JP2006148608A - Sound signal discriminating device, tone controlling device, broadcast receiving unit, program, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sound signal discriminating device which accurately discriminating speech from non-speech when sound signals are inputted into it. <P>SOLUTION: The sound signal discriminating device is equipped with: a speech/non-speech judging means 12 for judging whether the inputted sound signals are corresponding to speech or non-speech; a monaural/stereo judging means 13 which judges whether the inputted sound signals are monaural or stereo; and a reference optimizing means 14 which optimizes a judgment reference in the speech/non-speech judging means 12 resting on a judgment result given by the monaural/stereo judging means 13. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an audio signal discriminating device, a sound quality adjusting device, a broadcast receiver, a program, and a recording medium, and more particularly, an audio signal discriminating device that discriminates speech / non-speech from an audio signal, and the apparatus. The present invention relates to a sound quality adjusting device, a broadcast receiver including the sound quality adjusting device, a program thereof, and a computer-readable recording medium on which the program is recorded.

  Conventionally, in general audio devices, various sound quality adjustment devices such as bass adjustment that adjusts the output frequency characteristics of the low frequency range, treble adjustment that adjusts the output frequency characteristics of the high frequency range, and loudness adjustment that emphasizes the low and high frequency ranges. Is provided.

As such a sound quality adjusting device, by detecting the presence or absence of the periodicity from the sound information itself of the input sound signal, it is determined whether the input signal is music information or other information, and the result is A device that controls acoustic parameters in response to this has been proposed (see, for example, Patent Document 1).
JP-A-61-93712

  However, in particular, in devices that receive television broadcasts or radio broadcasts, an unexpected misjudgment may occur when judging whether or not music information is appropriate only from audio information.

  For example, if a cappella flows in a music program, it is not possible to detect the rhythm due to its style, it is determined that it is not music information, and the optimal acoustic parameter for this music information is not selected by an equalizer or the like A misjudgment occurs. As a result, for this music information, the equalizer may select, for example, the optimal acoustic parameters for speech, so the clarity of the words can be improved for the music information of a cappella that emphasizes the sound of raw sound. As a result, the sound characteristics that are emphasized (relatively emphasized in the middle sound range) are output, and the sound settings that the user originally wants to hear are not achieved.

  In addition, while watching a news program, it is preferable to select parameters that are optimal for speech that emphasizes language clarity, but depending on the content of the news, it may sometimes be in parallel with the announcer's speech. In some cases, the sound collected on site is output as it is. If music is mixed in such collected audio information, it is assumed that the music information output from the collected audio has an advantage over the speech of the news program depending on the balance of the volume of both. Therefore, such a case is also a problem that can occur sufficiently as an example opposite to the above-described a cappella example.

  The present invention has been made in view of the above circumstances, and an audio signal discriminating apparatus capable of accurately discriminating speech / non-speech from an input audio signal, and sound quality adjustment provided with the apparatus It is an object of the present invention to provide an apparatus, a broadcast receiver including the sound quality adjusting apparatus, a program thereof, and a computer-readable recording medium on which the program is recorded.

The present invention is constituted by the following technical means in order to solve the above-described problems.
The first technical means is a speech / non-speech determination means for making a determination to determine whether an input audio signal corresponds to speech or non-speech in an audio signal determination device; Determination unit in the speech / non-speech determination unit based on the determination result of the monaural / stereo determination unit and the determination result in the monaural / stereo determination unit And a standard optimizing means for optimizing.

  According to a second technical means, in the first technical means, the determination in the speech / non-speech determination means is performed by a plurality of signal analysis, and the reference optimization means is a threshold value for each signal analysis as the determination criterion. The set is changed based on monaural / stereo determination.

  A third technical means is a sound quality adjusting device including the audio signal discriminating device in the first or second technical means, and is adapted to perform speech on an audio signal discriminated as speech / non-speech by the audio signal discriminating device. It is characterized by adjusting the sound quality different between non-speech and non-speech.

  A fourth technical means is a broadcast receiver comprising the sound quality adjusting device and the broadcast receiving device in the third technical means, wherein the audio signal is transmitted from the broadcast signal received by the broadcast receiving device to the sound quality adjusting device. It is characterized by inputting, adjusting sound quality and outputting sound.

