Audio Analysis Principles

Publisher:美丽花朵Latest update time:2012-01-21 Keywords:audio Reading articles on mobile phones Scan QR code
Read articles on your mobile phone anytime, anywhere

1. Audio Signal and Audio Analysis
Audio is an important medium in multimedia. The frequency range of audio signals that we can hear is about 20Hz-20kHz, of which speech is distributed approximately within 300Hz-4kHz, while music and other natural sounds are distributed in the full range. Sound is recorded or reproduced by analog equipment to become analog audio, and then digitized to become digital audio. The audio analysis mentioned here is the process of extracting a series of characteristics of the signal in the time domain and frequency domain by taking digital audio signals as the analysis object and digital signal processing as the analysis method.
Audio analysis in various specific frequency ranges has different application fields. For example, the analysis of speech signals between 300-4kHz is mainly used in speech recognition, and its purpose is to determine the content of speech or judge the identity of the speaker; while the analysis of speech signals in the full range between 20-20kHz can be used to measure the performance of various audio devices. The so-called audio equipment is the various electronic devices needed in the whole process of picking up the actual sound and playing the sound, such as microphones, power amplifiers, speakers, etc. The main technical indicators for measuring audio equipment are frequency response characteristics, harmonic distortion, signal-to-noise ratio, dynamic range, etc.

2. Principles of audio analysis
The principles of audio analysis mainly involve the basic theory of digital signal processing, the basic methods of audio analysis, and the measurement and analysis of audio parameters, among which digital signal processing is the theoretical basis of audio analysis.
1. Technical basis of audio analysis
Fourier transform and signal sampling are the most basic technologies used in audio analysis. Fourier transform is the basis for spectrum analysis. Spectral analysis of signals refers to the establishment of various "spectra" with frequency as the horizontal axis, such as amplitude spectrum and phase spectrum, by obtaining the amplitude and phase of its components according to the frequency distribution law according to the frequency structure of the signal. In the signal, the periodic signal corresponds to a discrete spectrum after Fourier series transformation, while for non-periodic signals, it can be regarded as a periodic signal with an infinite period T. When the period approaches infinity, the fundamental spectrum line and the spectrum line interval (ω=2π/T) approach infinitesimal, so that the discrete spectrum becomes a continuous spectrum. Therefore, the spectrum of non-periodic signals is continuous.
In a computer-centered test system, the analog signal passes through an A/D converter before entering a digital computer to convert the continuous time signal into a discrete time signal, which is called signal sampling. Then it is converted into a discrete digital signal through amplitude quantization. In this way, a series of new problems will appear in the frequency domain, and the spectrum will change. After the analog signal is converted into a digital signal, its Fourier transform also becomes a discrete Fourier transform, which involves a series of problems such as sampling theorem, frequency aliasing, truncation and leakage, windowing and window function.
2. Audio analysis method
Usually when measuring and analyzing the audio of an audio device, the device is regarded as a black box system with input ports and output ports. Input a known signal into the system, and then obtain the output signal from the output end for analysis, so as to understand some characteristics of the system. This is the general method of audio analysis. The signal input into the audio device is called the excitation signal. The excitation signal can be a periodic signal such as sine or square wave, or a random signal such as white noise or pink noise, or a signal such as dual tone, multi-tone, or sine burst. The most commonly used detection and analysis methods include sine signal detection, pulse signal detection, and maximum length sequence signal detection.

3. Audio parameter measurement and analysis
Audio measurement generally includes basic parameters such as signal voltage, frequency, signal-to-noise ratio, harmonic distortion, etc. Most audio parameters can be composed of these basic parameters. Audio analysis can be divided into several categories, such as time domain analysis, frequency domain analysis, and time-frequency analysis. Since the harmonic distortion of the signal is more important for audio measurement, it is separately classified as distortion analysis. The following introduces various audio parameter measurements and audio analysis.
1. Basic parameter measurement
The basic parameters that need to be measured in audio measurement are mainly voltage, frequency, and signal-to-noise ratio. Voltage testing can be divided into several types, such as root mean square voltage (RMS), average voltage, and peak voltage.
Frequency is one of the most basic parameters in audio measurement. Usually, a high-frequency precision clock is used as a reference to measure the frequency of the signal. When measuring the frequency, the input signal and the reference clock are counted simultaneously within a limited time, and then the count values ​​of the two are compared and multiplied by the frequency of the reference clock to obtain the signal frequency. With the improvement of the computing speed of microprocessor chips, the frequency of the signal can also be calculated by software using fast Fourier transform.
The signal-to-noise ratio is a basic performance indicator of audio equipment, which is the ratio of the effective voltage of the signal to the noise voltage. The calculation formula of the signal-to-noise ratio is:

