Detour 3

W hile we're on the subject of hi-fi equipment, lefs discuss a few more ierms

U s not enough for a piece of hi-fi equipment to cover a wide range ol frequencies; different frequencies in the range have to be treated equally Thai is. an amplifier or tape deck that covers 20 20,000 II/, bui provides a lot less gain above, say, 1,000 Hz than below, would sound very bassy. Ideally, hi-fi equipment should be able to handle signals of different frequencies equally well. Evenness of response is usually rated in decibels or dR For example, a typical amplifier might have a rating of "20-20,000 Hz ±1 dB." which means thai the gain (how much ii amplifies) does not van more than plus or minus 1 decibel from some midscafe value (We1 II have a detour later to evplain decibels \

In addition It) having a wide frequency range, ihe hi-fi device also should noi diston the is, its ouiput waveform should look like the input waveform (exccpl for possibly being larger or smaller). One to rale distortion is us THD or total harmonic distortion. Remember that it's the harmonics that make one signal ol a given frequency different from aitoLher signal of ihe same frequency. Hence if the output from an at a low speed. Then you rewind the tape and pla> it back, hut at double ihe speed so that ii takes only 10 minutes lo play. Can you thus send 20 minutes of speech, but pay for only a 10-minute phone call?

You can ccrtainly do thuL but your voice will sound like the Chipmunks (that's how they do their voices!) and may not be too understandable. Bui suppose your friend rccords your voice on another tape recorder, hut this time records at high speed and pla\s il back later at hall'-specd (try this with a Chipmunks record!) This stretches the 10 minute tape back into 20 minutes. Will this work? (And if il does, can you speed up the tape by a factor of 10 and pa> for only a 2-minute call?)

Yes , . , and no. What happens is that as you double the speed ol your lape, every frequency on the tape gels doubled Loo. A 1,000-Hz component of your voice becomes 2,000 Hz, and so on. The problem is that every component of your voice that is above 1.750 Hz or so gets doubled to above 3,500 Hz. and therefore doesn't make it through the phone line. In other words, your friend will only hear those components in your voice that are below 1,750 Hz, (And if you tried to speed things up by a factor of 10, he would only hear those parts of your voice that are below 350 Hz.)

In other words, it's the bandwidth of the telephone line that limits how amplifier or recorder looks different from the input. its harmonics mu>t somehow have been changed. The standard way of measuring this is to insert a pure ^me wave test signal ithat has no harmonies), arid look to see whether there arc any harmonics in the output. IT so, then the signal sol distorted. Hie THD number is a percentage *hat tells how much harmonic voltage got added to the pure signal. For instance, if the outpui from an amplifier (with a sine wave inpm thai should have no harmonics) is 10 volts of fundamental and 2 volts of harmonics, then there would he 20% THD (a terribly high number, by the way. I IIT) values of under one or two percent are more desirable).

Actually, though, harmonic distortion is not nearly as bad you think. Since music and speech normally ha\e harmonics anvwa>, adding an extra percent or so of harmonic to I hem is not too noticeable Amplifiers and other all-electronic hi-fi equipment tend to have low distortion, but tape recorders Lind mechanical components such as phonograph cartridges and speakers often have i high THD < sometimes as much as Vr to \iY> for speakers I.

Much more dangerous is IM or intennoiiulatwn distortion^ which introduces new frequencies not much information you can get across in 10 minutes. It' you used a higher-bandvMdth line—such as the special lines that broadcast stations lease from the phone company lor studio-to-tr&nsmittcr links, which cover up to 10,000 or 15,000 Hz—you could easily speed up your tape by a factor of 3 or 5, and still transmit all of your message (though siill only ai normal telephone-line quality).

So the idea is that there is a tradeoff between bandwidth anil time. If you have a fixed amount of information lo send, you can send it fast if you have a lot of bandwidth. But you have to send ii more slowly if the bandwidth is small. That explains why. lor example, a lax transmission can go through a regular telephone line, but a full-motion TV video image can't. The fax takes up to a minute to send one picture, Whereas the TV has to send il in 1/30 of a second.


Although the discussion has rambled off and on about various aspects of audio, we've actualK covered a lot of ground. We have seen the characteristics that make up an audio ua\c—the frequency. wave shape, and amplitude of the signaL We have looked at how harmonics affect the wave shape, and how the bandwidth of a system affects the sound quality that you can send through it. Next lime we w ill tackle transmission of video.

in the original at all. E\en 1/2\'t or 1/4*3 IM distortion is grating and unpleasant Unfortunately, IM distortion is net ver> often listed in *pee sheets for equipment; fortunately. however, IM distortion sort of goes hand in hand with THD, and a hi-11 device with low THD probably also has Uiu IM

FinalK, hi-fi equipment should have very little noise,, Noise can appear in the form of a low* pitched hum (often caused by a bad power supply; bad grounding, or bad shielding of a wire) and a high-frequency hiss. Either one is bad. Hifi equipment specs therefore often list the SNK or signal lo-noise ratio. This is the ratio between the loudest music it can handle and the noiie. For exampiet in a CD recording, the loudest music voltage is typical!)/ about 65,000 times higher than the noise voltage, while in a cassette recording it might only he 300or4<H) limes stronger In a telephone circuit, the ratio between the loudest voicc signal and the noise mi £ tot be a> low as I (J to 1 Since we're holding oil on our discussion oi decibels, lei me just say at this poult thai the 65,000 ratio is equivalent to about 95 dB, the ratio of 300 to 400 is about 50dK. and a ratio of Id is only about 20 dB.

Number 79 on your Feedback card events

0 0

Post a comment