JP4804376B2 - Audio equipment - Google Patents

Description

  The present invention relates to an audio device that generates a multi-channel audio signal from a two-channel stereo audio signal, and more particularly to an audio device that can improve an input signal to generate a clear multi-channel audio signal.

  Conventionally, various techniques for generating a multi-channel audio signal from a 2-channel stereo audio signal have been proposed. For example, as shown in FIG. 5, the left and right input signals L and R are output as L and R as they are. A technique is also used in which both signals are added in the opposite direction, + and-, respectively, so that LR is the left surround output signal SL and RL is the right surround output signal SR. However, in this technique, since the left surround output signal SL and the right surround output signal SR are in opposite phases, the reproduction gives a sense of opposite phase to the user, and there is a sense of incongruity during reproduction, which is unnatural. There was a problem of giving an impression.

  As a countermeasure, a technique as shown in FIG. 6 has been proposed (Patent Document 1: Japanese Patent Laid-Open No. 2003-333698; Patent No. 3682032). The present invention mainly corresponds to a technique obtained by further improving this technique, and the details thereof are described in detail in Patent Document 1, but the outline is as follows. That is, FIG. 6 shows an embodiment in which this technique generates a 4-channel signal from a 2-channel stereo signal, and L and R, which are 2-channel stereo audio signals, are input. The input signals L and R generate four-channel signals L, R, SL, and SR to be output.

  Among the four-channel signals that are output, the output signals L and R output the input signals as they are. L and R are input to the adaptive decorrelator 1L, and a signal to be ASL is generated. The signal ASL is passed through the band limiting filter 2L and the delay processor 3L to perform band limiting and delay processing, and then set to the left surround output signal SL. On the other hand, L and R are also input to the adaptive decorrelator 1R to generate an ASR signal, and this signal ASR is passed through the band limiting filter 2R and the delay processor 3R to perform band limiting and delay processing. After that, the right surround output signal SR is obtained. This technology is based on the processing method as described above, and surrounds of more channels such as 5 channels and 5.1 channels can be generated by further processing other channels.

As the adaptive decorrelator used here, for example, adaptive signal processing using an FIR filter can be performed. FIG. 7 shows a configuration example of an adaptive decorrelator that employs adaptive signal processing using the FIR filter. This adaptive decorrelator includes an input terminal for an input signal Y on the addition side and an input terminal for an input signal X on the subtraction side, and an output terminal for an output signal O that becomes a surround signal. The input signal Y on the addition side is input to the arithmetic unit A via the delay processor Z- ^m .

On the other hand, the input signal X on the subtraction side is sequentially subjected to delay processing by delay processors Z- ¹ provided in multiple stages constituting the FIR filter, and then RES (n) = X (n) ^T W (n) Is superimposed on a predetermined coefficient in a coefficient processor W having W0, W1,..., Wk as elements. . Here, k represents a tap length (the number of delay processes).

The response signal RES obtained in this way is input to the arithmetic unit 4, and the response signal RES is subtracted from the input signal Y of the other channel that is also input to the arithmetic unit 4, so that the error signal e and the output signal O are can get. The coefficient processor W is updated by a coefficient update processor B having an adaptive algorithm so as to extract a component having a high correlation with the component of the input signal Y from the components of the input signal X. That is, the coefficient update processing unit B receives the input signal X and the error signal e from the arithmetic unit A every moment, and the input signal X and the error signal e are processed by the update algorithm, so that the coefficient update processing is performed. Coefficient update commands are output from the processor B to the coefficient processors W0, W1,..., Wk at each stage, and based on this, the coefficient superposed on the output signal from the delay processor Z- ^{1 at} each stage is output. The value changes.

  For example, an LMS algorithm is employed as an update formula employed in such a coefficient update processor B. This is an algorithm using the instantaneous square error as an evaluation quantity, and the coefficient processor W is updated by the equation W (n + 1) = W (n) + 2μ · e (n) · X (n). Here, μ is a step size parameter and is a quantity that greatly affects the performance of the adaptive decorrelator realized.

