JP2010118978A - Controller of localization of sound, and method of controlling localization of sound - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve sound recognition by clear sound regardless of the type of a content and even if monaural sound such as dialogs, announcement or the like is included in L and R channels. <P>SOLUTION: A controller 150 of localization of sound is equipped with: a broadcast receiving unit 210 which can receive television broadcast; an audio coded signal extraction unit 216; a sound determination unit 218; a degree-of-correlation derivation unit 222 for deriving a degree of correlation between L and R channels based on a flag indicating whether or not coding using a correlation between the L and R channels is performed on an audio coded signal; a storage unit 224 for storing a table 268 of coefficient groups wherein degrees of correlation and coefficient groups of head-related transfer functions are associated with each other; a table extraction unit 226 for extracting a coefficient group of a head-related transfer function from the table of coefficient groups according to the degree of correlation derived by the degree-of-correlation derivation unit; and a localization of sound processing unit 228 for executing localization of sound processing using the head-related transfer function which reflects the extracted coefficient group. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a sound image localization control device and a sound image localization control method that perform sound image localization processing on a television broadcast that supports surround.

  Conventionally, there is a surround system as a technique for improving the realistic sensation of sound, and a wide space has been reproduced in movies and the like. For example, in the 5.1 surround system, in addition to the left / right speaker and the left / right rear speaker, the sound is reproduced with the five-channel audio signal emitted from the five speakers of the center speaker and the low-frequency signal of one channel. is doing.

  In contrast to such a surround system, there is a sound image localization technique such as a virtual surround system that reproduces a virtual surround space by, for example, two speakers installed in front of a listener without arranging a speaker behind the listener. Has been developed. In the virtual surround system, the calculation of the head-related transfer function and the like for localization of the sound image was complicated, but the calculation has become easier due to recent advances in computers, and it is necessary to install a speaker behind the listener. Therefore, it has become widely known from the viewpoint of housing conditions and ease of wiring.

  In recent years, television broadcasting has also been digitized, and in conventional analog broadcasting, monaural and stereo systems have been mainstream, whereas in digital broadcasting, content corresponding to surround is increasing. Therefore, not only movies but also regular television broadcasts, such as news programs and sports broadcasts, can be easily enjoyed surround, and in the future, television broadcasts that support surround will become more common. It is done.

Such television broadcasting includes various contents. However, if the same sound image localization process is uniformly applied to all the contents, the desired sound image localization effect may not be obtained. Therefore, a technique for controlling the sound image localization effect according to the content content based on the content information is disclosed (for example, Patent Document 1).
No. 2006-148664

  Among various contents, for example, in a movie, a plurality of sound images can be imaged by the above-described virtual surround system, and a realistic sound can be enjoyed. On the other hand, speech such as dialogues and announcements (hereinafter referred to as “single speech”) produced by commentators such as news programs and sports broadcasts are required to be clearer than a sense of reality. Therefore, when such content corresponds to surround, a single sound by a commentator is often assigned only to a center channel of a plurality of channels.

  Also in the technique of Patent Document 1 described above, in the case of a news program, on the premise that the sound is assigned only to the center channel, the reverberation is suppressed by reducing the number of center channel filter processing taps. Increases clarity.

  However, in digital broadcast news programs and sports broadcasts, the role of the channel is different from DVD, etc., which is defined in accordance with the creator's intentions. In many cases, surround sound is included. Therefore, in the virtual surround system, when the sound image localization processing with a sense of spread is applied to the L / R channel, even if the center channel reverberation can be suppressed using the technique of Patent Document 1, the L / R channel The single voice included in the channel remains, and the clarity is deteriorated.

  In view of such a problem, the present invention aims to improve the audibility of a clear voice regardless of the type of content, and even when a single voice such as a dialogue or an announcement is included in the L / R channel. It is an object of the present invention to provide a sound image localization control device that can perform the above-described operation.

  In order to solve the above problems, a typical configuration of the sound image localization control device of the present invention includes a broadcast receiving unit capable of receiving a television broadcast, and an audio for extracting the audio data of the television broadcast received by the broadcast receiving unit. When it is determined that the encoded signal extraction unit, the audio encoded signal is compatible with surround, and the surround determination unit that determines whether it is compatible with surround, from the auxiliary information of the audio encoded signal, A flag indicating whether or not correlation encoding, which is encoding using correlation between L / R channels, is applied to an audio encoded signal, is extracted, and correlation between L / R channels is based on the extracted flag. A correlation degree deriving unit for deriving a degree, a storage unit for storing a coefficient group table in which the degree of correlation and the coefficient group of the head related transfer function are associated, and a correlation degree derived by the correlation degree deriving unit A table extraction unit that extracts a coefficient group of a head related transfer function from a number group table, and a sound image localization processing unit that executes a sound image localization process using a head related transfer function that reflects the extracted coefficient group, To do.

