O.B. Popov1, T.V. Chernysheva2, V.A. Abramov3, A.A. Borisov4

1–4 Moscow Technical University of Communications and Informatics (Moscow, Russia)

1 olegp45@yandex.ru, 2 krba2012@yandex.ru, 3 vabramov44@mail.ru, 4 a.borisov199@gmail.com

The methods of digital representation of an audio signal are constantly becoming more complicated, more and more complex algorithms for describing the signal and eliminating redundancy are being used. At the same time, signal changes introduced by the basic procedures of analog-to-digital conversion and processing of the audio signal in the transmission and storage channel are often not taken into account. Thus, the spectrum of the audio signal error signal contains large levels of these errors in the high frequency region, and, consequently, is poorly masked by the signal itself and is noticeable to the listener. It is shown that signal distortions occur not only during analog-to-digital conversion, but also during transmission over the channel as a result of the need to change the sampling frequency, while oversampling errors increase at moments of signal unsteadiness (attack), leading to changes in its spectrum. To reduce such distortions, the possibility of using a sampling frequency conversion algorithm not in the time domain, but in the frequency domain after a fast Fourier transform is considered. In cases where the sequence of changes in sampling frequencies in the transmission channel is known in advance, conversion errors can be predicted and the possibility of their introduction into the original signal as pre-orders is considered. Attention is drawn to the fact that with a compact representation of the audio broadcast signal, the task of preserving the waveform is obviously not set and only the preservation of a relatively high subjective transmission quality is ensured, but at the same time the difference between the original and compactly presented signal can be quite large and reach 15–20% of the original signal.

It is shown that for an objective assessment of the quality of audio signal transmission, a set of parameters can be used that make it possible to predict the listener's assessment of sound quality – these are energy, rhythmic, envelope shape parameters (attacks and recessions), as well as spectral parameters. In order to assess the distortion of the audio signal in communication channels with various types of treatments, along with spectral analysis, the possibility of using kepstral parameters is also considered, allowing for the separation of mixed audio signals located in the common frequency range, when a large number of modulation components of the signal are converted into their reduced number, while the kepstral representation makes it possible to analyze modulated oscillations with isolation various combinational frequency components.

It is also shown that cepstral analysis opens up the possibility of assessing the connectivity of individual spectral components in an audio signal, as well as determining the relationship between harmonic and noise components in the signal and, in addition, cepstr allows you to determine the basic tone of the signal, its position and movements on the frequency axis, while the advantage of cepstral analysis is also expressed in the fact that any processing and distortion of the audio signal is manifested on the kepstral characteristic and can be easily tracked. The use of the results of this work makes it possible to improve the quality of broadcast signals and information programs, as well as to better control and regulate sound broadcasting equipment and thereby increase the popularity ratings of broadcasting stations, as well as increase the economic efficiency of these stations.

Popov O.B., Chernysheva T.V., Abramov V.A., Borisov A.A. Distortion of the audio signal in the transmission channel. Electromagnetic waves and electronic systems. 2023. V. 28. № 6. P. 13−25. DOI: https://doi.org/10.18127/j15604128-202306-02 (in Russian)

1. Aladinskiy V.A., Kuzminskiy S.V. Method for automatic evaluation of the voice signals quality with low bit rate voice coding. Achievements of modern radioelectronics. 2021. V. 75. № 4. P. 40–47. DOI 10.18127/j20700784-202104-(in Russian)
2. ITU-T Rec. P.862. PESQ an objective method for end-to-end speech quality assessment of narrowband telephone networks and speech codecs, February 2001.
3. GOST R52742-2007. Channels and paths of audio broadcasting. Typical structures. The main quality parameters. Measurement methods. (in Russian)
4. Taktakishvili V., Ovchinnikov A., Popov O., Abramov V. Objective Assessment of the Quality of Transmission and Informativeness of a Speech Signal According to Statistical Parameters. Conference of Open Innovations Association, FRUCT. 2019. № 24. P. 741–746.
5. Abramov V.A., Popov O.B., Richter S.G. Audio processing of digital broadcasting signals and its consequences. T-Comm: Telecommunications and transport.2016. V. 10. № 6. P. 17–20. (in Russian)
6. Patent for the invention RUS2691122 dated 11.06.2019. Method and device for companding sound broadcast signals. Abramov V.A., Popov O.B., Orlov V.G. (in Russian)
7. Patent for the invention RUS2731339 dated 01.09.2020. Method and device for measuring the power and steepness of the rise of non-stationary acoustic signal sites. Abramov V.A., Popov O.B., Taktakishvili V.G., Ovchinnikov A.A. (in Russian)
8. Certificate of state registration of a computer program № 2022666583 dated 05.09.2022. Compander system for transmitting a sound broadcasting signal on "Compander 1". Popov O.B., Abramov V.A., Prosecutor G.A. (in Russian)
9. Patent for the invention RUS2405262 dated 27.11.2010. A method for changing the transmission rate of a digital audio signal of broadcasting and a device for its implementation. Abramov V.A., Popov O.B., Richter S.G. (in Russian)
10. Patent for the invention RUS2773261 dated 01.06.2022. Method and device for measuring rhythmic frequencies, power and duration of declines of non-stationary acoustic signal sections. Abramov V.A., Popov O.B., Vlasyuk I.V., Saker A.V. (in Russian)
11. Patent for the invention RUS2756934 dated 07.10.2021. Method and device for measuring the spectrum of information acoustic signals with distortion compensation. Abramov V.A., Popov O.B., Vlasyuk I.V., Saker A.V. (in Russian)
12. Certificate of state registration of a computer program № 2019667523 dated 24.12.2019. Complex representation of signals by "Complex". Abramov V.A., Popov O.B., Taktakishvili V.G., Chernysheva T.V., Ovchinnikov A.A. (in Russian)
13. Certificate of state registration of a computer program № 2022616423 dated 19.04.2022. A program for high-precision spectral analysis of an audio signal. PO "PREP-Spectrum". Abramov V.A., Popov O.B. (in Russian)
14. Patent for the invention RUS2584462 dated 06.10.2014. A method for measuring the spectrum of information acoustic signals of broadcasting and a device for its implementation. Abramov V.A., Popov O.B. (in Russian)