The adaptive decorrelators 1L and 1R shown in FIG. 6 as described above perform adaptive signal processing by the FIR filter updated by the step size parameter μ of the LMS algorithm as shown in FIG. B corresponds to an adaptive algorithm. When this is incorporated into the surround signal generator of the audio apparatus, the result is as shown in FIG. That is, when the left and right input signals L and R are directly output as the left and right output signals L and R in FIG. 8, and the surround signals are generated by the left and right surround signal generators L10 and R10, the FIR shown in FIG. The adaptive signal processing is performed by a filter, and FIG. 8 shows a generalized processing system of FIG. Therefore, here, a 2-channel stereo signal is input, and for each channel input signal, the difference between one input signal processed by an adaptive control unit having an adaptive algorithm and the other input signal is calculated. An audio apparatus comprising: a surround generation unit that updates an adaptive filter coefficient (ADF coefficient) according to the difference, the one input signal, and a step size parameter in the adaptive control unit, and outputs the difference as left and right surround signals, respectively. The present invention assumes this configuration. The surround signal obtained as described above can solve the problems of the prior art, can be a natural surround system without a sense of reverse phase, and is attracting attention as AST (Adaptive Surround Technology).
JP 2003-333698 A (Patent No. 3682032)

  On the other hand, at the product design stage of the current audio apparatus, the maximum value of the input level of the 2-channel signal to the DSP is approximately 1 Vrms = 0 [dB] for CD, 0.14 Vrms = −17 [dB] for AM tuner, FM The tuner is 0.14 Vrms = −17 [dB], and considering the CD as a reference, the maximum input level of the tuner is relatively −17 [dB] lower.

  In this regard, in the AST surround system as described above, the input level necessary for constantly updating the coefficient of the adaptive filter serving as the AST core, that is, the lower limit value of the input level is related to the step size parameter μ. It can be calculated as a theoretical value. In the adaptive filter coefficient update equation, W (n + 1) = W (n) +2 μ · e (n) · X (n), when the adaptive filter coefficient update stops, that is, μ · e (n) · X ( n) = 0 is the condition. Here, the minimum value 0x000001 (24-bit expression) of the DSP value that can be expressed by a fixed point is the closest value to “0”. Further, when e (n) = X (n) and μ is variable, the lower limit value of the input level for each μ is derived.

  This is shown in FIG. 2, and in the adaptive control using the LMS algorithm as described above in FIG. 2 (a), the step size parameter μ in the LMS algorithm is as shown in FIG. 2 (b). = 0.001 −39.25 [dB], μ = 0.0001 -29.24 [dB], μ = 0.00001 -19.24 [dB], When μ = 0.0000001, it is −9.24 [dB].

  Therefore, as described above, the input level of the DSP of the two-channel signal in the audio apparatus is set to a low maximum input level particularly in an audio source such as a tuner. Therefore, the adaptive filter coefficient is updated in the AST surround system. It will not be done and its effectiveness will be reduced.

  As a result, even if the same music is used, the surround effect differs between the CD and the tuner even if the same audio device is used. In addition, since the surround effect varies depending on the input level of the audio signal, the surround effect is reduced for music with a low input signal level. In addition, even for music with a wide dynamic range such as classical music, the sound collection level is extremely low compared to the maximum value depending on the musical instrument such as a violin, and the surround effect by the AST may be similarly reduced. .

  Therefore, the present invention can sufficiently exhibit the surround effect even when the input signal level is low, in an audio device that generates the surround effect by adaptive control including an adaptive algorithm such as the LMS algorithm, and the surround due to the difference in the input signal level. An object of the present invention is to provide an audio apparatus capable of performing processing so as not to differ.

  The audio apparatus according to the present invention inputs a 2-channel stereo signal, and for each channel input signal, the difference between the signal obtained by processing one input signal by an adaptive control unit having an adaptive algorithm and the other input signal In an audio apparatus comprising: a surround generation unit that calculates an adaptive filter coefficient according to the difference, the one input signal, and a step size parameter in the adaptive control unit, and outputs the difference as left and right surround signals A dynamic range compression unit for compressing the dynamic range of the input signal provided on the signal input side of the surround generation unit, and control for increasing the compression in the dynamic range compression unit as the level of the input signal is lower And a section.

  Another audio device according to the present invention is characterized in that, in the audio device, the control unit performs control to perform greater compression as the value of the step size parameter is smaller.

In another audio device according to the present invention, in the audio device, the control unit performs an operation of X = 1− (1− | x |) ^{n on} the output signal X when the input audio signal is x. The compression characteristic is added by making the value of n larger as the level of the input signal is smaller.

  Another audio device according to the present invention is characterized in that, in the audio device, in addition to the dynamic range compression processing, the control unit also performs control to increase an input gain when an input signal level is a predetermined value or less. .

