JP2010230972A - Voice signal processing device, method and program therefor, and reproduction device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a reproduction system which makes human voice easy to listen to, without sense of incongruity. <P>SOLUTION: A sound signal in which human voice is included, in sound signals of channels corresponding to a plurality of loudspeakers arranged in a surrounding of a reference point, is detected by comparing each sound characteristic such as a sound volume and a frequency band. A dynamic range of only the detected sound signal of the channel is compressed. A listener easily listens to words of a movie, which is human voice, in a simple configuration for comparing the sound volume and the frequency band without a sense of discomfort, and excellent contents are appreciated. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a sound signal processing device that processes sound signals of channels corresponding to a plurality of speakers arranged around a reference point, a method thereof, a program thereof, and a playback device.

Conventionally, a reproduction system that reproduces multi-channel sound data using a plurality of speakers is known. In this reproduction system, for example, image data is displayed on a monitor, a plurality of speakers are arranged around the viewer, and sound data is reproduced from around the viewer. Sound data to be reproduced by these reproduction systems is recorded on a package medium such as a DVD (Digital Versatile Disc), or distributed via a network such as the Internet.
Then, the conventional sound data processing apparatus performs a process of outputting corresponding sound data from each speaker arranged around the viewer, such as 5.1ch (channel).

By the way, most contents such as movies are created on the premise of playback in an environment such as a movie theater where the volume is dark and there is little noise. To produce the maximum expressive power in such an environment. Recording with a large dynamic range has been implemented.
On the other hand, when the same content is played back by a home audio / video playback device, it may be difficult to hear, particularly when the volume is reduced at night or when the surrounding noise is high. Specifically, when the content is played at a low volume, such as in the case of whispering dialogue, the playback sound drops to the vicinity of the so-called masking threshold and is often very difficult to hear. is there.

  Conventionally, for content with a large dynamic range such as a movie, it has been proposed to compress the dynamic range of sound data (see, for example, Patent Document 1).

JP 2006-524968 A

However, when dynamic range compression is performed on the sound data of the content as described in Patent Document 1 described above, for example, if the content includes music such as a musical or an opera, the contents of the content may be impaired.
In view of the above, it is an object of the present invention to provide a sound signal processing device, a method thereof, a program thereof, and a playback device that make it easy to hear a human voice without a sense of incongruity. .

  The sound signal processing device according to the present invention is a sound signal processing device that processes a sound signal of a channel corresponding to the speaker from a plurality of speakers arranged around the reference point. And a sound detection means for detecting a sound signal including a human voice among the sound signals of the channel based on a comparison of the respective sound characteristics, and a sound signal of a channel including the sound detected by the sound detection means. And compression processing means for compressing only the dynamic range.

  The sound signal processing method according to the present invention is a sound signal processing for processing sound signals of a channel corresponding to the speakers from a plurality of speakers arranged around the reference point by an arithmetic means in order to reproduce the sound signals. In the method, the calculation means detects a sound signal including a human voice among the sound signals of the channel based on a comparison of respective sound characteristics, and detects the sound signal in the sound detection step. And a compression processing step of performing dynamic range compression only on a sound signal of a channel including sound.

  The sound signal processing program according to the present invention causes the calculation means to function as the sound signal processing device according to any one of claims 1 to 7.

  A playback device according to the present invention corresponds to the sound signal processing device according to any one of claims 1 to 7 and a plurality of speakers arranged around a reference point by the sound signal processing device. And output means for outputting the sound signal of the processed channel through the speaker.

  The calculation means in the present invention is not limited to a single computer, but may include a configuration in which a plurality of computers are combined in a network, an element such as a CPU or a microcomputer, or a circuit board on which a plurality of electronic components are mounted. Is included.

It is a block diagram which shows schematic structure of the reproducing | regenerating apparatus which concerns on one Embodiment in this invention.

Hereinafter, an embodiment of the present invention will be described with reference to FIG.
In the present embodiment, the configuration of a playback system including a plurality of speakers is illustrated, but the present invention is not limited to this.

