JP2008249971A - Speech signal processor, speech signal processing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a speech signal processor of which the processing capability can be well demonstrated in speech signal processing, and to provide a speech signal processing method and a program. <P>SOLUTION: The speech signal processor 1 performs interpolation processing by any of filters 11 to 14 which are different in characteristics for the speech signal input from input sections A to C. According to a load status onto a digital signal processor (DSP) 10, the filter for performing interpolation processing is switched by a CPU 15. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an audio signal processing device, an audio signal processing method, and a program.

The audio device receives input signals reproduced from various storage media. Examples of the storage medium include a CD (Compact Disk), an MD (Mini Disc), and a USB (Universal Serial Bus) memory. The reproduction signals from these storage media may be subjected to various types of filter processing such as sound range complementation and equalization as necessary.
These filters are configured by programs in a DSP (Digital Signal Processor).

For example, Patent Document 1 discloses a technique related to digital arithmetic processing performed by a filter to obtain a desired acoustic characteristic.
JP 2000-91865 A

  However, in the conventional audio signal processing apparatus, the same filter is applied to the filter processing for the same purpose regardless of the load state of the DSP. Therefore, depending on the selected filter load and the number of filters, the DSP processing is performed. In some cases, there is a vacancy and the DSP does not fully demonstrate its original processing capability.

  An object of the present invention is to provide an audio signal processing device, an audio signal processing method, and a program capable of sufficiently exerting the DSP processing capability in audio signal processing.

The invention according to claim 1 is an audio signal processing apparatus,
An audio signal input unit for inputting an audio signal;
An audio signal processing unit that performs audio processing on the audio signal input from the audio signal input unit and includes two or more audio processing units having different calculation loads for the same purpose audio processing content;
A control unit that performs control to switch the audio processing unit so as to preferentially select the audio processing unit having a large calculation load among the two or more audio processing units according to the processing load of the audio signal processing unit; It is provided with.

According to a second aspect of the present invention, in the audio signal processing device according to the first aspect,
The control unit performs control to switch the audio processing content by the audio signal processing unit in accordance with the type of the audio signal input from the input unit.

The invention according to claim 3 is the audio signal processing apparatus according to claim 1 or 2,
Two or more audio processing means having different calculation loads for the same purpose audio processing in the audio signal processing unit are different in the type of filter used.

According to a fourth aspect of the present invention, in the audio signal processing device according to the first or second aspect,
The two or more sound processing means having different calculation loads for the same purpose sound processing in the sound signal processing unit have different filter orders.

The invention according to claim 5 is an audio signal processing method,
An audio signal input process for inputting an audio signal;
The voice signal input step is a voice processing step for voice processing, and the voice signal processing unit having two or more voice processing means having different calculation loads for the same target voice processing content performs voice processing. An audio signal processing step for processing;
Control is performed to switch the sound processing means so as to preferentially select the sound processing means having a large calculation load among the two or more sound processing means according to the processing load by the sound processing in the sound signal processing step. And a control process.

The invention according to claim 6 is a program,
An audio signal input means for inputting an audio signal to a computer;
A voice signal processing means for performing voice processing on the voice signal input from the voice signal input means, and having two or more voice processing means having different calculation loads for the same purpose voice processing content;
Control means for performing control to switch the sound processing means so as to preferentially select the sound processing means having a large calculation load among the two or more sound processing means according to the processing load by the sound signal processing means. It is made to function.

  ADVANTAGE OF THE INVENTION According to this invention, the audio | voice signal processing apparatus, the audio | voice signal processing method, and program which can fully demonstrate the processing capability of DSP in an audio | voice signal processing can be provided.

  Hereinafter, a preferred embodiment of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to this embodiment.

[Constitution]
As shown in FIG. 1, the audio signal processing apparatus 1 includes input units A to C, a DSP 10, a DAC (Digital Analog Converter) 20, and a volume 30.

  The input units A to C are input devices for inputting audio signals, and various audio signals are input thereto. Here, as the type of the audio signal to be input, for example, in this embodiment, a PCM (Pulse Code Modulation) format audio signal input from the CD to the input unit A. Examples include an audio signal compressed in a format such as MP3 or WMA input to the input unit B, an audio signal converted into a digital format by an ADC (Analog Digital Converter) input to the input unit C, and the like. Note that the format of the input audio signal is not limited to these formats.

  The DSP 10 is a processing device for performing digital signal processing on an audio signal, and includes filters 11 to 14, a CPU 15, a load measuring unit 16, switches s1 to s7, delay elements d1 to d3, and adders t1 and t2. Configured.

  The filters 11 to 14 each perform sound processing on the input sound signal. Each filter is a filter having different characteristics, but the filter 11 and the filter 12 are filters that perform the same-purpose sound processing, and the filter 13 and the filter 14 are filters that perform the same-purpose sound processing.

