Basics of Sound

I'm sure you all know from school physics lessons that sound is a pattern of compressions in air. And biology would have given you the introduction on how the human ear works.

So, if you measured the air pressure at a certain point in a noisy room you would see something a little like this (time on the x axis, pressure on the y):

We generally call `pressure', amplitude. Now, if you call represent amplitude by voltage you have sound going down a wire. Take any two core cable and modulate (vary) the voltage across the wires so that it looks exactly like the pressure trace and you have an `unbalanced connection' (more on that later) like a phone lead.

Now, any waveform can be broken down into a sum of sine waves of differing frequencies and amplitudes. (You do know what a sine wave is, right?). That may sound complex, but you have all seen it when looking at the pretty display on your stereo system; the one with the bars that go up and down to the music. The ones on the left (low frequences) are the bass notes and the right (high frequencies) are the treble. Now imagine that you had lots of bars, then it could look a little like this:

(a Hertz (written Hz) is "one per second")

That's a frequency-amplitude representation (as opposed to the time-amplitude one we had before). Also known at frequency domain and time domain (FD and TD).

Digital Sound

In order to use computers to process sound we need to represent sound as a stream of bits. This is done by choosing a range of numbers as a max and min pressure level and then recording the pressure at frequent time intervals. The number of pressure levels (samples) taken per second is called the sampling rate (which is, itself, a frequency). The size of the numbers used is given in bits:

In that image you can see the time-amplitude graph as before (in yellow) and the digital approximation in purple. As you can see, as we increase the sampling rate (thus decreasing the width of the bars) and increase the number of bits used to store the number (thus decreasing the size of the steps) we get a better approximation of the original waveform.

Just as a guide, CD quality is 44.4KHz (44400 samples per second) and 16 bits. DVD-A stereo is 192KHz, 24-bit and 96KHz, 24-bit for 5.1 multichannel.

Nobody uses more than 32-bit. On good equipment you can tell the difference between 44KHz and 96KHz. I doubt I could tell the difference between that and 192KHz though.

Some terms you should know: ADC (Analoge to Digital Convertor/Conversion) is the process described above and DAC (Digital to Analoge Convertor/Conversion) is the reverse.

Still confused? Maybe these will help:

Reproduction

At some point, all this sound information has to be turned back into pressure in the air again via the wonders of speakers. I'm sure that you're roughly aware how a moving coil driver works so I'm not going to cover it here.

Now, assume that we connected our wire to a speaker - what problems would we have? Firstly, the voltage and current that we are supplying is minimal - it's only the signal. We would need to amplify the signal a huge amount in order to drive the speaker correctly. (More later).

But we would also find that high frequencies caused standing waves on the surface of the cone if it were too large. But also, if we make the cone small enough to deal with high frequences, we then cannot move enough air to reproduce the low frequences very well. Which leads us to the understanding that no speaker is good for all frequencies. Generally we have a number of different drivers for different frequeny bands. This is fine, but we need someway to split them up.

Since I'm talking about frequencies, you need to have the frequency-amplitude representation of sound in mind. Say that we had two drivers, we would wish to split a signal into the "left" half and "right half". The frequency at which we split is called the crossover frequency. Then we send the "left" (low freq) half to the sub (larger) driver and the "right" (high freq) to the mids/tweeter (smaller) driver.

There are two ways to acheive this splitting: passive crossover networks and an active crossover.

Passive crossover networks are circuits of capacitors, resistors and inductors (mainly) which, due to impedience charactoristics of these components, can split a signal. They are cheap, but burn energy off as heat and must be designed specifically for the exactly combination of drivers that will be connected.

Active crossovers are digital devices that ADC the signal, process it, then DAC two or more outputs. They can be configured for any drivers and work on an unamplified signal so don't waste power.

Amplification

So far I've glossed over the fact that our very low current signal gets converted into a room-shakingly powerful one before going into the drivers. This is done by (and no great surprise) amplifiers.