  The fifth technical means includes a monaural / stereo determination step for determining whether the input audio signal is a monaural signal or a stereo signal, and a reference optimization for optimizing a predetermined reference based on the determination result A program for causing a computer to execute a step and a speech / non-speech determination step for determining whether an input audio signal corresponds to speech or non-speech based on the predetermined criterion. .

  According to a sixth technical means, in the fifth technical means, the determination in the speech / non-speech determination step is performed by a plurality of signal analysis, and the reference optimization step is a threshold for each signal analysis as the predetermined reference. The set is changed based on monaural / stereo determination.

  According to a seventh technical means, in the fifth or sixth technical means, the sound quality which is different between speech and non-speech for the speech signal discriminated as speech / non-speech by the speech / non-speech determination step in the computer. The program further includes a program for executing the adjusting step.

  The eighth technical means is a computer-readable recording medium on which a program according to any of the fifth to seventh technical means is recorded.

  According to the present invention, it is possible to accurately determine speech / non-speech for an input audio signal.

  The audio signal discriminating apparatus according to the present invention includes speech / non-speech determination means, monaural / stereo determination means, and reference optimization means. Hereinafter, a sound quality adjustment apparatus that includes such a sound signal determination apparatus and performs sound quality adjustment based on the determination here will be described. However, the sound signal determination apparatus according to the present invention is based on, for example, determination other than sound quality adjustment. The present invention can also be applied to separate recording (recording) of content (content including the audio signal).

FIG. 1 is a block diagram showing a configuration example of a sound quality adjustment apparatus according to an embodiment of the present invention. In the figure, 1 is a sound quality adjustment apparatus, 11 is an audio signal input means, 12 is a speech / non-speech determination means, 13 mono / stereo decision means, reference optimization means 14, 14a switch, setting means 14b is the threshold _{(threshold) V SL1,} 14c is set means to the threshold _{V SL2,} 15 is the sound quality adjustment section, 16 It is an audio signal output means.

  The speech / non-speech determination unit 12 performs a determination for determining whether the audio signal input by the audio signal input unit 11 corresponds to speech or non-speech. In the audio signal input means 11, the input source and input method are not limited. The speech / non-speech determination unit 12 may be configured in whole or in part by hardware or software. In the speech signal discriminating apparatus according to the present invention, the speech / non-speech judging means 12 judges whether the speech signal is speech or non-speech.

  In the audio signal discrimination device according to the present invention, an empirical rule such as “News programs are generally monaural broadcasts, while music and music programs in which music flows are often set to stereo broadcasts” is used. By detecting the monaural / stereo signal superimposed on the audio signal, it is determined whether the currently broadcast program is suitable for speech / non-speech (music). For this reason, the audio signal discriminating apparatus according to the present invention includes a monaural / stereo determination unit 13 and a reference optimization unit 14. The sound quality adjusting apparatus according to the present invention controls the acoustic parameters based on the determination.

  The monaural / stereo determination means 13 determines whether the input audio signal is a monaural signal or a stereo signal. The monaural / stereo determination means 13 may be configured in whole or in part by hardware or software, and is determined by information such as mono / stereo switching when a sound signal is input. May be. Furthermore, if the original content of the audio signal is posted in an electronic program guide (EPG) and can be reserved for recording, the mono / stereo information in the EPG is also posted together. It is also possible to perform monaural / stereo determination by acquiring.

  The reference optimization unit 14 optimizes the determination criterion in the speech / non-speech determination unit 12 based on the determination result in the monaural / stereo determination unit 13. Therefore, in the audio signal discriminating apparatus according to the present invention, in order to accurately determine the speech / non-speech of the audio signal, it is assumed that monaural / stereo determination and reference optimization have been performed on the audio signal in advance. However, a delay or the like may be used, or the speech / non-speech determination may be performed by sequentially performing monaural / stereo determination and reference optimization every time an audio signal is input.

