ROM-ANCING THE CLONE
As the name implies, Audio is a magazine about audio. Thus, theoretically, the articles should deal mainly with audio topics. But in the case of digital technology, the distinction between audio and data is increasingly blurred. This month, and next, let's consider the growing CD family. While you weren't looking, the music CD has been cloned into two new media, CD-ROM for computer in formation and CD-I for interactive, multi-media uses. Both carry that silvery surface a long way from Mozart or Manhattan Transfer.
Rather than storing music, the CD format can be treated as a read-only memory system for any kind of pro gram material, and CD-ROM (Com pact Disc Read-Only Memory) is the logical extension of digital audio CDs into that much broader application of general information storage. It is in tended mainly as a medium for data distribution and mass storage for computer-related applications.
To clarify: The format with which you are familiar, CD-Audio, is a specific application of Compact Disc technology; it is a standard for recording PCM audio. The CD-ROM standard is de rived from the CD-Audio standard, but defines a format for general data storage. Unlike CD-Audio, CD-ROM is not tied to any specific application. Both use optical discs, of the same diameter (12 cm, or 4.7 inches) but with different data formats.
Apart from modulation and error-correction overhead, a CD-ROM holds more than 650 megabytes (1 byte = 8 bits) of user information. This is a large storage area, equivalent to 1,500 half-megabyte floppy disks, 275,000 pages of alpha-numerics, 18,000 pieces of computer graphics, or 3,600 still video pictures. A CD-ROM can efficiently store information such as computer application software, operating systems, data bases, published reference materials, directories, back issues of journals, encyclopedias, libraries of still pictures, parts catalogs, or other types of information not requiring frequent updating. For example, how about a catalog of every music recording ever made, and every record re view ever written? Easy-it would occupy a disc or two.
The amount of data stored on a CD ROM is, of course, equal to the amount of information stored on a CD-Audio disc. But a ROM makes you more aware of just how information-hungry a music recording is. If you were to download all the information from one fully loaded CD-ROM at a rate of 300 baud (transmitting it to a friend, for example, using a modem and a telephone line) for 24 hours a day, it would take 184 days to transmit the contents.
(Hopefully, it wouldn't require a long distance call!) Think of CD-ROM as electronic paper (arid a lot of it); anything publish able is a candidate for CD-ROM.
However, a CD-ROM is much more efficient than paper. For example, the text of the Grolier Academic Encyclopedia is available on a CD-ROM, and it occupies only about 60 megabytes, or one-tenth of the disc's storage space.
A comprehensive index, occupying another 60 megabytes, greatly facilitates data access. Enter a subject, and a listing of entries comes up on the screen. These entries are far more comprehensive than those a conventional encyclopedia because the system's catalog can compare your re quest with the contents of the entire encyclopedia. A phrase like "freedom of speech" might have only a few citations in the index of a conventional encyclopedia, but in a CD-ROM sys tem, every mention of that phrase in the entire encyclopedia would be displayed. Then you could scroll to any entry and read the full text. The words you searched for would be highlighted in the text.
Like an audio CD player, the CD ROM player contains laser optics, a disc drive, and demodulation and error-correction circuits. However, in stead of an audio output section, it has a computer interface. (Combination players, with both audio and computer outputs, could also be made.) Although a CD-ROM disc looks exactly like a music CD, it identifies itself (through the Q subcode channel) as differing from an audio CD, and it em ploys a modified data format. Data in a music CD is derived from the 44.1-kHz sampled signal, with 16-bit quantization. The 16 bits are divided into higher and lower 8-bit bytes. These data bytes (before EFM modulation) are grouped into frames of 24 bytes, and parity bits and a synchronization word are added. While satisfactory for music applications, a 24-byte frame is inconveniently short for massive transfers of computer data or programs. Further more, the audio CD makes no provision for locating any specific frame by its address.
The solution to both problems was to design the CD-ROM system around blocks of 98 frames. Addressing information is handled by treating the sub-code information for each block as eight 98-bit subcode words instead of the 98 eight-bit words (bytes) used in the audio format. The effective data block becomes 2,352 bytes (24 bytes x 98 frames). This is sufficiently long to handle data in units of 2,048 (2K) bytes-convenient for computer use-- without totally discarding the audio CD format.
Each CD-ROM disc is divided into 330,000 blocks, and each block is di vided into four fields. The first 12 bytes from each 2,352-byte block are used as a synchronization word. The next four bytes form a header field used for time and address flags. Specifically, the header contains a mode byte (which shows which of two data for mats is in use) and three address bytes. These three bytes store location as time: The first address byte holds minutes (0 to 74), the next address byte holds seconds (0 to 59), and the last holds the block number (0 to 75) within the second. For example, an ad dress of 59-20-45 denotes the 45th block in the 20th second of the 59th minute on the disc. This repeats information already found in the Q subcode channel, but it makes searching faster and more accurate.