For example, in the present embodiment, the filter 11 and the filter 12 perform high frequency interpolation processing. The high-frequency supplement processing is performed by, for example, estimating and supplementing a high-frequency region that is missing from a compressed sound source or the like. The filter 13 and the filter 14 perform processing for equalizing the audio signal.
Equalizing includes, for example, a bass enhancement process. In addition, the filter 12 performs processing with higher voice processing accuracy and a greater DSP load than the filter 11. Similarly, the filter 14 performs processing with higher accuracy of audio processing and a greater DSP load than the filter 13.

  In order to obtain a filter having the above characteristics, for example, in the present embodiment, an FIR (Finite Impulse Response) filter is used as the filter 11. As the filter 12, an FIR filter having a high-order tap is used. The filter 13 is an IIR (Infinite Impulse Response) filter. The filter 14 is an FIR filter.

  In general, the processing by the FIR filter can easily eliminate the phase shift because the phase changes linearly as compared with the IIR filter. Therefore, the load on the DSP increases. Also, increasing the number of filter taps improves the performance as a filter, but increases the load on the DSP.

  The CPU 15 reads out and expands a control program from a memory (not shown), and performs centralized control and various calculations of operations of each unit provided in the DSP 10 in cooperation with the control program.

  The load measurement unit 16 measures the load of processing performed by the DSP 10 and outputs the measurement value. For example, the load is measured by measuring a MIPS (Million Instructions Per Second) value that is a processing speed of the DSP 10. That is, it is measured by measuring the time for the DSP 10 to execute one instruction and calculating the number of instructions executed per second.

  The switches s1 to s7 are switches that switch whether or not to pass the input audio signal.

  The delay elements d1 to d3 are elements that delay the input audio signal by a predetermined period and output it. The adders t1 and t2 are elements that add and output a plurality of input audio signals.

The DAC 20 converts the digital audio signal output from the DSP 10 into an analog signal, and outputs the analog signal to an amplifier (not shown) via the volume 30.
It is a device to do.

  The volume 30 includes an amplifier and amplifies the audio signal input from the DAC 20 for volume adjustment.

[Operation]
Next, with reference to FIG. 2, a control flow by the CPU 15 when audio processing is performed by the DSP 10 will be described.

  First, input A, input B, or input C is selected according to the format of the input audio signal, and the audio signal is input to the DSP 10 (step S1).

  Next, when the sound signal input in step S1 needs to be adjusted in the low frequency range (step S2; Yes), the process proceeds to step S3.

  Next, when the high frequency range complementation is necessary for the audio signal input in step S1 (step S3; Yes), a normal low range adjustment filter is selected and low range adjustment is executed (step S4). Next, a normal high range complement filter is selected, and high range complement is executed (step S5). Details of the filter selection method will be described later (the same applies to the subsequent filter selection methods).

  Next, the audio signal processed by the selected filter is output (step S13), and this process ends.

  If the audio signal input in step S1 does not require high-frequency range interpolation (step S3; No), the process proceeds to step S6.

  Next, when there is a margin in the operation of the DSP 10 (step S6; Yes), a heavy bass adjustment filter is selected, a bass adjustment is performed (step S7), and the process proceeds to step S13.

  If there is no margin in the operation of the DSP 10 (step S6; No), a normal low range adjustment filter is selected, low range adjustment is executed (step S8), and the process proceeds to step S13.

  If the low frequency adjustment is not necessary for the audio signal input in step S1 (step S2; No), the process proceeds to step S9.

  Next, the audio signal input in step S <b> 1 proceeds to step S <b> 10 when high-frequency range complementation is necessary (step S <b> 9; Yes).

  Next, when there is a margin in the operation of the DSP 10 (step S10; Yes), a high-frequency region supplement filter with heavy processing is selected, high-frequency region interpolation is executed (step S11), and the process proceeds to step S13.

  If there is no margin in the operation of the DSP 10 (step S10; No), a normal high-frequency range complementary filter is selected, high-frequency range processing is executed (step S12), and the process proceeds to step S13.

  In addition, when the high frequency range supplement is unnecessary for the audio signal input in step S1 (step S9; No), the input audio signal is output (step S14), and this process ends.

  Next, an example of a filter selection method performed in the flowchart shown in FIG. 2 and a combination of filters selected according to the type of the input audio signal will be described with reference to FIG. In the figure, “◯” indicates a filter used for the input audio signal, and “X” indicates a filter not used.

