First, I think it might be good to try and define what "audio" is. To us Humans, audio refers mainly to information that can be heard, or perceived with our ears. Audio has to do with sound, and is therefore a branch of Mechanical Physics.

Sound is produced when an object or a substance is set into motion, and that motion is communicated to our ears, usually by the air around us. Our eardrum vibrates in response to the disturbance, and nerves in our inner ear send signals to the brain that we decode as 'hearing'. Of course it's actually a little more complicated, but this will do for now. Sound generally has a duration in time, and therefore the waves that produce it also exist for some time, like thunder, or the note of a piano dying away, or listening to continuous music. All these sounds are continuously varying in time. If we use a microphone and record them, we turn the air vibrations into electrical ones. These electrical signals are an 'analog', or an electrical copy, of the original sound pressure waves in the atmosphere. Thus, when we speak of 'Analog Audio', we refer to electronic equipment which is designed and built to handle continuously varying signals that represent sound to us People types.

Many different and varying ways have been devised by clever people, for the last nearly two hundred years, to record (and eventually reproduce) the phenomena of sound vibrations. It is possible to direct sound at an open flame, and it will respond to the sound by changing it's shape. If fine sand or powder is sprinkled on the body of a musical instrument, say a violin, when the instrument is played, the powder will arrange itself in patterns according to the sound vibrations existing in the wood of the soundbox. Mirrors can be attached to stretched rubber diaphragms, and a beam of light shone on them, which will deflect and vibrate according to sound waves which strike it. With the coming of the age of Electricity, newer and better methods were discovered and refined for capturing and playing sound... the first of which was a Danish invention that used a strip of steel tape to record telephone conversations, constructed by Valdemar Poulsen in the 1880s. Then came Edison with his tin-foil cylinder, and Berliner with the wax disc, and Western Electric with the first widely-used optical recording on photographic film. All of these methods, up to today's cassette tapes and vinyl records, use an 'analog' of the sound to record and play back the continuous program material.

BUT: there are these things called computers. Once upon a time, they were huge and slow and unbelievably expensive. No one outside of a few well-funded researchers, mostly in the various Phone Companies, would have dared to try and store sound on a computer. Then, as Technology and Computers advanced, they became smaller and faster and ever more powerful. It became increasingly feasible to think about recording audio (and video) into a computing system. There's just one small catch... computers work with numbers, and numbers only. Modern digital computers work with only two numbers: Zero and One. This is called the Binary system, and 0 and 1 are called BITS, for BInary digiTs. Computers only know on and off, yes/no, high, low... only two states. We need to somehow get our continuous sound into this state. Here's a short description of how that might happen.

Everyone has seen a movie. A movie film is composed of many many individual still pictures, one after the other. The movie camera opens it's shutter, light travels down the lens and exposes one frame of the film, the shutter closes, the film is advanced in the camera by one frame, the shutter opens again, the new frame is exposed, and so on. In 35MM cinema work, this happens 24 times a second. onversely, when the movie is shown in the theatre, it is run in a projector, which has a powerful light source and a lens to focus the image on the screen. Like the camera, the projector shutter opens, the image of one frame is shown on the screen; the shutter closes, the projector mechanism pulls the film to the next frame, and the shutter opens, showing it, and so on. We do not record the scene continuously, rather we break it up into a succession of 'snapshots' that, played back at 24 frames per second, appear to us to have continues motion. In actuality, you can see that much of the action occurring before a movie camera is 'thrown away'... the camera only records during the instant that the shutter is open.