[Playback system]
In FIG. 1, reference numeral 100 denotes a reproduction system. The reproduction system 100 has a so-called multi-channel system configuration in which sound signals are reproduced from a plurality of speakers arranged around a reference point.
The sound signals to be reproduced are not limited to those including video signals such as movies, musicals, plays, and music promotion videos, but are included in various content data including only sound signals that do not include video signals such as music. In particular, it is intended to include human voices, but does not exclude reproduction of content data that does not include human voices.
The playback system 100 includes a playback device 200 that processes content data, and a plurality of speakers 300 that output sound signals processed by the playback device 200. In the case where the playback device 200 is configured to process a video signal, the playback system 100 may include a display device that outputs the video signal, that is, displays the screen.

The playback device 200 is a device that processes content data in order to output a sound signal of the content data through a speaker.
The playback apparatus 200 includes a content acquisition unit 210, a display unit 220, an input unit 230, a calculation unit 240, and an output unit 250.

The content acquisition unit 210 acquires content data. For example, the content acquisition unit 210 is connected to a drive device that reads content data stored on an optical disc such as a CD (Compact Disc) or DVD (Digital Versatile Disc) or a magnetic disc, or a network such as the Internet, an intranet, or a LAN. An interface for acquiring content data from outside the playback apparatus 200 can be exemplified.
Then, the content acquisition unit 210 outputs the acquired content data to the calculation unit 240. The content data to be acquired is output while being sequentially acquired as stream data.
If the content data includes a video signal and a sound signal, only the sound signal is output to the calculation means, and the video signal is processed by another processing device.

The display unit 220 includes a monitor that is visible from the outside in a housing (not shown) of the playback device 200. As the monitor, for example, various display devices such as a liquid crystal or an EL (Electro Luminescence) panel are used.
Then, the display unit 220 displays the processing status of the content data, the playback output status, the input operation content of the input unit 230, and the like based on the signal output from the calculation unit 240 under the control of the calculation unit 240.

The input unit 230 has a plurality of various switches such as operation buttons and operation knobs (not shown) that can be input by the user. The input means 230 outputs predetermined signals to the calculation means 240 by input operations of these switches, and inputs various conditions to the calculation means 240.
Note that the input unit 230 is not limited to a configuration in which setting input is performed by an input operation of a switch, and any input method such as voice input can be used. Further, it may be configured as a remote controller which is a so-called remote controller, and may be configured to transmit a signal corresponding to the input operation to the calculation means 240 via a wireless medium for setting input.
Although the details will be described later, the input unit 230 includes a switching unit 231 that switches the degree to which the dynamic range compression of the sound signal is performed, that is, the strength, by the user's input operation.

As the calculation means 240, for example, a system microcomputer is used, and the entire playback apparatus 200 can be controlled.
The calculation unit 240 includes a sound signal acquisition unit 241, a sound detection unit 242, and a compression processing unit 243 as a program for processing a sound signal.

The sound signal acquisition unit 241 acquires the sound signal of the content data output from the content acquisition unit 210 and extracts the sound signal for each channel.
For the extraction of the sound signal for each channel, various methods such as bandpass, mixing processing, effect processing, and delay processing are appropriately applied.

The sound detection unit 242 detects a sound signal of a channel including a human voice among the sound signals for each channel acquired by the sound signal acquisition unit 241. That is, the sound detection unit 242 determines whether or not sound is included based on the sound characteristics of each sound signal.
Specifically, the voice detection unit 242 includes a volume comparison unit 242A, a correlation determination unit 242B, and a frequency analysis unit 242C, and the sound signal of the channel determined by these includes a human voice. Judge as.