  The sound quality adjusting means 15 adjusts the sound signal discriminated as speech / non-speech by the audio signal discriminating apparatus having such components to a different sound quality between speech and non-speech. The sound quality setting method here is arbitrary, and it suffices if the setting value, the increase / decrease setting value, or the setting value in each frequency band differs depending on speech / non-speech. For example, the sound quality setting where the center frequency of the equalizer such as a graphic equalizer and the Q value of the filter (the sharpness of peaks and valleys in the curve for one band of the graphic equalizer) are fixed, and these can be changed as in the case of a parametric equalizer. Sound quality setting may be used. The audio signal output unit 16 outputs the audio signal adjusted by the sound quality adjustment unit 15.

  2 is a flowchart for explaining an example of the sound quality adjustment process in the sound quality adjustment apparatus of FIG. 1, and FIG. 3 is a diagram showing an example of the sound quality setting equalizing used in the sound quality adjustment process in the sound quality adjustment apparatus of FIG. is there.

  For the sake of simplicity, the description will be made assuming that the determination criterion for speech / non-speech is made by a single threshold process. First, when an audio signal is input, monaural / stereo determination means 13 performs monaural / stereo determination (step S1). In this determination, for example, assuming that L is a left input signal and R is a right input signal, a calculation of (LR) / (L + R) is performed on the input signal, and the phase difference determination may be performed.

If it is determined by this determination that the signal is a monaural signal, the reference optimization means 14 connects the switch 14a to the setting means 14b side for the threshold value _VSL1, and the determination threshold value in the speech / non-speech determination means 12 _Is set to _VSL1 (step S2). On the other hand, if it is determined in step S1 that the signal is a stereo signal, the reference optimizing unit 14 connects the switch 14a to the setting unit 14c side for the threshold value _VSL2, and the determination in the speech / non-speech determining unit 12 is performed. _Is set to _VSL2 (step S3). By optimizing the threshold setting in this way, control is performed so that it is easy to determine speech for monaural signals such as news, and it is easy to determine non-speech for stereo signals with a lot of music including BGM. Can do. The configuration of the reference optimization unit 14 is not limited to that shown in the figure.

Next, the speech / non-speech determination unit 12 determines speech / non-speech based on the threshold value V _SL1 / V _SL2 set in any of steps S 2 / S 3 (step S 4). If it is determined to be speech, the sound quality setting A is selected (step S5), and the process ends. On the other hand, if it is determined in step S4 that the speech is not speech, the sound quality setting B is selected (step S6), and the process ends.

  Here, an example of the difference between the sound quality setting A and the sound quality setting B will be described with reference to FIG. In the case of the sound quality setting A (speech), the frequency characteristic of the equalizer is set as shown in the graph 21, and in the case of the sound quality setting B (non-speech), the frequency characteristic of the equalizer is set as shown in the graph 22. The difference between the graph 21 and the graph 22 is that in the case of non-speech, the vicinity of the predetermined low frequency 22a and the vicinity of the predetermined high frequency 22b are emphasized compared to the case of speech.

  As described above, according to the audio signal determination device according to the present embodiment, the determination standard of the automatic speech detection function can be optimized and the accuracy of the detection function can be improved by performing monaural / stereo determination. Therefore, it is possible to accurately determine speech / non-speech with respect to the input audio signal, that is, suitable speech / non-speech detection according to the monaural / stereo signal of the audio signal. For example, optimization control can be performed so that a monaural signal such as news can be easily determined as speech, and a stereo signal with a lot of music including BGM can be easily determined as non-speech.

  In addition, according to the sound quality adjusting apparatus according to the present embodiment, the optimum acoustic parameter for the audio signal is determined not only from the audio information of the audio signal but also in accordance with the gist of the program (program including the audio signal) (speech / By simultaneously determining (non-speech determination), it is possible to reduce erroneous determination of acoustic parameter control such as an equalizer based on the characteristics of the input audio signal as much as possible, and it is possible to perform accurate acoustic parameter control and accurate sound quality adjustment. In addition, in a device equipped with the audio signal discriminating apparatus according to the present invention, including the sound quality adjusting apparatus according to the present embodiment, for example, the gist of the program is indicated by a monaural / stereo signal superimposed on the audio signal simultaneously with the audio information. By determining and optimizing the criteria for determining whether the input audio signal is speech or non-speech (music) according to the result, speech / non-speech according to the contents and characteristics of the broadcast program Free control of speech detection and control of equipment based on the control (for example, sound quality adjustment and separate recording) can be performed.