  In movies and the like, sound images can be assumed in any number and at any position, and such a plurality of sound images improve the sense of presence of the listener. On the other hand, a single sound in a news program or a sports broadcast should be assumed to be a single and central sound image, and should be so. In the present invention, the degree of correlation between the L / R channels is derived, and a voice with a high degree of correlation is regarded as a single voice emitted from a single sound image in the center, and the coefficient group makes the voice clear. The speech is output using the head-related transfer function based on it. With this configuration, it is possible to improve the audibility with clear sound even when a single sound such as a dialogue or announcement is included in the L / R channel regardless of the type of content. Further, in the case where correlation encoding is employed in the encoding process, the present invention uses a flag indicating whether or not the correlation encoding has been performed. The degree of correlation can be easily derived without complicated complicated calculations, and the processing load can be reduced.

  When the correlation encoding is performed for each of the plurality of frequency bands of the audio encoded signal, the correlation degree deriving unit determines whether the L / R channel is based on the ratio of the frequency bands subjected to the correlation encoding with respect to the entire frequency band. The degree of correlation may be derived.

  Here, depending on the degree of correlation, not only simply whether or not the voice is a single voice, but also the probability that it is a single voice, that is, all frequencies in the frequency band on which correlation coding is performed. The coefficient group of the head-related transfer function can be changed according to the ratio to the band. With this configuration, it is possible to further improve the audibility with clear sound.

  The audio encoded signal may be encoded by an AAC (Advanced Audio Coding) method, and the correlation encoding method may be an MS (Mid Side) stereo method.

  AAC is widely used as an audio encoding method. In AAC, the MS stereo method is used as the encoding method using the correlation between channels, and it can be easily derived whether or not the encoding is performed by the flag. With this configuration, it is possible to reduce the processing load of the sound image localization control device.

  The correlation degree deriving unit may derive the correlation degree by making the weighting of the predetermined frequency band larger than other frequency bands. At this time, the predetermined frequency band may be 200 to 4000 Hz.

  An object of the present invention is to clarify a voice when the target voice is a single voice. Therefore, the accuracy of deriving the degree of correlation can be improved by making the weighting for 200 to 4000 Hz, which is the frequency band of human voice, larger than other frequency bands.

  Here, the head related transfer function can independently adjust the direct sound and the reflected sound of the sound signal by the coefficient group.

  When the impulse response through the head-related transfer function is viewed in the time domain, the response waveform can be distinguished into direct sound and reflected sound. In the present invention, the coefficient of the direct sound and the reflected sound in the head-related transfer function is adjusted, and the ratio of the reflected sound to the direct sound is lowered, thereby suppressing reverberation due to reflection and generating clear sound.

  The storage unit includes a coefficient group table corresponding to a plurality of head related transfer functions corresponding to the arrangement of the speakers, and the table extracting unit is a head selected and determined according to the arrangement of the speakers among the plurality of head related transfer functions. A group of coefficients of the partial transfer function may be extracted.

  The head-related transfer function changes depending on the positional relationship between the listener's head (both ears) and the speaker. Therefore, by selecting an optimal head-related transfer function corresponding to the position of the speaker, the user can obtain more appropriate clear sound.

  In order to solve the above-described problems, a typical configuration of the sound image localization control method of the present invention receives a television broadcast, extracts an audio encoded signal that is audio data of the received television broadcast, and performs audio encoding. It is coding using the correlation between the L / R channels from the auxiliary information of the audio coded signal when it is determined whether the signal is compatible with surround and it is determined that the signal is compatible with surround. A flag indicating whether or not correlation encoding is applied to the audio encoded signal is extracted, and based on the extracted flag, a correlation degree between the L / R channels is derived, and a correlation is determined according to the derived correlation degree. The coefficient group of the head related transfer function is extracted from the coefficient group table that associates the degree and the coefficient group of the head related transfer function, and the sound image localization processing is executed by the head related transfer function reflecting the extracted coefficient group. It is characterized in.

  The components based on the technical idea of the sound image localization control device described above and the description thereof can be applied to the sound image localization control method.