The volume comparison unit 242A determines a sound signal of a channel whose volume is relatively higher than that of the other channels among the sound signals of each channel. For example, in a movie or the like, when there is a voice of a person who is a dialogue, the volume of the voice in a scene with the dialogue is almost larger than the volume of other sounds. As described above, the sound signal of the channel including the sound has a higher volume than the sound signals of the other channels, and therefore, the sound signal of the channel having a higher volume is likely to include the sound. Therefore, the volume comparison unit 242A recognizes the sound signal of the channel with a high volume as a candidate including sound.
In particular, it is preferable to compare the sound signals of channels corresponding to adjacent speakers 300 among the speakers 300 arranged around the reference point. For example, the sound volume comparison unit 242A recognizes a sound whose volume is larger than the sound signal corresponding to the adjacent speaker 300 as a candidate including sound. That is, in many scenes of the movie, there is one character who speaks the line, and when the character appears, for example, when the character speaks on the left side from the center of the screen, the left speaker 300 Since the dialogue is remarkably reproduced in the corresponding channel, it is particularly preferable to compare the sound signals corresponding to the adjacent speakers.
Note that in the case of content in which dialogue is localized at an intermediate position between two adjacent speakers 300, that is, when dialogue is included in the channels corresponding to the two adjacent speakers 300, the corresponding two adjacent speakers 300. It may happen that the volume of both is higher than the volume of other channels. In this case, if the corresponding two adjacent speakers 300 are regarded as candidates for a channel that reproduces sound together, and it is detected that the corresponding channel is reproducing a dialogue by other means described later, the two speakers 300 are also detected. May be determined as channels that are both playing dialogue. In this case, the volume of the two adjacent speakers 300 is higher than that of the adjacent speakers 300, and can be detected by this comparison.
The comparison between the sound signals of the channels corresponding to the adjacent speakers 300 is merely a means for detection, and the dynamic range compression described later is applied to the detected sound signal in any of the adjacent sound signals. The present invention is not limited to the speaker 300 or both speakers 300, and may be, for example, a sound signal of the speaker 300 that is located at a distance from each other such as a center speaker and a rear speaker. That is, the dynamic range compression is performed on a channel including human speech such as dialogue.

The correlation determination unit 242B determines a sound signal of a channel having low correlation in sound characteristics with sound signals of other channels among the sound signals of each channel. For example, in a certain genre such as an action movie or war movie, there is an explosion sound in the middle of a dialogue, such as a loud sound production, the same explosion sound will be played even with the sound signal of other channels, In such cases, the correlation between channels is high, but in the case of dialogue, dialogue speech is often assigned only to a specific channel. In this case, it is inevitably necessary to communicate with other channels. Correlation is low. Accordingly, the correlation determination unit 242B recognizes a sound signal of a channel having low correlation with a sound signal of another channel as a candidate including sound.
The correlation determination unit 242B is also preferably configured to compare the sound signals of the channels corresponding to the adjacent speakers 300 in the speakers 300 arranged around the reference point, similarly to the sound volume comparison unit 242A. For example, the correlation determination unit 242B compares the frequency band and volume that are sound characteristics with the sound signal corresponding to the adjacent speaker 300, and the correlation between the adjacent ones is high, but the channel corresponding to the other speaker 300 has a high correlation. A sound signal having a low correlation with the sound signal, that is, a frequency signal whose level and volume are not similar is recognized as a candidate including sound. An example of the similarity in volume level is judgment based on a volume difference.
Note that, in the case of content in which the dialogue is localized at an intermediate position between two adjacent speakers 300, the correlation between the reproduced audio of the corresponding two adjacent speakers 300 is high, and each of the two adjacent speakers 300 is adjacent to the adjacent speaker 300 on the opposite side. There can be a situation where the correlation is low. In this case, as in the case of the above-described volume, by comparing the sound signals of the channels corresponding to the adjacent speakers 300 with respect to the adjacent speakers 300, the corresponding two adjacent speakers 300 can hear the sound. Considered as a candidate for the channel being played. Then, when it is detected that the corresponding channel is reproducing the dialogue by other means described later, it is preferable that the two speakers 300 be determined as the channels reproducing the dialogue.
Further, as described above, the comparison between the sound signals of the channels corresponding to the adjacent speakers 300 is only a means for detection, and the dynamic range compression described later is applied to the detected sound signal. The present invention is not limited to any one of the adjacent speakers 300 or both speakers 300, and may be, for example, a sound signal of a speaker 300 located at a distance from each other such as a center speaker and a rear speaker. .