In another audio device according to the present invention, in the audio device, in addition to the dynamic range compression process, the control unit performs a process of increasing a value of the step size parameter when an input signal level is a predetermined value or less. It is characterized by that.
This is the main feature.

  By configuring the present invention as described above, the surround effect can be sufficiently exerted even when the input signal level is low, in the audio device that generates the surround effect by the adaptive control provided with the adaptive algorithm. Further, it is possible to perform processing so that the surround effect due to the difference in the input signal level is not different, and it is also possible to listen to music with a low sound collection level clearly and with good surround effect even in a noisy environment. Furthermore, the surround effect for MP3, AAC, and WMA compressed audio can be improved.

  The audio device according to the present invention is an audio device that generates a surround effect by adaptive control provided with an adaptive algorithm, and has the purpose of sufficiently exhibiting the surround effect even when the input signal level is low. A stereo signal is input, and the difference between a signal obtained by processing one input signal by an adaptive control unit having an adaptive algorithm and the other input signal is calculated for the input signal of each channel, and the adaptive control unit In an audio apparatus including a surround generation unit that updates an adaptive filter coefficient according to a difference, the one input signal, and a step size parameter, and outputs the difference as left and right surround signals, respectively, on the signal input side of the surround generation unit Dynamic that compresses the dynamic range of the input signal And Nji compression unit is obtained by realized by a control unit for compressing the increased control at the lower level of the input signal the dynamic range compressor.

  FIG. 1 shows an embodiment of an audio apparatus according to the present invention. In the figure, an AST core surround processing unit 2 is provided as an AST (Adaptive Surround Technology) processing unit that performs audio surround processing using adaptive control having the above-described adaptive algorithm, and this AST core surround processing is provided. A signal input unit to the unit 2, that is, a dynamic range compression unit 1 is provided in front of the processing unit. The dynamic range compression unit 1 is incorporated in an audio processing device (DSP) and is a controller as a function of a microcomputer. It is controlled by the control unit 3.

  The control unit 3 performs the coefficient control of the step size parameter μ used for the adaptive filter of the adaptive algorithm in the AST core surround processing unit 2, and particularly when the compression processing as described later in the dynamic range compression unit 1 cannot sufficiently deal with, Control is also performed to increase the input level by a predetermined amount, which is partly performed conventionally. For the dynamic range compression unit 1, the compression rate is controlled according to the input level. Various methods conventionally proposed can be used as the method of the dynamic range compression rate at this time.

FIG. 3 shows an example of a dynamic range compression processing technique, which is a technique previously proposed by the present applicant as a compression processing means for dealing with noise levels (Japanese Patent Laid-Open No. 5-344078). ). In the dynamic range compression example shown in FIG. 3, when the input audio signal is x, the DSP calculates x = (1− | x |) ⁿ , adds compression characteristics, and uses the order switching means to reduce the noise level. The output value X is generated by changing the compression characteristic by changing the value of the order n according to. Hereinafter, the previously proposed compression method will be briefly described.

  In the example shown in FIG. 3, as shown in FIG. 3A, whether the input signal x is positive or negative is determined (11). When x ≧ 0, a = 1, and when x <0, a Is output (= 12) and is output to the multiplier 17. On the other hand, the processing signal y1 = x is set using the input signal x (13), then y2 = 1−y1 is calculated (14), and this is output to the power calculation unit 15. The power calculation unit 15 is switched by the processing of the order switching unit 20, and the order switching unit 20 can perform switching according to ambient noise, so that ambient sound collected by the microphone 21 is detected as a noise level. The noise level is detected by the device 22 and is output to the order switching circuit 23.

  The order switching circuit 23 switches the order by dynamic range compression characteristics as shown in FIG. That is, the logarithmic display input / output characteristics are indicated by a solid line E, and the conventional analog input / output characteristics are shown as A and the digital input / output characteristics as B. As is clear from this, in this technique, the order n value can be varied based on the noise level, thereby making it closer to the analog input / output characteristic A, and the order n is determined based on the detected noise level. By switching, a compression characteristic having an optimal compression ratio according to the noise level can be added to the audio signal.

The order switching circuit 23 operates by such an order switching method, and the power calculation unit 15 calculates the power according to the switched order to obtain the output signal y3. Thereafter, y4 is obtained by the calculation of y4 = 1−y3, which is output to the multiplier 17 and multiplied by a which is +1 or −1 depending on whether the input signal x is positive or negative. Thereby, when x ≧ 0, X = 1− (1− | x |) ⁿ , or when x <0, the audio signal X calculated by X = −1 + (1− | x |) ⁿ is output. The relationship between X and x at this time is similar to the analog method as described above.