  In the present invention, it is possible to improve the audibility with clear sound regardless of the type of content, and even when a single sound such as dialogue or announcement is included in the L / R channel.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The dimensions, materials, and other specific numerical values shown in the embodiment are merely examples for facilitating understanding of the invention, and do not limit the present invention unless otherwise specified. In the present specification and drawings, elements having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted, and elements not directly related to the present invention are not illustrated. To do.

  In recent years, together with the start of digital broadcasting, television broadcasting employing the surround system has been broadcast. Therefore, other than movies, for example, news programs, sports broadcasts, etc., are uniformly expanded to the surround system, and single voices such as dialogues and announcements may be assigned to L / R channels other than the center channel. . However, unlike a movie whose one of its purposes is to enjoy a sound with a sense of presence and a plurality of sound images, the single sound emitted by the commentator in these programs is required to be clearer than the presence.

  In the present embodiment, it is determined whether or not a single sound is included in the sound signal, and a head related transfer function corresponding to the degree to which the single sound is included is applied. Even when a single sound such as an announcement or an announcement is included in the L / R channel, it is possible to improve the audibility by a clear sound. Here, the head-related transfer function is an impulse response obtained by measuring an impulse signal emitted from an arbitrarily arranged speaker at the entrance to the ear canal of the listener. Here, in order to facilitate understanding, a virtual surround system adopting virtual surround will be described, and then a sound image localization control device constituting the virtual surround system will be described in detail.

(Surround System 100, Virtual Surround System 110)
FIG. 1 is a schematic diagram showing a configuration of a surround system 100 based on a 5.1 surround system. The surround system 100 includes a surround control device 148 that receives a television broadcast and extracts video and audio, a monitor 152 that displays the extracted video, and a speaker 154 that outputs the extracted audio. Is done.

  In the 5.1 surround system, a plurality of speakers 154 are arranged so as to surround the listener 160. For example, a center speaker 154a is located at the front center of the listener 160, and an L speaker 154b and an R speaker 154c are arranged on the left and right sides of the listener 160. SL speakers 154d and SR speakers 154e are arranged on the left and right sides. In addition, a subwoofer (LFE: Low Frequency Effect) 154f that outputs low-frequency sound is disposed at an arbitrary position. With the surround system 100, it is possible to improve the sense of presence of sound and reproduce a wide space.

  FIG. 2 is a schematic diagram showing the configuration of the virtual surround system 110. The virtual surround system 110 includes a sound image localization control device 150, a monitor 152, and two speakers 154 that output sound.

  In the virtual surround system 110, a virtual surround is reproduced by two speakers (L speaker 154 b and R speaker 154 c) installed in front of the listener 160 without arranging speakers behind the listener 160. The virtual surround system 110 employs a sound image localization technique using a head-related transfer function for localizing a sound image.

  In the virtual surround system 110, the sound image localization control device 150 can be formed integrally or separately with the monitor 152 and the speaker 154. Hereinafter, a specific configuration and operation of the sound image localization control device 150 will be described.

(Sound image localization control device 150)
FIG. 3 is a functional block diagram showing a hardware configuration of the sound image localization control device 150. The sound image localization control device 150 includes a broadcast receiving unit 210, an antenna 212, a video processing unit 214, an audio encoded signal extraction unit 216, a surround determination unit 218, a decoding unit 220, a correlation degree deriving unit 222, The storage unit 224, the table extraction unit 226, the sound image localization processing unit 228, and the amplification unit 230 are configured.

  The broadcast receiving unit 210 receives a television broadcast radio wave through the antenna 212 and separates the received signal into a compressed video signal and a compressed audio signal.

  The video processing unit 214 converts the compressed video signal received by the broadcast receiving unit 210 into a video signal and outputs the video signal to the monitor 152.

  The audio encoded signal extraction unit 216 extracts an audio encoded signal that is audio data from the television signal received by the broadcast receiving unit 210. Here, it is assumed that MPEG-2 AAC (standard ISO / IEC 13818-7), which is an international standard and adopted in a digital broadcasting system in Japan, is used as an audio encoding system.

  The surround determination unit 218 determines whether the audio encoded signal from the audio encoded signal extraction unit 216 corresponds to surround (multi-channel signal of 3 channels or more).

  The decoding unit 220 decodes the audio encoded signal and generates a linear PCM signal. Hereinafter, in order to facilitate understanding, the linear PCM signal is simply referred to as an audio signal.

