character of the digital world works to an advantage because the copy is equally absolute unless the digital recording is reconverted to an analog signal. Thousands of copies can be made from a digital master without distortion; similarly, digital media on CDs can be played back thousands of times without distortion.