Digital audio works this way, too. Let's take the example of recording a piece of music on a CD. We'll try and record someone playing a guitar on our CD. First, we need to use a microphone, to convert the mechanical sound vibrations of the air surrounding the guitar into corresponding electrical analog signals. In this example, for simplicity, we will record directly into our computer, much like you might do at home by hooking a microphone into the jacks on your computer's soundcard. Now, we have a varying electrical voltage at the input of our Digital system. First, we need to break this up into numbers that our computer can understand... that is, we need to make 1s and 0s out of the sound waves. We can use a type of circuit called a 'Sample and Hold' device. This is sort of like the shutter in our camera, and it takes 'snapshots' of the input voltage and holds each one for further processing. In the case of a CD, we want to do this 44,100 times each second. Okay, wow! We've got quite a lot of info already! Each of these 'snapshots' is actually the voltage reading taken the instant the sample and hold circuit opened its "shutter" and looked at the input analog audio. But, the poor computer can't deal directly with any kind of voltage, so we need next, the services of the 'Analog to Digital Converter' (or ADC). This little bit of magic takes these successive voltage snapshots and converts them into the Binary representation that our computer can easily understand and process. It also does this 44,100 times every second. For example, if the current snapshot from the Sample and Hold circuit is +1.00 volts, the result from the Analog to Digital Converter might be 1111 1000 0000 0000. If +1.5 volts, then 1111 1000 1100 000, and so on. Now we have a continuing stream of binary numbers, each representing a snapshot of an analog voltage, coming along in our computer at 44.1K/sec.

To the computer, this is now just information, it is just a file, no different from any other file. If we record 1 minute of guitar solo, the resulting file will be 60 x 44,100, or 2.646 [16 bit] million samples in length. Seeing as how most computers represent internal data in Bytes [of 8 Bits each], in the case of our home computer it would be 5.292 Million Bytes of data. If stereo, we double that; about 10 Mb. Again, internally to the computer, there is no difference between this file, a picture file containing an image, or a long document. Now, our guitar audio is 'numbers' and we can manipulate and process those numbers anyway we like. So lets burn a CD, shall we?

A CD is a plastic sandwich with a very thin aluminium layer in the middle. By using a strong laser beam, we can burn tiny holes in this layer. We tell the computer to go fetch the Guitar Solo File, and send it to the CDR to be recorded. The file, which is a 5.2Mb long, 16 bit wide block of data, is first read into a circuit that takes each individual Byte of data, and reads it out on bit at a time, in order. This is sent to the recording laser. We spin the CD (at 300 RPM) and position the recording beam in near the center of the CD. The beam is switched on and off, and data from our file is read into it bit after bit... and a spiral track of microscopic pits are left behind as the file is read out. We burn short pits for zeros, and long pits for ones, and we keep this up until we come to the end of our file. Thus we now have the 10 minute file recorded onto the CD layer as a pattern of sequential bits.

Oh, now you want to actually hear what we recorded? Okay.. let's try that:

We spin the CD at 300 RPM and shine a weak laser beam onto the pitted aluminium layer on the CD, and use a photocell to record the reflections coming back from the spiral track. If there are no pits, the light is reflected completely from the aluminium layer, and if there is a long or a short pit, the light does not get reflected. Thus, coming from the photocell, we have a signal switching on or off (1 or 0) at a rate corresponding with the pits previously burned into the CD. We need first to put them back together into Bytes, and we need to assemble them into a file in our computer's memory. Computers are good at this sort of thing; its what they were designed to do. So we have reconstructed the original file which we made previously, containing binary numbers representing analog voltages in snapshots taken from our input signal. So the process is reversed to hear the recording: we take the numbers in the file and present them, one after another in sequence, 44,100 times per second, to a circuit called a ..... wait for it!!!... Digital to Analog Converter. This device accepts digital bytes and makes a voltage on it's output that corresponds exactly to the voltage that made that particular byte pattern in the first place. The output of this DAC is then routed to your headphones, or speakers, or wherever else you might wish.

And that's it! In actual practice, of course the process is a little more complicated, and there are many practical electronic and digital considerations that need to be taken into account, but this will give you just an idea of what is meant by 'Digital Audio'.

In closing, also remember that an audio file, once represented as data in a digital system, can be read out, replaced, copied, manipulated, and all of the other things we've come to expect, with out being degraded by the process. Analog recordings, on the other hand, all get a little worse with each reproduction until at some point they become unusable. Bits, however, are pretty much forever..