  As another embodiment of the present invention, it is preferable that the determination in the speech / non-speech determination unit 12 is performed by a plurality of signal analyses. In this form, a plurality of signal analyzes are performed on the input audio signal in step S4 of FIG. Signal analysis includes, for example, signal energy change analysis with respect to time, syllable uniformity analysis, and frequency vs. sound intensity analysis. By such signal analysis, for example, (I) signal energy change over time, (II) frequency versus voice intensity, (III) vowel and consonant order, (IV) syllable length, (V) consonant and vowel. The amount of energy can be obtained.

For example, in consideration of the following points, speech / non-speech may be determined. (I) In speech, there is a segment with low speech energy between syllables (with high speech energy) and syllables, and such segments are often absent in non-speech. (II) The mid-range intensity of speech is 100 Hz to 3 kHz, and the non-speech intensity is low and high. (III) Speech often follows the order in a syllable from a consonant to a vowel. (IV) Speech often has uniform syllable lengths. In (V) speech, the amount of vowel energy is often greater than the amount of consonant energy. Furthermore, (I) to (V) are weighted and summed, and statistical processing is performed to obtain a final signal analysis result. If the numerical value is monaural, it is stereo with the threshold value _VSL1 . In this case, the speech / non-speech determination may be performed by determining with the threshold value _VSL2 . As another method, in this embodiment, the reference optimization unit 14 may change a set of threshold values for each signal analysis as a determination criterion based on monaural / stereo determination.

Also in the present embodiment, as described above, in step S1, for example, (LR) / (L + R) is performed on the input signal, and phase difference determination is performed. Threshold (V _SL1 / V _SL2 ) is determined. Then, sound quality setting (A or B) is performed based on the result of signal analysis (step S5 / S6), and an audio signal is output.

  Further, the audio signal discriminating apparatus and the sound quality adjusting apparatus 1 described above with reference to FIG. 1 to FIG. 3 and each means as the constituent elements may be configured by hardware as described above, but a part thereof is configured by software. May be. For example, it may be configured by incorporating a program into a general-purpose computer such as a PC (personal computer), and various processes in that case will be described with reference to a configuration example of a general information processing apparatus shown in FIG. . FIG. 4 is a block diagram showing a configuration example of a general information processing apparatus, in which 3 is an information processing apparatus, 31 is a CPU (Central Processing Unit), 32 is a RAM (Random Access Memory), and 33 is a rewritable memory. A ROM (Read Only Memory) 34 is an input device, 35 is a display device, 36 is an output device, and 37 is a bus.

  A program for causing a computer to function as the device and each means in the present invention is stored in the ROM 33 and is executed by being read out by the CPU 31. This program when mounted on a computer or the like is a program (a program that causes a computer to function) that controls the CPU 31 or the like of the computer as each of the means described above. Information handled by the apparatus and each means according to the present invention is temporarily stored in the RAM 32 at the time of processing, and then stored in various ROMs 33, and is read out, corrected, and written by the CPU 31 as necessary. Here, examples of the information related to the present invention include a threshold value and an audio signal when the signal is input and analyzed by an audio signal input unit as one of the input devices 34. Further, for example, a threshold value set among the threshold values stored in the ROM 33 may be read out to the RAM 32 so that the threshold value setting is maintained.

  Further, the progress of the process and the result of the process are presented to the device user through the display device 35 such as an LCD, PDP, organic EL, CRT, and the like, if necessary, from the input device 34 such as a keyboard and a mouse (pointing device). The apparatus user may input and specify parameters necessary for processing (for example, specification of an audio signal to be input or contents including the input signal). In addition, this program may be provided with a graphical user interface (GUI) for the display device 35 so as to be easy for the device user to use. The output device 36 includes a speaker that is an audio signal output device, a communication device such as a network board for communication by connecting to a network, and an output device for an output device such as a printing device. . Note that the CPU 31, RAM 32, ROM 33, input device 34, display device 35, and output device 36 may be connected by a bus 37 or the like.