In actual measurement, for convenience, the total voltage of the signal with noise is usually used instead of the signal voltage to calculate the signal-to-noise ratio.
2. Time domain analysis
Time domain analysis usually involves inputting a test signal into the audio device to be tested and observing the time domain waveform of the device output signal to evaluate the relevant performance of the device. The most commonly used time domain analysis test signals include sine signals, square wave signals, step signals, and single-tone mutation signals. For example, inputting a sine signal into the device and observing the distortion of the output signal time domain waveform is a time domain analysis method.
Square wave analysis has good mutation and periodicity. By observing the output signal waveform of the device to the square wave signal, the various performances of the device can be well detected. Therefore, the square wave signal has become the most commonly used time domain analysis signal. Figure 1 is a specific description of the response signal of the audio device to the square wave within half a cycle (rising edge). The most important parameters for describing the square wave response are rise time, peak oscillation, overshoot, and slope.

Step signal analysis is relatively simple and is mainly used to detect the response sensitivity of audio equipment to signal mutations. There are usually two parameters for step signal analysis, namely the rise time and pulse width of the step response signal. The smaller the rise time, the more sensitive the device is to the signal mutation and the better the transient characteristics; the smaller the pulse width, the better the damping characteristics of the device and the more stable the system.
The peak value of a sinusoidal signal suddenly rises at a certain moment, forming a mutation, which is a single-tone mutation signal. Since the energy of a single-tone mutation signal is concentrated in a very narrow frequency range, a single-tone mutation signal is often used to detect the response of an audio device at a specific frequency. The main purpose of a single-tone mutation signal is to quickly determine the damping characteristics of certain audio devices, such as speakers.


3. Frequency domain analysis
Frequency domain analysis is an important part of audio analysis. The main basis of frequency domain analysis is the frequency response characteristic curve. The sine detection, pulse detection and maximum length sequence signal detection mentioned above can all obtain the frequency response of the device. The frequency response curve reflects the distribution of the frequency response of the audio device in the entire audio range. Generally speaking, the frequency component at the peak of the curve has a large playback sound pressure and a strong sound pressure; the frequency component at the bottom of the curve has a small sound pressure and a weak sound. If the peaks and troughs fluctuate too much, it will cause serious frequency distortion.
4. Time-frequency analysis
The time-frequency characteristics describe the changes in the frequency domain characteristics of the audio device on the time axis as time changes. The time-frequency characteristics not only describe the response state of the audio device in the process of frequency changes, but also describe the response state of the audio device in the process of time changes, that is, comprehensively describe the response characteristics of the audio device from a three-dimensional perspective. For playback equipment, the subjective listening evaluation, such as whether the bass is clean, whether the background is clear, whether the layers are clear, and the depth of the sound field, are closely related to the time-frequency characteristics of the audio device. The time-frequency characteristics of audio equipment are a very important aspect of objectively evaluating the performance of audio equipment.
5. Distortion analysis
The distortion of audio equipment includes harmonic distortion, intermodulation distortion, phase distortion and transient distortion. The most important thing in audio measurement is harmonic distortion. Harmonic distortion, in simple terms, is the extra harmonic components after the sound signal is reproduced by the audio equipment. From the audience's point of view, the sound emitted by different sound-emitting objects is composed of different fundamental waves and harmonic waves, and the audience can distinguish the sound-emitting objects according to the characteristics of the sound. If the power amplifier amplifies the music sound emitted by a certain instrument (the music sound is composed of fundamental waves and harmonic waves), and after the sound is played by the speaker, the waveform shape, amplitude and phase of the fundamental wave and each harmonic can be reproduced without distortion, it can be considered as high-quality playback; otherwise, the sound emitted by the speaker sounds irritable and awkward, and the harmonic distortion has reached an unbearable level, and even makes it impossible for people to distinguish the type of sound-emitting instrument. Therefore, harmonic distortion is an important performance indicator of audio equipment.
There are two methods for measuring harmonic distortion. One is to input a sinusoidal signal into the device to be tested, and then analyze the frequency components of the device response signal to obtain harmonic distortion. Another simpler measurement method is to first use a band-stop filter to filter out the fundamental frequency component in the response signal, and then directly measure the voltage of the remaining signal, and compare it with the original response signal to obtain harmonic distortion. Obviously, the harmonic distortion obtained by the second method is THD+N. Since the total voltage value of the signal is used instead of the voltage value of the fundamental frequency component, the harmonic distortion obtained is smaller than the actual value, and the greater the actual harmonic distortion, the greater the error.
In actual audio measurement, several frequency points are usually selected within a certain frequency range, and the harmonic distortion of each point is measured separately. Then, the harmonic distortion values ​​are connected into a curve with frequency as the horizontal axis, which is called a harmonic distortion curve. Figure 2 is a graph of the total harmonic distortion and 2nd, 3rd and 4th order harmonic distortion of a power amplifier in the range of 100-1OKHz.

