JP4382626B2 - Audio signal level control apparatus and method - Google Patents

Description

  In order to transmit an audio signal within a certain dynamic range, signal level compression or expansion processing is performed. When compression or expansion processing is performed on such a level, distortion of the audio signal or an abnormal sense of hearing may occur. A level control method that eliminates as much as possible is required. As an example of such a level control system, there is a technique disclosed in Patent Document 1, and FIG. 5 shows a functional block diagram thereof.

  In the figure, the input audio signal is divided into two by the input unit 101 into a control signal and a controlled signal. The control signal among the divided audio signals is guided to the control interval determination unit 102, and the control signal is used for each interval or a plurality of intervals between the points where the signal waveform of the audio signal intersects the zero-level time axis. The control section is set sequentially by dividing the signal.

  On the other hand, the controlled signal is guided to the storage delay unit 105 and stored by being divided into sections corresponding to the control sections described above under the control of the control section determination section 102. In addition, the control signal divided for each control section in the control section determination unit 102 is guided to the control information extraction unit 103 and is used for level information used for controlling the signal level in each control section, for example, the peak of the signal amplitude. A value or a mean square value is extracted for each control section.

  The extracted output from the control information extraction unit 103 is guided to the compression (decompression) ratio determination / control signal creation unit 104, and the compression (decompression) ratio set in advance corresponding to the signal level value is input. Select the compression (decompression) ratio of the value corresponding to the extracted level information, and control by corresponding to the compression (decompression) ratio and forming a constant level control signal that does not change within each control section Guide to section 106.

  On the other hand, the controlled signal divided in correspondence with the control section and stored in the storage delay unit 105 is also read out in correspondence with the start of the control signal from the compression (decompression) ratio determination / control signal creation unit 104, and is also controlled by the control unit 106. Lead to. The control unit 106 multiplies the level control signal and the controlled signal described above, thereby completing the level control of the input audio signal for each sequential control section. The multiplication output from the control unit 106 is guided to the output unit 107 and temporarily stored, and then sequentially extracted as a level control output signal.

  As described above, the input audio signal is divided into sequential control intervals, and is subjected to level control that is constant and unchanged within each control interval. It occurs only at the boundary of each control section that becomes zero level. Therefore, the signal waveform in the sequential control interval of the level control output audio signal is completely similar to the signal waveform of the input audio signal, and the change of the signal waveform occurs only at the boundary of each control interval where the signal waveform itself becomes zero level. As a result, there is no abnormality in hearing. That is, the level-controlled output audio signal waveform is zero level at the boundary of each control section, and only the slope of the signal waveform before and after that is different from the input audio signal waveform. And the feeling of abnormality is greatly reduced.

  An example of the signal waveform shown in FIG. 6 will be described with respect to the aspect of the sound signal level control according to the present method described above. Between the signal waveform in the input sound signal waveform (A) and the sequential crossing point (zero cross point) of the zero-level time axis. For example, when the peak level is extracted as level information, a waveform indicated by a solid square in the level control information (B) is obtained. Thus, with respect to the number of periods (zero-cross sections) divided for each zero-cross point of the input audio signal, for example, the signal waveform is divided every four periods to make the above-described control section, and four of the above-described control sections in the sequential control section. If the highest level of the solid line squares is used as the level information in each control section, the square waveform shown by the dotted line of the level control information (B) is obtained.

  For example, when level compression is performed on an audio signal, the higher the signal level is, the larger the compression rate is. Therefore, the signal level is controlled in the opposite relation to the level of the dotted rectangular waveform described above. A control signal (C) corresponding to the level control information (B) is obtained. When the original signal level is compressed in proportion to the size corresponding to each control section of the control signal (C), the input signal indicated by the dotted line in the level control output signal waveform (D) is compressed as indicated by the solid line, and the original signal level is compressed. The higher the level of the signal, the greater the compression, but the entire output signal waveform is almost similar to the input signal waveform, and no distortion or abnormal change is observed.

Further, as a technique improved from the above-described audio signal level control method, there is a technique disclosed in Patent Document 2.
Japanese Patent Publication No.57-41861 JP 59-2221017

  In the conventional technology described above, in order to always hold the audio data for the so-called zero-cross section, which is a segment of the audio signal at the zero-cross point of the signal (the point where the signal waveform intersects the zero-level time axis). A storage delay unit (memory) is required. Here, the transmission frequency band of the audio signal is generally a band of 20 Hz to 20,000 Hz. Therefore, as shown in FIG. 7, if the audio delay of the zero cross section in the audio of 20 Hz is held, 25 ms (1/20 Hz). There will always be a voice delay of half the period. In addition, it is necessary to have a memory necessary for the delay. When the frequency band is expanded to a lower component, further audio delay occurs.

