Digital audio encoding
In fact, this or that digital form of representation of analog audio signals is already a coding method. – a sequence of numbers that describes an analog audio signal is itself a digital code.
However, the encoding that we are going to talk about now is something else. Now let’s look at the methods of encoding digital audio signals.
A digitized audio signal “in its pure form” (for example, in the form of one of the PCM variations discussed above) is a fairly accurate, but not the most compact, way of recording the original analog signal.
Judge for yourself. To obtain complete information about the original analog signal in the frequency range 0-20 kHz (in the audible frequency range), the analog signal must be sampled at a frequency of at least 40 kHz. Thus, the CD – DA standard (the standard for recording data on audio CDs familiar to all) establishes the following encoding parameters: recording of two or one channel in PCM format with a sampling frequency of 44.1 kHz and a depth 16-bit quantization bits. One hour of music in this format takes approximately 600 MB (60 minutes * 60 seconds * 2 channels * 44100 samples per second * 2 bytes per sample = approximately 605 MB). Considering that, for example, an ordinary music lover’s music collection may have 5000 tracks with an average length of about 3 minutes each, the amount of memory required to store it in its original digital form turns out to be very impressive. . Therefore, storing relatively large amounts of audio data, ensuring fairly good sound quality, requires the use of various “tricks” to compress the data.
In general, all existing methods for encoding audio information can be conditionally divided into only two types.
1. Lossless data compression (“Lossless encoding”) is a method of encoding (compacting) digital audio information, which enables one hundred percent recovery of the original data from the compressed stream (the term “data Original “here means the original form of the digitized audio data). This method of data compression is used in cases where one hundred percent absolute preservation of the quality of the original audio data is required. Lossless compression algorithms that exist today can reduce the volume of data occupied by 20-50% and at the same time guarantee a 100% recovery of the original digital material from the compressed data. The operating mechanisms of such encoders are similar to the operating mechanisms of general data archivers, such as ZIP or RAR, but at the same time they are specially adapted to compress audio data …. Lossless encoding While it is ideal in terms of preserving the quality of audio materials, it cannot provide a high level of compression.
2. There is another more modern form of data compaction. This so-called lossy data compression (Engl. “Lossy encoding”) The purpose of encoding is to achieve the highest data compression rate by all means while keeping sound quality at an acceptable level. The idea behind lossy encoding is based on two simple underlying considerations:
original digital audio data is redundant: it contains a lot of unnecessary information that is useless to the ear, which can be removed, thereby increasing the compression ratio;
Requirements for the sound quality of audio material may vary and depend on specific purposes and areas of use.
Lossy encoding is therefore called “lossy”, which results in the loss of some of the audio information. Such encoding leads to the fact that the decoded signal, when reproduced, sounds similar to the original, but in reality it is no longer identical to it. Most lossy coding methods rely on the use of psychoacoustic properties of the human auditory system, as well as various tricks associated with resampling and resampling the signal. In frequency, during the compression process, the encoder analyzes the audio data to identify various details of the sound that can be ignored. Disguised frequencies, inaudible and inaudible sound details can be sacrificed for a higher compression ratio. There, where only intelligibility is important in sound (for example, in telephony, where the presence of frequencies above 4 kHz is not necessary), the audio information in the encoding process is seriously “simplified”.