As I've said, the mode byte distinguishes between two CD-ROM data formats. (Audio and other CD formats are told apart by Q-channel subcode data, while CD-ROM is distinguished from CD-I by data in the disc's table of contents; however, these standards aren't final.) Only one of the mode byte's eight bits is needed for this; the rest may be used for future formats.
The CD-ROM Mode 1 format provides for 2,048 bytes of each block-exactly 2 kilobytes-to be devoted to user data. The Mode 2 format allows for 2,336 bytes of user data. Fewer bytes are given to error detection and correction (EDC/ECC) in Mode 2 than in Mode 1, as shown in Fig. 1.
Fig.
1--Data formats for Mode 1 and Mode 2 CD-ROM. The area shown is equivalent
to 98 CD audio frames or 1/75 S.
Because of extended error correction, Mode 1 will have the greatest application. The 2,048 bytes of user data is where the actual CD-ROM data re sides. (Recapitulating, this is formed from part of the data area gathered from 98 frames of data formerly used for music data.) The remaining 288 bytes, called auxiliary system data, provide for a higher degree of error detection and correction; this is a pre requisite for successful numerical data storage, which is more demanding than audio data. This system data forms an extended error-correction code (EDC and Layered ECC) existing independently of, and in addition to, the cross-interleave Reed-Solomon error-correction code already watching over each CD frame; thus, the error rate is improved over that of an audio CD. Typical CD-ROM bit-error rates will be between 10-16 and 10-17, or one uncorrectable bit in every 1016 to 1017 bits. The extended code may be processed either by hardware circuitry contained in the CD-ROM player or by a host computer.
CD-ROM production begins with a master data tape(s), recorded in computer magnetic-tape format. The data is entered into a pre-mastering mini computer, where the data is placed into the CD-ROM format. (Header, error correction, subcode, etc. are generated.) Following pre-mastering, actual disc production is identical to that of audio CDs, albeit with greater quality control to ensure accuracy of data, because computer data is less forgiving of errors than music.
While CD-ROM uses a data format similar to that of music CDs, the discs are not compatible. A CD-ROM drive dispenses with D/A conversion, output filtering, and audio output stages, but requires a special interface and a computer for output. A CD-ROM drive alone costs less than an audio CD player; the consolidation of both functions into one player is ideally cost effective. A combination disc holding both music and ROM material could then be implemented.
Player design, even for a combination system, is straightforward. An intelligent interface and system controller is included in the drive, ready for connection to the host computer. Data is transferred to the host computer in blocks of 2 kilobytes. Using the header field, the sector address (in minutes, seconds, and blocks) is verified by the host computer. The disc table of con tents can be transferred to the host computer. After receiving an "audio" command from the host, the player can output audio data, accessed by the subcodes in the disc. The disc rotates only when a drive-select "ready" command is given; otherwise the disc is stationary to save power and motor wear. Multiple drives can be controlled by one host computer's interface, thus leaving open possibilities for further expansion.
The average data-transfer rate is 75 sectors per second, and therefore 153.6 kilobytes per second for the constant linear velocity (CLV) of 1.2 meters per second (same as the CD). Proprietary CD-ROM systems can achieve faster rates with special components. Access time for any point on a disc is less than 1 second.
A CD-ROM drive is a rather elegant data storage/retrieval system. Its possible applications are numerous. Extensive amounts of information can be conveniently stored on a few discs; for example, a CD-ROM telephone book for the entire U.S. would be feasible.
Assuming 100 million listed telephones, and 40 bytes of information associated with each one, the total of 4 gigabytes could be stored on about five discs--a stack about 1/4 inch high.
The CD-ROM will form the basis for a new electronic publishing medium applicable to books, dictionaries, technical manuals, business catalogs, computer storage, expert systems, artificial intelligence, and so on. It represents an entirely new technology of information dissemination.
Finally, an interesting analogy. Just how much information can be stored on a CD? It has been estimated that the capacity of a single disc is some 1 or 2% of that of the human brain. The equivalent of a year of human experience could be stored on a single disc, available for instant recall. This remark ably illustrates the storage potential of the CD-or the shallowness of this view of human existence, depending on your point of view.
Well, what the heck, you're saying. Artificial intelligence? The shallowness of human existence? Does CD-ROM have anything to do with audio? It sure does. One forthcoming CD-ROM product, CD-I, will do absolutely everything the audio CD does, and much, much more. I'll begin with that story in my next installment.
(adapted from Audio magazine, Dec. 1986)
= = = =