  As shown in FIG. 6, the control unit is a zero-cross section, and in order to detect the section, a voice delay depending on the lowest frequency of the transmission frequency band is generated. In addition, as shown in FIG. 6, there is a control section that crosses the zero-cross section (two periods), and more delays occur (a delay memory is required) to realize this.

An object of the present invention is to provide an audio signal level control device and an audio signal level control method capable of minimizing audio delay and minimizing audio distortion.

According to the present invention, an audio signal level control device forming the level control of the audio signal,
Frequency separating means for separating the frequency of the audio signal into a high frequency component and a low frequency component;
When the high frequency component is larger than the low frequency component, first control means for controlling to a predetermined constant level for each at least one zero cross section of the audio signal;
Second control means for dynamically processing level control of the audio signal when the low frequency component is equal to or higher than the high frequency component;
The sound level control apparatus characterized by including this is obtained.

According to the present invention, an audio signal level control method for accomplishing the level control of the audio signal,
A frequency separation step of separating the frequency of the audio signal into a high frequency component and a low frequency component;
A first control step of controlling to a predetermined constant level for each at least one zero cross section of the audio signal when the high frequency component is greater than the low frequency component;
A second control step of dynamically processing level control of the audio signal when the low frequency component is equal to or higher than the high frequency component;
An audio signal level control method characterized by comprising:

  The operation of the present invention will be described. When the high frequency component of the frequency component of the audio signal is large, the delay is small, so that the level control is performed with a constant multiplier corresponding to the peak value in the zero cross section, for example, as in the conventional case. When the band component becomes large, level control is performed by dynamic real-time processing so as not to cause a delay. This enables voice level control with little delay and little distortion.

  According to the present invention, when the high frequency component level is large as the frequency component of the audio data, the zero cross section data is multiplied by a predetermined constant, and when the low frequency component level is large, By performing the dynamic processing, there is an effect that the voice delay can be minimized and the voice distortion can be minimized.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the present invention. In the figure, HPF1 and LPF2 are a high-pass filter and a low-pass filter for performing frequency division on the input digital audio signal a, and the frequency characteristics are as shown in FIG. The zero-cross detection unit 3 receives the high-frequency component audio data b that has passed through the HPF 1 and outputs a zero-cross detection signal e according to a change in the sign flag of the audio data. Note that the sign flag is a flag at the head of data for indicating the sign of the data.

  The peak data detection unit 4 detects the peak value data d in the zero cross section, and the peak data detection unit 5 receives the low-frequency component audio data c that has passed through the LPF 2 as input, and obtains peak data (peak value) f. It is to detect. The multiplier calculation unit 7 uses the outputs d, e, and f of the peak data detection units 4 and 5 and the zero cross detection unit 3 and the ATK (attack) constant g and REL (release) constant h supplied from the outside. A multiplier j for multiplying the audio data is calculated. The delay memory 6 delays the audio data a for the time T required to calculate the multiplier j. The arithmetic unit 8 multiplies the output i of the delay memory 6 by a multiplier j and outputs audio data k that is level-compressed.

  The input digital audio signal a is input to the HPF 1 and LPF 2 for performing frequency division and also input to the delay memory 6. The audio signal a passes through each filter according to the frequency component by the HPF 1 having the cutoff frequency fCH and the LPF 2 having the cutoff frequency fCL shown in FIG. Zero cross section control is performed on the audio data b that has passed through the HPF 1. Therefore, the zero cross detection unit 3 outputs the zero cross detection signal e to the multiplier calculation unit 7 due to the change of the sign flag of the audio data b.

  Until the zero-cross detection signal e is output, the peak data detection unit 4 constantly updates the largest value d in the zero-cross detection interval. At the generation timing of the zero-cross detection signal e, the multiplier calculation unit 7 Peak data d is taken in. On the other hand, the peak data detection unit 5 detects the peak data f of the voice data c that has passed through the LPF 2 and outputs it to the multiplier calculation unit 7.

  The multiplier calculation unit 7 selects a larger value of the peak data at the zero-cross timing e from the zero-cross detection signal e, the peak data d of the HPF region, and the peak data f of the LPF region, and multiplies the audio data by the multiplier. j is determined. Now, assuming that the peak data d in the HPF region is larger than that f in the LPF region, the same multiplier j that is predetermined according to the peak data e is output for the audio data in the zero-cross section. That is, in the case of d> f, the high frequency component is large and the delay of the audio data does not become large. Therefore, the same level control as the conventional one explained with reference to FIGS.