  Further, as a recording medium on which the program as described above is recorded, specifically, a CD-ROM, a magneto-optical disk, a DVD-ROM, an FD, a flash memory, and various other ROMs (including a rewritable ROM) and a RAM The above functions can be easily realized by recording and distributing a program for causing a computer to execute the functions of the above-described embodiments of the present invention on these recording media. Then, the information processing apparatus such as a computer is loaded with the recording medium as described above, and the program is read by the information processing apparatus, or the program is stored in the recording medium included in the information processing apparatus, By reading in response, the function according to the present invention can be executed.

  Furthermore, the present invention includes an audio signal discriminating device (or a sound quality adjusting device) as in each of the above-described embodiments, and a broadcast receiving device that receives a broadcast signal of television broadcast or radio broadcast as well as digital / analog. It can also be applied to a broadcast receiver. Here, an audio signal is input from the broadcast signal received by the broadcast receiving apparatus to the sound quality adjusting apparatus, and the sound quality is adjusted and output as sound. In addition, as described above, when the discrimination of the audio signal according to the present invention is applied to content recording, such a broadcast receiver not only receives the broadcast but also records or reserves the received broadcast. May be added. Furthermore, when the discrimination of the audio signal according to the present invention is applied to content recording, it is suitable not only for received content but also for re-recording content acquired via a network or a recording medium. For example, speech / non-speech determination can be used for CM determination and other separate recordings with various recorders.

It is a block diagram which shows the example of 1 structure of the sound quality adjustment apparatus which concerns on one Embodiment of this invention. It is a flowchart for demonstrating an example of the sound quality adjustment process in the sound quality adjustment apparatus of FIG. It is a figure which shows an example of the sound quality setting equalizing used by the sound quality adjustment process in the sound quality adjustment apparatus of FIG. FIG. 11 is a block diagram illustrating a configuration example of a general information processing apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Sound quality adjustment apparatus, 3 ... Information processing apparatus, 11 ... Audio | voice signal input means, 12 ... Speech / non-speech determination means, 13 ... Monaural / stereo determination means, 14 ... Reference | standard optimization means, 14a ... Switch, 14b ... Threshold setting means for V _SL1, setting means to 14c ... threshold _{V SL2,} 15 ... sound quality adjustment section, 16 ... audio signal output means, 31 ... CPU, 32 ... RAM, 33 ... rewritable ROM, 34 ... input device, 35 ... display device, 36 ... output device, 37 ... bus.

Claims (8)

  Speech / non-speech determination means for determining whether the input audio signal corresponds to speech or non-speech, and the input audio signal is either a monaural signal or a stereo signal A monaural / stereo determination unit that determines whether the signal is non-speech and a standard optimization unit that optimizes a determination criterion in the speech / non-speech determination unit based on a determination result of the monaural / stereo determination unit. An audio signal discrimination device.   The determination in the speech / non-speech determination means is performed by a plurality of signal analysis, and the reference optimization means changes a set of thresholds for each signal analysis as the determination reference based on monaural / stereo determination. The audio signal discrimination device according to claim 1, wherein   A sound quality adjusting device comprising the audio signal discriminating device according to claim 1, wherein the audio signal discriminated as speech / non-speech by the audio signal discriminating device has different sound quality between speech and non-speech. A sound quality adjusting device characterized by adjusting.   A broadcast receiver comprising the sound quality adjusting device according to claim 3 and a broadcast receiving device, wherein the audio signal is input to the sound quality adjusting device from the broadcast signal received by the broadcast receiving device, and the sound quality is adjusted. A broadcast receiver characterized by outputting sound.   A monaural / stereo determination step for determining whether the input sound signal is a monaural signal or a stereo signal, a reference optimization step for optimizing a predetermined reference based on the determination result, and the input sound A program for causing a computer to execute a speech / non-speech determination step of determining whether a signal corresponds to speech or non-speech based on the predetermined criterion.   The determination in the speech / non-speech determination step is performed by a plurality of signal analysis, and the reference optimization step changes a set of thresholds for each signal analysis as the predetermined reference based on monaural / stereo determination. The program according to claim 5.   The computer program further includes a program for causing the computer to execute a step of adjusting the sound quality determined by the speech / non-speech in the speech / non-speech determination step to a sound quality different between the speech and the non-speech. Item 7. The program according to item 5 or 6.   The computer-readable recording medium which recorded the program of any one of Claim 5 thru | or 7.

Images