Digital audio is the representation of sound signals through a set of binary data. A complete digital audio system usually begins with a transceiver (microphone) that converts the pressure wave that represents the sound to an analog electrical signal.
This analog signal goes through an analog signal processing system, in which limitations in frequency, equalization, amplification and other processes such as compaction can be performed. The equalization aims to counteract the particular frequency response of the transceiver used so that the analog signal closely resembles the original audio signal.
After analog processing the signal is sampled, quantified and encoded. Sampling takes a discrete number of analog signal values per second (sampling rate) and quantification assigns discrete analog values to those samples, which means a loss of information (the signal is no longer the same as the original). The coding assigns a sequence of bits to each discrete analog value. The length of the bit sequence is a function of the number of analog levels used in the quantization. The sampling rate and the number of bits per sample are two of the fundamental parameters to choose when you want to digitally process a certain audio signal.
The digital audio formats try to represent that set of digital samples (or a modification) of them efficiently, so that it is optimized depending on the application, either the volume of the data to be stored or the processing capacity necessary to obtain the starting samples. In this sense there is a very widespread audio format that is not considered digital audio: the MIDI format. MIDI does not start from digital samples of sound, but stores the musical description of the sound, being a representation of the score of the same.
The digital audio system usually ends the reverse process to that described. The set of samples they represent are obtained from the stored digital representation. These samples go through a digital-analog conversion process providing an analog signal that after a processing (filtering, amplification, equalization, etc.) affects the output transceiver (speaker) that converts the electrical signal to a pressure wave that represents Sound.
Digital audio quality
The quality of the digital audio depends strongly on the parameters with which that sound signal has been acquired, but they are not the only important parameters for determining the quality.
One way to estimate the quality of digital sound is to analyze the signal difference between the original sound and the sound reproduced from its digital representation. According to this strategy we can talk about a specific signal to noise ratio. For audio systems that perform lossless digital compressions, this measure will be determined by the number of bits per sample and the sampling rate.
The number of bits per sample determines a number of quantification levels and these a signal-to-noise ratio of carrier peak that depends quadratically on the number of bits per sample in the case of uniform quantification. The sampling rate establishes a higher level for the spectral components that can be represented, and linear distortion may appear in the output signal and aliasing (or spectral overlap) if the signal filtering is not adequate.
For digital systems with another type of compression, the signal to noise ratio can indicate very small values even if the signals are identical to the human ear.
The reason is that the signal to noise ratio is not a good parameter of sound quality measurement because the quality perceived by the listener is determined by the response of the human ear to the sound waves, which does not perceive many of the possible differences Logically, if the signals are very similar, the ear cannot differentiate them, but they can also be very different and can be perceived as the original signal. Therefore, the evaluation of the quality of a digital system through sensitivity parameters of the human ear and specific tests with specialized listeners seems more appropriate.
It is in this sense that the quality of digital audio systems is evaluated today. Both MPEG and Dolby Digital (AC-3), which establish perceptual compressions, perform test benches to estimate the quality of the encodings.
