Audio: Audio Signals

AUDIO SIGNALS

People with karaoke sets and home hi-fis know that they've got to connect their stuff together with dinky cables before it'll work. But what actually goes through the cables? How does the sound from the CD player go to the amplifier to the speakers?

A basic primer on sound

Sound travels through the air as longitudinal waves. Molecules of air vibrate, changing their distances between each other. When they come closer, they are known as compressions. When they move further, they are known as rarefactions.

These alternating rarefactions and compressions of air reach your eardrum, and cause the eardrum to vibrate too. Once your eardrum is vibrating, your ear will hear the sound.

Audio signals are measured in Bels, or in the more convenient Decibels (dB), which are one-tenth of a Bel. The Decibel or Bel standard is used to measure sound in the same way humans judge sound. For example, if we were to perceive a doubling of a standard audio volume, the sound level would have increased by about 6dB. However, how much extra energy is required to achieve this doubling would vary according to how loud the sound already was to begin with.

Analogue electric audio signals come in various standards, but typically, they are direct currents that vary in strength. The quickly changing high and low voltages in an audio signal correspond to the rarefactions and compressions of sound, though not necessarily in that order.

When the diaphragms of microphones vibrate, they create a little current into the audio cable. The same can be said for electric guitar pickups, that detect the movement of metal strings over a magnetic field. This current is your audio signal.

A loudspeaker does precisely the same thing in reverse. The current reaches the speaker, and the coils of wire in the speaker turn it back into actual movement, that causes the air surrounding the speaker to vibrate, which results in sound again.

But a microphone produces a very low voltage signal, while a loud speaker relies on a very very high voltage signal to work. The device that sits between the two is an amplifier. The amplifier will increase the voltage in for a higher-power audio signal proportionately to the voltage that it is receiving.

Noise & Distortion

In simple PA Systems, it is possible to work with just one microphone, one amplifier and one speaker. However, for more complicated setups, it might be necessary to alter the sound slightly. For example, karaoke sets add echo to the voice before pumping the sound to the loudspeaker. Also, you may want to add a voice-over to music, just like a DJ speaking over music.

It is overkill to do all this at a signal level powerful enough to drive speakers. To process audio signals that have come straight from the microphone would not be a good idea either, as the signal is barely a trickle of current. It would be like gathering a few scraps of heiroglyphs from an archealogical dig and trying to write a thesis on what the Egyptians thought of Barney the Dinosaur in ancient times. Not only will you have to add a lot of (probably erroneous) information, it will probably be completely different from the truth.

Erroneous sound signals are called noisy, because that's how they will sound. You can often hear this in badly tuned radios as static crackles or a quiet hiss in the background. This arises because your radio is not set up to receive all the audio information to give you a clear sound.

Distortion also damages sound signals. This arises when the equipment producing the sound cannot adequately handle the sounds to give you a fair representation of what it is supposed to sound like. The most well-known example is probably of the Distorted Guitar, the kind of sound produced by heavy metal guitarists. That grungy, grating sound used in guitar solos is interesting, but I'm pretty sure that's not what a guitar actually sounds like. A twang on a guitar string produces a twang, not a Deep Purple power chord.

That previous example also shows that Distortion and Noise are not always undesireable. They can be manipulated to good use.

Clipping is a form of distortion. Clipping occurs when a sound signal's voltage increases past a point that the equipment cannot output. As such, instead of maintaining the real wave-form of a sound, the sound is abruptly truncated where equipment meets its limits. This can sometimes cause the grating-guitar sort of sound.

The Line-Level

To minimise noise and distortion when processing sound, transferring it from one component to another, or reproducing recorded sound, there is a standard type of audio signal called the Line-Level signal. This electronic representation of sound is similar to the types used to drive speakers or coming from microphones and guitars. The only difference is in its strength. It is supposed to be rated at about 1 volt for professional applications, although I have no idea how they rated it. The professional term for this signal is called +4dBu.

Then some bright spark came along and said, "Hey, why should consumer products use the same sort of signal quality as professional products?" As a result, the totally redundant (but woefully popular) -10dBV signal was invented. This signal is rated at a tenth of a volt. There is actually no reason why there should be two types of signal. The two standards coexist in most studio setups, causing occasional conflicts.

Why was the line-level necessary? It is not so powerful to require high-power equipment to withstand the energy in the signal. In fact, line-level signals come out of just about every home hi-fi system (other than most amplifiers) so that they can be interconnected. It is also not so powerful that it would overload some circuits and cause distortion.