  When the peak data f in the LPF region is equal to or higher than the peak data d in the HPF region, that is, when f ≧ d, the multiplier j is dynamically determined and output using the ATK multiplier g and the REL multiplier h. The attack constant, which is an ATK constant, is a constant that determines the time until the suppression amount reaches 90% when the audio data is suppressed when the audio data of a certain threshold level or higher is input. The release constant, which is a REL constant, is a constant that determines the time until the suppression amount is released by 90% when the suppression amount is released from the suppression state when the voice data is below the threshold level. Using these constants g and h, voice data suppression control is performed dynamically, that is, in real time.

  FIG. 3 is a block diagram of another embodiment of the present invention, and the same parts as those in FIG. 1 are denoted by the same reference numerals. In the configuration of FIG. 1, the multiplier calculation unit 7 compares the frequency components of the high frequency and low frequency of the audio data a. In FIG. 3, a level comparison unit 9 is provided, and this level comparison unit is provided. 9, the HPF output level b and the LPF output level c are compared to determine which frequency band contains a lot of audio data components. By supplying this determination result m to the multiplier calculation unit 7, when f ≧ d, the multiplier calculation unit 7 can execute dynamic processing without waiting for the zero-cross signal e, thereby obtaining higher speed. It is done.

  FIG. 4 is a block diagram of still another embodiment of the present invention, and the same parts as those in FIGS. 1 and 3 are denoted by the same reference numerals. In this example, frequency characteristics can be set from the outside with respect to HPF1 and LPF2 in FIG. 3, and the cutoff frequencies fCH and fCL shown in FIG. 2 can be arbitrarily set by external control signals n and p. It is designed to be variably controlled. As a result, it is possible to perform zero-cross section control up to a band that can tolerate the maximum voice delay. In the example of FIG. 1 as well, the frequency characteristics of the HPF 1 and the LPF 2 can be controlled as well.

  In the above embodiment, when audio data has a lot of high frequency components, a constant level control corresponding to the peak data of the zero cross section is performed, but a constant level control according to the mean square value of the zero cross section is performed. In addition, it is obvious that the level control may be performed in units of a plurality of zero cross sections instead of one zero cross section as a unit.

It is a block diagram which shows one embodiment of this invention. It is a figure which shows the frequency characteristic of HPF1 and LPF2 of FIG. It is a block diagram which shows other embodiment of this invention. It is a block diagram which shows another embodiment of this invention. It is a block diagram for demonstrating a prior art. It is a figure which shows operation | movement of the block of FIG. It is a figure explaining the problem of a prior art.

Explanation of symbols

1 HPF
2 LPF
3 Zero cross detection unit 4, 5 Peak data detection unit 6 Delay memory 7 Multiplier calculation unit 8 Calculation unit 9 Level comparison unit

Claims (6)

An audio signal level control device forming the level control of the audio signal,
Frequency separating means for separating the frequency of the audio signal into a high frequency component and a low frequency component;
When the high frequency component is larger than the low frequency component, first control means for controlling to a predetermined constant level for each at least one zero cross section of the audio signal;
Second control means for dynamically processing level control of the audio signal when the low frequency component is equal to or higher than the high frequency component;
An audio level control apparatus comprising: The frequency separation means includes a high-pass filter that passes the high-frequency component of the audio signal, and a low-pass filter that passes the low-frequency component,
2. The audio level control apparatus according to claim 1, wherein the first control means controls the audio signal to a predetermined level that is predetermined according to peak data of the zero-cross section.   3. The sound level control apparatus according to claim 2, wherein each frequency characteristic of the high-pass filter and the low-pass filter is changeable. An audio signal level control method for accomplishing the level control of the audio signal,
A frequency separation step of separating the frequency of the audio signal into a high frequency component and a low frequency component;
A first control step of controlling to a predetermined constant level for each at least one zero cross section of the audio signal when the high frequency component is greater than the low frequency component;
A second control step of dynamically processing level control of the audio signal when the low frequency component is equal to or higher than the high frequency component;
An audio signal level control method comprising: The frequency separation step is performed by a high-pass filter that passes the high-frequency component of the audio signal and a low-pass filter that passes the low-frequency component,
5. The audio signal level control method according to claim 4, wherein in the first control step, the audio signal is controlled to a predetermined level that is predetermined according to peak data in the zero-cross section.   6. The audio level control method according to claim 5, further comprising the step of making each frequency characteristic of the high-pass filter and the low-pass filter freely changeable.

Images