  The frequency analysis means 242C determines whether or not sound is included by analyzing the frequency band of the sound signal of each channel. For example, human voice is generally 300 Hz to 4 kHz. Thus, the frequency analysis unit 242C analyzes the frequency band of the sound signal and recognizes the sound signal of the channel whose frequency band of 300 Hz to 4 kHz is significantly higher than the other frequency bands as a candidate including sound. .

Then, the voice detection unit 242 detects the sound signal of the channel recognized as including the human voice by the volume comparison unit 242A, the correlation determination unit 242B, and the frequency analysis unit 242C.
The channel of the sound signal to be detected is not limited to one channel. That is, when human voices are included in the sound signals of a plurality of channels, for example, when a plurality of characters speak at different standing positions, the sound signals of the plurality of channels are detected.

The compression processing unit 243 performs dynamic range compression only on the sound signal of the channel detected by the sound detection unit 242.
Various methods can be used as the compression method. In general, a method of converting the output level relative to the input level of the sound signal by performing a kind of logarithmic conversion that compresses the level difference at the time of large input while converting the small input level to the large output level can be exemplified. Specifically, it can be realized by using a DSP (Digital Signal Processor), making a reference for each sound signal of each channel, and reading the output level relative to the input level.
The compression processing means 243 can change the degree of dynamic range compression. For example, when the user inputs and operates the switching unit 231 of the input unit 230 to set and input a dynamic range compression level, the dynamic range compression is performed in accordance with the input level. .

The output unit 250 includes a digital / analog converter (DAC) 251 and an amplifier 252.
The DAC 251 is connected to the playback apparatus 200 and converts the processed digital audio signal output from the playback apparatus 200 for each channel into analog. Then, the DAC 210 outputs the sound signal converted into analog to the amplifier 252 respectively.
The amplifier 252 is connected to the DAC 251 and also to the speaker 300. These amplifiers 252 process the sound signal of the analog signal output from the DAC 251 so that it can be appropriately output from the speaker 300, and output to the speaker 300 for reproduction.

[Playback system operation]
Next, the operation of the playback system 100 will be described.
First, the user inputs a setting for requesting reproduction processing of desired content data. With this setting input, the playback apparatus 200 acquires content data recorded on a recording medium, for example, by the content acquisition unit 210 and sequentially outputs the content data to the calculation unit 240.
The sound signal acquisition unit 241 of the calculation unit 240 acquires the sound signal of the content data output from the content acquisition unit 210, and acquires the sound signal for each channel from the sound signal.
With respect to the sound signal of each channel acquired by the sound signal acquisition means 241, the sound detection means 242 of the calculation means 240 detects the sound signal of the channel including the human voice.

That is, the sound detection means 242 uses the volume comparison means 242A to determine the sound signal of the channel whose volume is relatively larger than the sound signals of the other channels among the sound signals of each channel. In particular, comparison is made between the sound signals of the channels corresponding to the adjacent speakers 300, and the sound characteristics of the channel with a high volume are specified.
In addition, the sound detection unit 242 uses the correlation determination unit 242B to determine, among the sound signals of each channel, the sound signal of the channel having low correlation with the sound signal of the other channel. In particular, when the sound characteristics such as the level of the frequency band and the volume level of the sound signals of the channels corresponding to the adjacent speakers 300 are compared, the sound characteristics are not similar, that is, the level distribution in each frequency band is different. If the volume difference is large, the sound characteristics of the channel with low correlation are specified.
Further, the sound detection means 242 analyzes the frequency band of the sound signal of each channel by the frequency analysis means 242C, that is, identifies the sound signal of the channel whose frequency band of 300 Hz to 4 kHz is significantly higher than other frequency bands. To do.
When the channels specified by the sound volume comparison unit 242A, the correlation determination unit 242B, and the frequency analysis unit 242C are the same, the sound detection unit 242 detects the sound signal of that channel as a candidate.

