Digital audio is the name of digitized audio signals ( music, etc.).
Analog- to-digital conversion
Tones are sound waves. They can to a sound transducer such as a microphone, are converted into analog electrical signals to be transferred, processed and recorded.
Digitizing means to transform the analog signals into discrete levels. For this purpose, the values of the audio signals must be sufficiently often sampled and stored. According to the sampling theorem known in the digital signal processing, the sampling frequency must be more than twice as great as the highest frequency of the signal to be sampled. If too rarely sampled, the result is the so-called aliasing. Since people only listen to about 20 kHz signals, Audio signals are sampled at 44.1kHz for CDs. Higher frequencies in the audio signal must be filtered before ( analog), otherwise there will be errors. Aliasing would " shut reflect " the high frequencies to low frequencies.
High sampling rates not only increase the maximum displayable frequency, but also facilitate the digital processing of the audio: is the available sample rate higher than the required, so you can digital audio effects use more samples to calculate and thus, for example interpolate tempo dilations better. More information at oversampling.
Since the digital -to-analog conversion a mirrored signal above the maximum usable frequency arises this with analog filters ( Hicut / low pass) must be removed. Ideally, all shares would be at 44.1 kHz sample rate below 22.05 kHz happen lossless ( low-pass ) are above 22.05 kHz, however, completely suppressed. Since such steep-sided filters are expensive to produce, can be at a higher sampling rate (eg 48 kHz) use the range of 22.05 kHz to 24 kHz as sliding for the low pass and thus use at a similar frequency range Hicuts cheaper.
More important than the sampling rate of the dynamic range, which is described by the PCM word length. It indicates how many changes in volume can be digitally encoded. While one can hardly perceive an increase in sample rate of 48 kHz to 96 kHz due to the low frequency portion of the top of the range, an extension of the sample width of 16 to 24 bits for the ear is clearly audible.
For uncompressed PCM encoded audio data produces large amounts of data. The memory required per second is given by:
- Bits per second = sample rate * sample width * channels
For a CD ( 44.1 kHz, 16bit, 2 channels) hence resulting in 1411200 bits per second, or 10.1 MB per minute.
44100 Hz * 16bit * 2 = 1411200 bits per second
/ 8 = 176400 bytes per second
/ 1024 = 172.2656 Kilobytes per second
/ 1024 = .1682 MB per second
* 60 = 10.0936 MB per minute
To reduce these amounts of data formats such as Ogg / Vorbis or MP3 have been created that achieved by lossy processing a reduction in the amounts of data. Depending on the desired output size the quality varies here between alienated not noticeable and strong; one speaks of the compression artifact, it can be heard, the resulting bit rate is chosen to be small, and thus clearly distinguish the result from the original.
The emergence and private use of the Internet allowed for a proliferation of digital audio files in previously unprecedented scale. Sharing networks such as Napster contained hundreds of thousands of music files. They are often the only source for no longer offered by record labels recordings.
Audio players allow playback of compressed audio files on the computer. But even portable players are the last few years with growing storage and functionality on the market. Often, however, these devices support only MP3 or WMA. Since WMA is a proprietary data format and patent- encumbered MP3, the manufacturers can restrict by a change in licensing policy, the use of audio files, prevent or provided with additional costs. For WMA has long pay- per-play in conversation, which every single play of the song is free of charge, or music only rented. Therefore better devices use the patent-free Ogg / Vorbis format, for never license fees may be charged, as it is publicly available to everyone.