This passage was written in 1984.
It is now possible to hear a recording of Caruso's singing that is far superior to any made during his lifetime. A decades-old wax-cylinder recording of this great operatic tenor has been digitized, and the digitized signal has been processed by computer to remove the extraneous sound, or “noise,” introduced by the now “ancient” wax-cylinder recording process. Although this digital technique needs improvement, it represents a new and superior way of recording and processing sound which overcomes many of the limitations of analog recording. In analog recording systems, the original sound is represented as a continuous waveform created by variations in the sound's amplitude over time. When analog playback systems reproduce this waveform, however, they invariably introduce distortions. First, the waveform produced during playback differs somewhat from the original waveform. Second, the medium that stores the analog recording creates noise during playback which gets added to the recorded sounds. Digital recordings, by contrast, reduce the original sound to a series of discrete numbers that represent the sound's waveform. Because the digital playback system “reads” only numbers, any noise and distortion that may accumulate during storage and manipulation of the digitized signal will have little effect: as long as the numbers remain recognizable, the original waveform will be reconstructed with little loss in quality. However, because the waveform is continuous, while its digital representation is composed of discrete numbers, it is impossible for digital systems to avoid some distortion. One kind of distortion, called “sampling error,” occurs if the sound is sampled (i.e., its amplitude is measured) too infrequently, so that the amplitude changes more than one quantum (the smallest change in amplitude measured by the digital system) between samplings. In effect, the sound is changing too quickly for the system to record it accurately. A second form of distortion is “quantizing error,” which arises when the amplitude being measured is not a whole number of quanta, forcing the digital recorder to round off. Over the long term, these errors are random, and the noise produced (a background buzzing) is similar to analog noise except that it only occurs when recorded sounds are being reproduced
Which of the following statements about the numbers by which sound is represented in a digital system can be inferred from the passage?
They describe the time interval between successive sounds in a passage of music.
They model large changes in the amplitude of the initial sound with relatively poor precision.
They are slightly altered each time they are read by the playback apparatus.
They are not readily altered by distortion and noise accumulated as the digital signal is stored and manipulated.
They are stored in the recording medium in small groups that can be read simultaneously by the playback apparatus.
GMAT官方解析回复可见
本帖隐藏的内容需要回复才可以浏览 | 