4. Audio analysis instruments The
audio analysis instruments mentioned here refer to analytical instruments that can measure various electroacoustic parameters of various single audio devices such as microphones, audio amplifiers, speakers, etc., and can also test the overall performance of combined audio devices such as combination audio and mixing consoles. At present, various analytical instruments that can be used to measure audio equipment have appeared on the market, such as distortion analyzers, spectrum analyzers, frequency counters, AC voltmeters, DC voltmeters, audio oscilloscopes, etc. These rack-mounted hardware instruments based on various functional circuits are easy to use and have high measurement accuracy. They have been widely used. Audio equipment manufacturers can use audio analysis instruments to check the performance of equipment and find out the defects, so as to improve the design and manufacture of equipment. Consumers can also use audio analysis instruments to evaluate equipment and choose suitable products. Taking combination audio as an example
, the term "timbre" is often used when evaluating its performance. The so-called timbre refers to the sound difference caused by different high-order harmonics of the sound. The so-called "sense of balance" of the sound refers to the degree to which the volume of the sound reproduced in the full frequency band sounds natural. The role of audio analysis instruments is to express various industry terms for evaluating equipment in the form of various quantitative characteristic parameters. The characteristic parameter corresponding to "tone" is the measurement of harmonic distortion, while "sense of balance" involves the distribution of the frequency response of the equipment in the entire audio range.
Generally speaking, a relatively complete audio analysis instrument should be able to measure signal AC and DC voltage, signal frequency, harmonic distortion, signal-to-noise ratio and other parameters. Powerful audio analysis instruments provide functions such as spectrum analysis, 1/3 octave analysis, octave analysis, and sound pressure level measurement. If you want to build an audio analysis system, you also need a standard audio signal generator as an excitation signal source.


5. Current status of audio analysis instruments
In the early days, there were few types of professional audio analyzers. When doing audio measurements, a multimeter, frequency meter, oscilloscope and spectrum analyzer were generally used to form an audio test system. This test system has many intermediate links, and the interface matching between each link is more difficult. It is troublesome to use, and the measurement results are often inaccurate.
The audio analysis instruments that have emerged in recent years are also consistent with the mainstream development trend of instruments, and are developing in the direction of high integration and intelligence. These instruments integrate complex audio signal generation devices, power amplifiers, etc., and have some preliminary graphical analysis functions, making it easy for users to build audio measurement systems. RS's "Elf Series" audio analyzer UP300&UP350 is a typical representative of this type of instrument, as shown in Figure 3. The instrument integrates AC and DC voltage measurement, frequency measurement, signal-to-noise ratio measurement, and can also measure frequency response, level linearity or harmonic distortion. It can measure bidirectional channel crosstalk and generate dual-tone signals for modulation distortion analysis and difference frequency distortion measurement. It has a wide measurement range and high accuracy. In addition to general audio parameter measurements, the UP350 can also analyze digital audio signals with a sampling rate of up to 192kHz, and can be used for digital audio device measurements and analog/digital hybrid interface related applications.

VI. Summary
Audio analysis uses time domain analysis, frequency domain analysis, distortion analysis and other methods to evaluate the performance of the audio system by measuring various audio parameters. Audio analysis is a comprehensive analysis involving many test instruments. For ordinary users, it is difficult to establish a complete audio test analysis system because they need to reasonably select test instruments based on the parameters they are interested in.

Keywords:audio Reference address:Audio Analysis Principles

Previous article:How to measure electric energy (KWh) with F43B
Next article:Introduction to Virtual Instruments

Latest Analog Electronics Articles
Change More Related Popular Components

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号