The job of an amplifier is to use its mains supply to output a powerful signal that has exactly the same waveform as its input signal. Usually an amplifier can work on two independant channels (a stereo amplifier). Otherwise it's a monoblock (one channel) or multichannel (>2 channels).

The design of amplifiers is a hell of subject itself, and I'm not going into it here. Most of the time you just don't care how it works. You just need to know what it's power rating is. This is usually given as a wattage into a number (often 8) of ohms. Speakers will usually have their impedience written on them, but remember that connecting a number of speakers in a chain alters it (and you don't just add them up).

If in doubt, ask someone when matching amplifiers to speakers but remember that you're more likely to damage a speaker with an under powered amp than an overpowered one. That's a little counter-intuitive so I'll explain.

If you have an overpowered amp then you'll hear as you overdrive the speakers (because you start with the amp set to minimum and bring it up slowly, right?). But with an underpowered amp you'll need to set it very high in order to get a decent volume out of it. This is all very fine until you ask the amplifier to climb a particually high peak and it finds that it doesn't have the power to do it. At this point it clips the output - imagine the top of a ssine wave cut off and flattened. Think about this and you'll realise that this output is asking the driver to perform a huge acceleration and this is very bad for it.

As a rough guide - 60-75W/channel will drive a home hifi system. 300-600W drives minor PA systems. 5KW and up are major PA systems.

Sound with Dramsoc

The Amp Rack

If you're reading this through you'll still have amplification fresh in your mind, so let's start with the Wavefront Amp Rack. This lives in the sound cupboard on level three of the Union building and looks a little like this:

The mains input is a 63-1 on the back, which can act as a pass through via a 63-1 male outlet, also on the back. It has four 16-1 outlets (used for each of the amps) and 4 13-amp sockets (two used for the crossovers).

The crossovers are XTA 224s. The amps are all Crown Macrotechs [Tweeter and High][Mid][Sub].

In normal operation, all volume controls should be at the 3 o'clock position, except the subs which should be at full.

It has master XLR inputs at the top and four speakon connections out of the back.

The crossovers are active and look a little like this:

They have two inputs, but we only ever use one ("A"), playing with input B achevies nothing. The input controls are the two LED meters in the middle. The controls for controling the input and each of the four outputs (on the right) are the same. The red button toggles the mute. If a channel is muted then the red LED over the button lights up. Pressing the grey button moves between Normal, Gain and Delay modes. This is reflected on the green LED display. In Gain and Delay modes the rightmost dial controls that aspect of the channel.

Generally, all channels should have no gain and some delay. (Where `some delay' read: we're not sure what it is. If in doubt, have no delay). Don't forget to zero any settings that you have played with after use. You shouldn't ever need to play with anything else.

Connecting up the Wavefronts

There are eight Wavefront cabs: four tops and four subs. They are almost always arranged in two pairs of stacks. Subs go under the tops and everything should be the right way up. Subs have one flyhook point and tops have two. Look for the "T" (on tops) and "B" (on subs) written on one of the catches of each cabinate. The "T" goes towards the sky and the "B" touches the floor. Keep the covers on them until they are in stacks.

There is a style to building the stacks - but I can't really explain it in on paper.

There are two long wavefront speakon runs (they are thick blue pairs of cable) - one longer than the other. Make sure you plug the connectors labeled "Tops" into the tops at one etc and the output labeled similarly at the amp rack, ok? Generally, connect everything up as in the diagram below. There are four short patch leads to connect between stack on the same side (as in diagram).

When you've connected them up you need to run though the drivers to check that everything is working correctly. Mute all the outputs on both crossovers then one person sits by them and someone else puts their ear to the drivers to check that they are all working. The person by the crossovers unmutes one output at a time (so that only one output is active at any time), and the other checks that both drivers are making a sound, yelling "Check" when they're happy with each one. Then you do it again with the other crossover/side of the stage (all the same thing).