Digital Audio Principle

Vibrations in matter. That is how we usually define sound. When playing the guitar, you strum a string, and it swings one way. The pressure is applied on the surrounding molecules and high pressure is created around the string. When the string swings back, the high pressure transforms into a low pressure zone. The pitch is defined by how many swings are made per minute. Here comes the term „frequency“. The frequency depends on the number of high and low pressure waves made within one second.

The purpose of the microphone is to converts the high and low pressure waves within the area where microphone takes effect. Microphone varies its voltage output accordingly. This is called an analog sound. The term “analog” has come from the fact that this voltage output is an analogue to the original sound waves.

Digital sound is quite different as you might have suspected. In order to receive digital sound there must be analog to digital converter working on the other end of the circuit. This converter takes periodic snapshots of the voltage levels coming from the microphone. The new term “sample” may be introduced here. Samples are these snapshots taken from the microphone.

This is possible to draw a parallel of digital audio and film. The rapidly moving sequence of still images is what we see in the movie. If few frames are shown per second, the flow of the movie would be less adequate, more roughly. Digital audio works by the same principle. Single snapshot of a movie is represented in a digital audio by a sample. Samples may be reproduced various ways. Each frame has a size and so does sample. In 16-bit digital audio each sample is a single 16-bit integer which value is ranged from –32768 through 32767. This number represents the amount of voltage of an analog sound for the same noise. In digital audio frequency is measured in hertz. Quality sound is 16-bit 44100Hz, or 44.1kHz.

Sixteen bits is usually not enough to represent a sound, and many musicians deal with the problem of this small size when modifying digital audio. When you apply effect on a sound, this can be viewed as solving a complex mathematical problem applying the solution on the series of samples. Every time you effects are applied on 16-bit integers, 3dB of sound are lost in the area where application took place. This math experiments with 16-bit integers usually lead to a lot of rounding and this is when the sound is lost. You may also lose sound at the upper end when samples start to clip, or when the 16-bit integer is unable to represent the high voltage of analog sound.

There are sound editing programs that are capable of retaining the fidelity of digital sound by using a 32-bit floating point numbers within the program. Floating point is defined as “representing decimal numbers in a computer that only understands integers”. When working with 32 bits for floating point number, some of those bits represent a number on the right side of the decimal place, and some represent a number on the left. When putting these two parts together a number like 2.2897 will appear. This method is very helpful if you are not satisfied with rounding decisions while applying effects to the sound. This does not exactly mean that processing in a 32-bit float system is lossless, yet this way you can minimize this losses so much that human ear would be unable to detect that the quality has worsened.