Copyright Michael Karbo, Denmark, Europe.
Chapter 5. Digital sound
In the previous chapters, we have looked at sound waves and loudspeaker units, which reproduce the sound waves of music.
Music CD with digital sound
In 1982, Philips and Sony introduced the latest digital media Compact Disc, which in the following years revolutionised the whole of the music industry. When we look back today, more than 20 years later, it is unbelievable how good a media, CDs have been.
The essence of a music CD is, that the music is digitised. Instead of containing a pattern of mechanical information (like in an LP record) or magnetic vibrations (like in a cassette tape), the CD disks contain bits – that is to say large collections of 0’s and 1’s. The advantages of the CD media compared with LP records are:
CDs have been and are an enormous success. But the computer revolution has created new problems for the music industry, which is in a crisis again.
Copying of CDs
When music – or other forms of data are digitised, then they can be copied without loss. Back in the start of the 1990’s, this wasn’t a big problem. But with the arrival of cheap CD burners, the mp3 format and the Internet completely new (and usually illegal) ways of copying have arisen. This development, which has seriously speeded up during the last five years, has been impossible for the music industry to stop. Digitizing of music has opened up for a lot of new and exciting possibilities for us users, while, at the same time, the record companies’ sales are undermined. Let us look a little closer at digitizing.
Sampling and coding
When music is digitised, the music’s sound waves have to, in one way or another, be measured and converted into data. The measurement of sound information takes place by sampling. Subsequently, there is a coding.
Coding converts sampled data to binary data, where there only are two data operators: logical 1 (high) and logical 0 (low). This, in fact, applies to every digital sound – no matter how complex, it might otherwise be. The dominant method of coding is called PCM, which stands for Pulse Code Modulation.
Taking samples with PCM takes place in telephone systems (all telephony is digital these days), in audio CDs, in MiniDiscs, in sound reproduction in Windows, in digital video cameras and much more.
Frequency and resolution
Through samples of sound wave data, the analog signals’ amplitude can be measured. These measurements (samples) are taken at certain intervals – also named sample rates, which are the number of samples that are taken per second.
This frequency of samples has to be at least double as high as the highest sound frequency in the material. Because our hearing reaches op to circa 20 kHz, then samples take place at 44,1 kHz.
The analog signals’ immediate amplitude is measured and rounded off to the nearest value within the chosen scale. This process is called quantization. Samples consist of thousands of small sound recordings.
Figure 23. Samples of 16 bit resolution.
The number of sample levels is the power of the number 2 (i.e. 2, 4, 8, 16, 32, 64, etc.). These values can be represented by a number on 1 bit, 2, 3, 4, 5, 6 bits etc. The sound of the CD is quantizised with a resolution of 16 bits. This gives a scale with more than 65.000 values. Every sample is the size of 16 bits and can assume a value between 0 and 65.535.
In Figure 23 you can see this illustrated. You see a section of a sound wave, which has been measured 26 times. Every measurement (sample) is given a 16 bits value. The overall amount of data (26 values each with a size of 16 bits) describes this little ”clip” of a musical piece.
A/D and D/A converters
Purely technically, a conversion takes place between digital and analog data. When sound is digitised, it is done in a so-called Analog to Digital Converter. When the digital sound is played back and reproduced in a loudspeaker system, then the CD’s digital sound track is converted into analog (electrical) signals, which is done in the player’s Digital to Analog Converter.
Figure 24. An ADC converts sound into data.
We can find both A/D and D/A converters in a computer’s sound card, because a sound card must be able to record and play back sound. During recording, the analog signals are sampled and encoded. During play back, the digital data is converted back to analog sound signals in the D/A converter. See a picture of a DAC chip in Figure 67 on page 3.
There are D/A converters in all forms of CD and mp3 players, because they all have to produce analog sound signals. Even a normal CD drive in a computer has a built in DAC, because the sound from a music CD has to be fetched through a headphone port:
Figure 25. Behind the little port to the headphone is a D/A converter.
Number of samples
To get a good quality of sound, it is necessary to sample many times. The more the samples, the more the digital recording will sound like the original sound. But irrespective of how many samples we use, the digital sound track will never be quite the same as the original, analog sound signals! You can see, below, a graphical illustration of the system. The fewer the samples there are, the more ”hacked” the sound curve is when reproduced. The hacks mean that there has been a rounding up in the individual samples and this gives reduced sound quality. The more samples, the nearer we come to the original sound.
Figure 26. The more the samples, the more precise the sound reproduction will be.
So sound recording has to be sampled before it can be converted into digital data. This sampling can be done in a sound card, as we will demonstrate later. But sampling also occurs in lots of other places.
When you, for example, buy an electric piano (see page 3), then the instrument contains sampled sounds. These are microphone recordings from big pianos, saxophones, guitars, etc., which are stored in the electric instrument and with which one can ”play together with”. When a record company records music for a new CD, the song and the instruments are recorded with microphones. This sound information is either recorded as analog or digital information, but in the end all the recordings are sampled because otherwise the music simply cannot be stored on a CD.
The uncompressed sound data is usually called PCM or wave data.