It is also not so low-power that noise begins to obscure all audio detail. As such, line-level signals are most appropriate for transferring audio information from component to component and sending to recording devices. How much power the devices use to record the sound is dependant on the individual method of recording.

The S/N Ratio

One reason why I don't like the -10dBV standard is because the Signal to-Noise Ratio (S/N Ratio) is for the -10dBV standard has to be lower by default. The S/N Ratio is used as a rough gauge to measure how noisy a signal or circuit is. Typically, for a given electronic circuit, the noise in that circuit will remain constant as long as nothing too drastic is done to it. The signal level in the circuit varies according to how much signal level you put into it. Therefore, it stands to reason that if you can get a higher-power clean signal to go into the circuit, there's no reason to use a low-power version, unless the high-power signal is going to overload and distort when it goes into the circuitry. But that is a simple problem that can be fixed with good design. A high-power signal will be less susceptible to noise corruption and cleaner in sound.

The S/N Ratio is basically determined by subtracting the average signal volume from the average noise level. As a result, the S/N Ratio is also measured in dBs. Most consumer equipment have S/N Ratios of 50dB and higher. Anything lower begins to have noticeable noise.

The S/N Ratio will vary depending on the type of signal you are putting into the circuitry. As such, a lot of professional measurements are used by A-weighting. This is measuring the S/N Ratio in comparison to typical audio signals. It does not take into account the S/N Ratio of sounds that only dogs can hear.

Impedance

This is probably the most misunderstood aspect of audio signals. I'm not too sure of my facts myself, but I'll give it a shot. Impedance is a measure of either an output's capability to drive inputs or an input's capability to receive signals from an output. It is measured in Ohms, so it is in some way related to the resistance of audio devices. It is often referred to as Z.

A perfect input device will have no impendance, and a perfect output device will have infinite impendance. In real terms, usually the impedances of outputs are 5 to 10 times that of standard inputs for acceptable performance. Anything less could result in distortion of sound. In effect, the output would not be capable of putting out enough clear signal for the input to pick up properly.

As an example, dynamic microphones typically have impedances of 300ohms to 600ohms. Condenser microphones have impedances in the thousands. Line outputs approach 10kohms.

Mixer inputs try to have impendances as low as possible to minimize distortion of the sound. Some mixers tout 'Very Low Impedance' or 'VLZ' as a feature of their mixers, which affects microphone inputs more audibly than line inputs. Line inputs are also known as 'Hi-Z' inputs, to accomodate signals that come from high-impedance equipment, i.e. Line-level signals.

The fact that output impedance is always much greater than input impedance also means that most high-impedance outputs can power multiple inputs at one time. A simple splitter cable would be able to allow two inputs to receive a clear signal from one output, or possibly even more. Thus, the reverse is not true...you shouldn't mix two outputs into one input by using a splitter cable. The impedance would be so badly offset that the sound would be audibly distorted.

Stereo sound

Most humans have two ears, duh. If sound was recorded with two separate microphones and played back with two separate speakers, you could possibly reproduce the same illusion of left-to-right direction for sound reproduction. This is why stereo was created: the common implementation of sounds recorded with two channels: left and right. In order fake the position of a sound somewhere in between, a proportion of the sound is distributed to each speaker.

Stereo sound doesn't succeed in giving the illusion that the band is right in front of you, but it does heighten the aesthetic interest of music or sound recorded. One real benefit of stereo is that you can separate the positions of different instruments, so that each instrument can be heard clearly while being part of a whole mix.

Stereo introduced some new terms of its own: Pan and Balance. Pan, short for panorama, dictates how much of a single sound should be given to the left speaker and how much to the right. This represents which direction in the stereo image the sound should seem to be coming from. Balance stands for how the volume of the left channel compares to that of the right channel in a stereo sound. For example, a stereo recording in which every instrument seems to be closer to the right speaker than the right has not been recorded with the correct balance (unless it was done deliberately). If you were to take a recording of a mono instrument, say, a saxaphone, and put the same sound into the left and right channels of a stereo mixer, adjusting the balance of the left and right channels would actually result in you changing the pan of that single saxaphone. It won't sound like two saxaphones, it would just sound like a single saxaphone moving from left to right as you alter the balance.

Back to the Bluffer's Guides