Thereafter, the calculation means 240 uses the compression processing means 243 to perform dynamic range compression on only the sound signal of the channel detected by the sound detection means 242.
Then, the sound signals of the respective channels are synchronized and output to the output means 250 corresponding to each speaker 300 and output from the speaker 300. With this output, the content data is reproduced.

[Effect of playback system]
As described above, in the above-described embodiment, among sound signals of channels corresponding to a plurality of speakers arranged around the reference point, a sound signal including human speech is obtained based on comparison of sound characteristics. Only the sound signal of the detected channel is compressed in the dynamic range.
For this reason, by performing dynamic range compression on the entire sound signal of the content data, for example, the compression of the sound that is easily affected by the balance of the sound, such as music, is applied and the contents of the music effect are changed. Therefore, it is easy to hear dialogues such as human speech without losing the balance of the sound characteristics of the entire sound, and it is possible to provide good content appreciation without any sense of incongruity.

As a detection of a sound signal including human voice, detection is performed by comparing sound characteristics of sound signals of channels corresponding to the adjacent speakers 300 in the speakers 300 arranged around the reference point.
In other words, since the line of a movie is often remarkably reproduced only on a specific channel, a simple structure that compares sound signals of channels of adjacent speakers 300 with each other, and inconvenience that is difficult to hear in the past has occurred. Can be easily heard without any sense of incongruity while suppressing malfunction.

Further, when detecting a sound signal including human voice, a sound signal of a channel whose volume is relatively higher than that of the sound signals of other channels is set as a sound signal candidate including the voice.
That is, for example, in a scene with a movie dialogue, the dialogue is generally localized to a sound correlated with the position of the person on the screen. In other words, the dialogue is conspicuous between a predetermined channel or two adjacent channels. In general, mixing is performed so as to be played back, and therefore, with a simple configuration for comparing the volumes, it is relatively difficult for malfunctions to occur and the dialogue can be easily heard without a sense of incongruity.

Further, when detecting a sound signal including human speech, a sound signal of a channel having low sound characteristic correlation with other channel sound signals, that is, having a low frequency band or volume similarity, is detected. It is a candidate.
That is, for example, in a scene with a movie dialogue, the dialogue is generally localized to a sound correlated with the position of the person on the screen. In other words, the dialogue is conspicuous between a predetermined channel or two adjacent channels. In general, the channel with dialogue has a different sound characteristic from the sound signal of the other channels, so it is relatively easy to compare the sound characteristics. It is easy to hear dialogue without feeling uncomfortable and difficult to act.
In particular, when determining the correlation, the sound signals of the channels corresponding to the adjacent speakers 300 are highly correlated with each other, but a channel having a low correlation with the sound signals of the other channels is used as a candidate. Operation can be prevented.

Further, when detecting a sound signal including human voice, the frequency band is analyzed, that is, a sound signal of a channel having a high frequency band ratio of 300 Hz to 4 kHz, which is a human voice, is selected as a sound signal candidate including the voice. It is said.
For this reason, with a simple configuration for comparing the level distribution status of the frequency bands, it is relatively difficult for malfunctions to occur and the dialogue can be easily heard without a sense of incongruity.

Then, as a detection of a sound signal including human speech, if the channels specified by the volume comparison means 242A, the correlation determination means 242B, and the frequency analysis means 242C are the same, the sound signal of that channel is Detected as a voice candidate including voice.
For this reason, it is possible to more reliably detect the sound signal of the channel including the human voice and to prevent malfunction.

[Modification]
In addition, this invention is not limited to one Embodiment mentioned above, The deformation | transformation shown below is included in the range which can achieve the objective of this invention.

For example, in the present embodiment, the sound volume comparison unit 242A, the correlation determination unit 242B, and the frequency analysis unit 242C are provided, and each of them is determined as to whether or not the sound signal includes a human voice. Absent. That is, determination may be made by providing at least one of the configurations. Furthermore, other judgments may be added to these configurations.
In addition, although the volume comparison unit 242A, the correlation determination unit 242B, and the frequency analysis unit 242C each detect and recognize what is recognized as a sound signal including a human voice, at least these are exemplified. Any one recognized as a sound signal including a human voice may be compressed.

