DRM is a modern digital broadcast standard
Consider the problem of a compact representation of Nia’s audio signals in standard DRM, the new wasp standard that has the som index ISO / IES 14496 (MPEG -4 standard).
The third part of this standard (“Sound”) describes the encoded representation of natural and synthetic sound samples. This distinction makes it possible to distinguish the main part, which is an individual characteristic of the given signal, from the transmitted sound compositions, and the part that can be synthesized on a computer. When transmitting an audio signal, it is necessary to transmit its main part and descriptionThis is the part of the signal that can be synthesized at the receiving end. The transmission of a description of the signal instead of itself allows to recreate in real time on the receiving side an exact analogy of the original sound fragment
at low traffic flow speeds. This was one of the problems that the DRM community solved based on the MPEG-4 standard.
DRM uses advanced audio coding (AAC) and harmonic vector linear predictive speech coding (CELP and HVXC). For a tangible improvement in sound quality (especially with AAS), a special method can be used to increase the efficiency (high frequency reconstruction) of sound coding (SBR).
The encoded audio transmission signal is represented as superframes (superframes) of constant length. In superframes, an information service UEP is provided (ie, purely voice). Specific information to configure the audio data stream is transmitted on the SDC channel. Note that if no special measures are taken, when encoding a channel, all information bits are equally protected against channel errors (EEP algorithm), that is, the protection is carried out with the same degree of redundancy. At the same time, it is known that human perception of sound is characterized by uneven sensitivity to errors arising in the digital information stream at the output of the encoder. Therefore, it is quite natural to want to provide unequal protection against errors, that is, to extend a higher degree of protection for that part of the information bits,
MPEG-4 AAC
Figure 1. Structure of the AAC audio superframe For universal audio encoding, the MPEG-4 AAC algorithm is used, the best of similar algorithms suitable for use in DRM system. In the standard application of a mono AAC encoder on the shortwave (KB) channel, a bit rate of 20 kbps is provided. Of the possible extensions to the standard, only SBR technology is allowed.
The MPEG-4 AAC audio coding standard is part of the MPEG-4 audio standard (ISO / IEC 14496-3 + ISO / IEC 14496-3 / Amd1). The AAC digital stream in the DRM system is a digital stream of the MPEG-4 audio standard, version 2 (designed for use on channels with a high level of interference). Of the possible types of audio encoders (ISO / IEC standard objects), only the low complexity (LC) version of the ER AAC encoder belongs to the high-quality encoding algorithms – it is used in the DRM system. Among the existing methods to organize a digital stream MPEG-4 AAC, version 2, the version immune to noise HCR (Huffman Codeword Reordering) is selected, which is characterized by a low sensitivity of the audio data to errors in the transmission channel and a minimum digital bit rate.
The characteristics of the digital stream formation at the output of the AAC encoder in the DRM system are as follows:
the bit rate can be arbitrary, however it must be changed in steps of 20 bps to ensure alignment of the 400 millisecond audio overframe;
sampling frequency values (f d) – 12 and 24 kHz;
the conversion length is 960 samples, which corresponds, depending on the sample rate, to the duration of a sound frame of 80 or 40 ms. This selection ensures that the duration of the audio frames is consistent with the logical frame at the MSC;
noise immunity. The MPEG-4 encoder has the means to protect the AAC – digital stream on channels with a high level of interference;
Super Audio Frames: 5 (f d = 12 kHz) or 10 (f d = 24 kHz) audio frames make up a super frame. An audio superframe has a constant length (400 ms), which determines the possibility of filling it with a number (5 or 10) of the simplest audio frames, each of which must also consist of two parts. An audio superframe is always transmitted in a logical frame (see Part 2, BC No. 8). Because of this, there is no need for additional synchronization during audio encoding. The structure of the audio superframe also provides for the implementation of the uneven protection function;