  As shown in FIG. 3, for example, when a PCM audio signal a input from a music CD is input, only the filter 14 is selected. In general, an audio signal input from a CD is less likely to be subjected to high-frequency range interpolation processing. Therefore, instead of not performing high-frequency range complementation processing, low-frequency range enhancement processing is performed by a filter with a high processing load, so that the usage rate of the DSP 10 can be increased and effective voice processing can be performed. Also, depending on the load on the DSP 10 when audio processing is actually performed, it is determined in step S10 in FIG. 2 whether to select a heavy filter or a normal filter.

  When the audio signal b compressed in a format such as MP3 is input, the filter 12 is selected. In general, an audio signal that has been compressed in a format such as MP3 has a high need for high-frequency range interpolation processing. Therefore, instead of not performing the enhancement process of the low sound range, the high sound range complementation process is performed by a filter with a high processing load, so that the usage rate of the DSP 10 can be increased and an effective complementation process can be performed. Further, in step S6 of FIG. 2, it is determined whether to select a heavy filter or a normal filter according to the load on the DSP 10 when the voice processing is actually performed.

  In other cases, that is, when an audio signal c whose characteristics are unknown is input, the filter 11 and the filter 13 are selected. Then, the filter is changed according to the load on the DSP 10 when the voice processing is actually performed.

In the above embodiment, two types of filters having different loads on the DSP 10 are provided for the purpose of high-frequency interpolation processing, and two types of filters having different loads on the DSP 10 are provided for the purpose of enhancing the low frequency range. Not. Each may have more than two types of filters,
It suffices if the filter for processing can be switched according to the load of the DSP 10. The processing of the filter is not limited to the above example, and high sound range enhancement, loudness correction, and the like may be performed.

  Moreover, although the example which changes the kind and order of a filter was demonstrated as an example of the filter from which calculation load differs, it is not limited to these. Depending on the load on the DSP 10, for example, by setting a filter that changes the variable type at the time of processing calculation from a float type to a double type, the resolution of the voice processing may be improved as long as the load on the DSP 10 is allowed.

[effect]
As described above, according to the present embodiment, the audio signal processing apparatus 1 is controlled by the CPU 15 by any one of the filters 11 to 14 which are filters having different characteristics with respect to the audio signals input from the input units A to C. The sound processing is performed below, and the filter for performing the sound processing is switched by the CPU 15 according to the load on the DSP 10, so that the sound quality after the sound processing is improved in order to fully demonstrate the processing capability of the DSP 10 in the sound signal processing. Can be improved.

  Further, since the CPU 15 switches a filter that performs audio processing according to the type of the audio signal input from the input units A to C, appropriate audio processing is performed according to the processing load on the DSP 10 and the characteristics of the audio signal. be able to.

It is a block diagram which shows the example of a structure and circuit of an audio | voice signal processing apparatus. It is a flowchart which shows the example of the control flow of CPU when audio | voice processing is performed by DSP. It is a table | surface which shows the example of the combination of the filter corresponding to the format of the audio | voice signal input.

Explanation of symbols

1 Audio signal processor 10 DSP
11, 12, 13, 14 Filter 15 CPU
16 Load measurement unit 20 DAC
30 Volume A, B, C Input part d1, d1, d1 Delay element s1, s2, s3, s4, s5, s6, s7 Switch

Claims (6)

An audio signal input unit for inputting an audio signal;
An audio signal processing unit that performs audio processing on the audio signal input from the audio signal input unit and includes two or more audio processing units having different calculation loads for the same purpose audio processing content;
A control unit that performs control to switch the audio processing unit so as to preferentially select the audio processing unit having a large calculation load among the two or more audio processing units according to the processing load of the audio signal processing unit; An audio signal processing device comprising: The control unit performs control to switch audio processing contents by the audio signal processing unit according to the type of audio signal input from the input unit,
The audio signal processing apparatus according to claim 1. Two or more audio processing means having different calculation loads for the same purpose audio processing in the audio signal processing unit are different in the type of filter used,
The audio signal processing device according to claim 1. Two or more sound processing means having different calculation loads for the same purpose sound processing in the sound signal processing unit are different in the order of the filters used,
The audio signal processing device according to claim 1. An audio signal input process for inputting an audio signal;
The voice signal input step is a voice processing step for voice processing, and the voice signal processing unit having two or more voice processing means having different calculation loads for the same target voice processing content performs voice processing. An audio signal processing step for processing;
Control is performed to switch the sound processing means so as to preferentially select the sound processing means having a large calculation load among the two or more sound processing means according to the processing load by the sound processing in the sound signal processing step. An audio signal processing method comprising: a control step. An audio signal input means for inputting an audio signal to a computer;
A voice signal processing means for performing voice processing on the voice signal input from the voice signal input means, and having two or more voice processing means having different calculation loads for the same purpose voice processing content;
Control means for performing control to switch the sound processing means so as to preferentially select the sound processing means having a large calculation load among the two or more sound processing means according to the processing load by the sound signal processing means. A program to make it work.

Images