  Moreover, although the audio | voice detection means 242 illustrated the structure compared between the sound signals of the channel corresponding to the adjacent speaker 300 in the speaker 300 arrange | positioned around the reference | standard point, it is not this limitation. For example, a sound signal that differs from the sound signal of another channel may be recognized as a sound signal including a human voice.

  Although a configuration in which the switching unit 231 is provided and the degree of dynamic range compression can be changed is exemplified, the degree of dynamic range compression may be constant.

The playback system 100 is exemplified, but for example, a playback device 200 that does not include the speaker 300, a circuit board on which the calculation unit 240 is mounted, and a program that causes the calculation unit such as a computer to function as the calculation unit 240 are configured. Also good.
Further, although the description has been made by processing the content data having the dialogue, the reproduction of the content data not containing the dialogue such as an orchestra performance is not excluded. That is, even if such content data is processed, the line is not included, so that the dynamic range compression process is not executed.

  In addition, the specific structure and procedure for carrying out the present invention can be appropriately changed to other structures and the like within a range in which the object of the present invention can be achieved.

200 …… Playback device 240 …… Calculation means 242 …… Audio detection means 242A ... Volume comparison means 242B ... Correlation determination means 242C ... Frequency analysis means 243 …… Compression processing means 300 …… Speaker

Claims (10)

A sound signal processing device for processing sound signals in order to reproduce sound signals of channels corresponding to these speakers from a plurality of speakers arranged around a reference point,
A sound detection means for detecting a sound signal including a human voice among the sound signals of the channel based on a comparison of respective sound characteristics;
Compression processing means for dynamic range compression only of the sound signal of the channel containing the sound detected by the sound detection means;
A sound signal processing apparatus comprising: The sound signal processing device according to claim 1,
The sound detection means compares sound signals of channels corresponding to adjacent speakers in speakers arranged around a reference point, and determines whether or not human sound is included. Processing equipment. The sound signal processing device according to claim 1 or 2,
The sound detection means detects a sound signal of a channel whose sound characteristics are relatively louder than sound signals of other channels among sound signals of each channel as a sound signal including sound. Sound signal processing device. The sound signal processing device according to any one of claims 1 to 3,
The sound detection means detects sound signals of channels whose sound characteristics are low in correlation with sound signals of other channels, among sound signals of each channel, as sound signals containing sound apparatus. The sound signal processing device according to claim 4,
The sound detection means has a sound characteristic of a channel having a high correlation in sound characteristics between sound signals of channels corresponding to adjacent speakers, and a sound signal of a channel having a low sound specific correlation in the sound signals of other channels. A sound signal processing apparatus, characterized in that the sound signal is detected as an included sound signal. The sound signal processing device according to any one of claims 1 to 5,
The sound signal processing apparatus, wherein the sound detection means determines whether or not sound is included by analyzing a frequency band in a sound characteristic of a sound signal of each channel. The sound signal processing apparatus according to claim 6,
The sound detection means detects a sound signal of a channel having a large proportion of a frequency band of 300 Hz to 4 kHz with respect to the entire frequency band in the sound signal of each channel as a sound signal containing sound. Signal processing device. In order to reproduce sound signals of channels corresponding to these speakers from a plurality of speakers arranged around a reference point, the sound signal processing method of processing the sound signal by a calculation means,
The computing means is
A sound detection step of detecting a sound signal including a human voice among the sound signals of the channel based on a comparison of respective sound characteristics;
A sound signal processing method comprising: performing a dynamic range compression only on a sound signal of a channel including sound detected in the sound detection step. A sound signal processing program for causing a calculation means to function as the sound signal processing device according to any one of claims 1 to 7. The sound signal processing device according to any one of claims 1 to 7,
Output means for outputting sound signals of channels processed by the sound signal processing device corresponding to a plurality of speakers arranged around the reference point by the speakers;
A playback apparatus comprising:

Images