  The dynamic range compression unit 1 shown in FIG. 1 employs the above-described compression method and is applied to the present invention, for example, as shown in FIG. That is, in FIG. 4A, the dynamic compression range process substantially similar to that in FIG. 3 is performed, but here a step size parameter (μ) detection unit is used instead of the order switching unit 20 in FIG. 3A. The order switching unit 24 having 25 is provided. The step size parameter (μ) detector 25 detects the value of the step size parameter (μ) in the LMS algorithm in the example shown in FIG. 8 in the AST core surround generator 2 of FIG.

  When the order switching circuit 26 detects that the compression ratio is set, the order switching data is obtained, for example, with the dynamic range compression characteristics shown in FIG. Set a power value like this, and change the compression characteristics. The example of FIG. 4B shows an example in which processing by this compression characteristic is performed when the step size parameter (μ) becomes smaller than the first predetermined value a, and the second smaller than this is shown. When the value is smaller than e, which is a predetermined value, the control unit 3 performs other processing such as input level control.

  In the example of FIG. 4B, in the characteristics of the straight line C1, processing is performed in which the compression ratio is increased in inverse proportion to the step size parameter (μ) from a or inversely proportional to the logarithm. C2 in the figure shows an example in which the compression rate is increased stepwise. The power multiplier is set to 2 when μ is b, the power multiplier is 3 when μ is c, and the power multiplier when μ is d. 1, and when μ becomes smaller than that, outside the range indicated by the line D in the figure, the control unit 3 in FIG. 1 performs other control such as input level control as described above. The characteristic of the curve C3 in FIG. 4B shows an example in which the step size parameter μ is gradually increased from the time a, and the compression ratio increase rate is set higher as μ decreases. Thus, the value of the step size parameter μ and the compression characteristic of the dynamic compression range can be set arbitrarily, and processing with a high degree of freedom is possible.

  When performing input level control as another control in the control unit 3 outside the range indicated by the line D in FIG. 4B, various conventional methods can be employed. When the input level falls below a predetermined level, a method of adding + α to the input signal or multiplying by × β may be employed. Further, as a method for dealing with a particularly low input level, in addition to the method for increasing the input gain as described above, the value of the step size parameter μ may be forcibly increased. This method is a method for increasing the input gain. It can also be used together or independently.

It is a functional block diagram of the Example of this invention. It is a figure which shows the theoretical value of the relationship between the lower limit of an input level, and the value of a step size parameter. It is a figure which shows the example of dynamic range compression corresponding to the noise level proposed previously. It is a figure which shows the dynamic range compression example corresponding to a step size parameter using the function of the said dynamic range compression example. It is a figure which shows the example of the conventional surround output. It is a basic block diagram which performs the conventional AST surround process. It is a figure which shows the process of the adaptive decorrelator part in the process. It is a functional block diagram when the same process is applied to an audio processing part.

Explanation of symbols

1 Dynamic range compression unit
2 AST core surround generator
3 Control unit

Claims (5)

A two-channel stereo signal is input, and a difference between a signal obtained by processing one input signal by an adaptive control unit having an adaptive algorithm and the other input signal is calculated for each channel input signal, and the adaptive control unit In the audio apparatus comprising a surround generation unit that updates the adaptive filter coefficient according to the difference, the one input signal, and the step size parameter, and outputs the difference as left and right surround signals,
A dynamic range compression unit for compressing the dynamic range of the input signal, provided on the signal input side of the surround generation unit;
An audio apparatus comprising: a control unit that performs control to increase compression in the dynamic range compression unit as the level of the input signal is low.   The audio apparatus according to claim 1, wherein the control unit performs control such that compression is increased as the value of the step size parameter is decreased. The control unit performs an operation of X = 1− (1− | x |) ⁿ for the output signal X when the input audio signal is x, and the value of n increases as the level of the input signal decreases. The audio apparatus according to claim 1, wherein a compression characteristic is added.   2. The audio apparatus according to claim 1, wherein the control unit performs control for increasing an input gain when the level of the input signal is a predetermined value or less in addition to the dynamic range compression processing.   2. The audio apparatus according to claim 1, wherein the control unit performs a process of increasing the value of the step size parameter when the input signal level is a predetermined value or less in addition to the dynamic range compression process.

Images