  When the surround determination unit 218 determines that the sound is compatible with surround broadcasting, the correlation degree deriving unit 222 performs encoding using the correlation between the L / R channels from the auxiliary information of the audio encoded signal. A flag indicating whether or not a certain correlation encoding is applied to the audio encoded signal is extracted, and a correlation degree between the L / R channels is derived based on the extracted flag. In the present embodiment, the degree of correlation is an index indicating the level of correlation, and is represented by an integer from 1 to 5, for example. The lower the value, the lower the correlation, and the higher the value, the higher the correlation. .

  Usually, in a movie or the like, a sound image can be assumed at any number and at any position, and such a plurality of sound images improves the sense of presence of the listener. On the other hand, a single sound in a news program or a sports broadcast should be assumed to be a single and central sound image, and should be so.

  Here, the degree of correlation between the L / R channels is derived by the degree-of-correlation deriving unit 222, and a voice having a high degree of correlation is regarded as a single voice emitted from a single central sound image, and thereafter Sound is output using a head-related transfer function based on a group of coefficients that make the sound clear. Conversely, a voice with a low degree of correlation is regarded as not a single voice, and the voice is output using a normal head-related transfer function with a sense of presence. With this configuration, it is possible to improve the audibility with clear sound even when a single sound such as a dialogue or announcement is included in the L / R channel regardless of the type of content.

  Further, when the correlation encoding is performed for each of the plurality of frequency bands of the audio encoded signal, the correlation degree deriving unit 222 calculates the L / L based on the ratio of the frequency bands that have been subjected to the correlation encoding with respect to the entire frequency band. The degree of correlation between the R channels may be derived.

  With such processing, depending on the degree of correlation, not only simply whether the speech is a single speech, but also the probability that it is a single speech, that is, all the frequency bands to which correlation coding is applied. The coefficient group of the head related transfer function can be changed according to the ratio to the frequency band. With this configuration, it is possible to further improve the audibility with clear sound.

  Here, the correlation degree deriving unit 222 derives the correlation degree between the L / R channels based on the ratio of the frequency bands that have been subjected to the correlation coding with respect to the entire frequency band. The degree of correlation between the L / R channels may be derived based on the ratio of frames that have been subjected to correlation coding with respect to a predetermined number of frames in the axial direction. Further, in order to avoid the switching of the correlation degree in a complicated manner, the correlation degree deriving unit 222 can be provided with a low-pass filter and a hysteresis characteristic.

  The correlation encoding method described above is an MS stereo method. In AAC, the MS stereo method is used as the encoding method using the correlation between channels, and it can be easily derived whether or not the encoding is performed by the flag. The degree of correlation can be easily derived without complicated complicated calculations such as derivation of numbers. With this configuration, the processing load on the sound image localization control device 150 can be reduced.

  In addition, as the format of the bit stream constituting the encoded signal including the auxiliary information described above, two types of AAC-ADIF (Audio Data Interchange Format) and AAC-ADTS (Audio Data Transport Stream frame) are defined. In any format, a flag indicating whether or not encoding using correlation between channels is performed. Hereinafter, an example using a flag in AAC-ADTS will be described.

  FIG. 4 is an explanatory diagram showing a frame configuration of AAC-ADTS. As shown in (B), one AAC-ADTS frame 240 shown in (A) has three ADTS header (adts_header) 242, error detection word (crc: cyclic redundancy check) 244, and raw data block (raw_data_block) 246. Consists of parts. In the ADTS header 242, various pieces of frame information such as a synchronization signal and a sampling frequency are written. The error detection word 244 is used for error check of the ADTS frame. In the raw data block 246, compressed audio data and an identifier indicating the type of the compressed audio data are written.

  The compressed audio data has a different compression method depending on the type of data. Here, as shown in (C), the internal structure of the raw data block when the identifier indicating the type of the compressed audio data of the raw data block is CPE (Channel Pair Element) 248 is shown in (D) and (E). . The CPE 248 is used when encoding two channels having a high correlation, such as the L / R channel, together.

  The raw data block 246 includes an element instance tag (element_instance_tag) 250 with a unique number for identification when a channel other than the two channels uses the same identifier (here, CPE 248), and a block used for frequency conversion. A common window (common_window) 252 that is a flag indicating whether or not a common window type for long and windowing is used, and an ICS info (area for storing the number of subbands (number of band divisions)). ics_info) 254, MS mask present (ms_mask_present) 256 represented by 2 bits as a flag indicating whether or not encoding by MS stereo is performed, and whether or not encoding by MS stereo is also performed As a flag to indicate, MS use which is a flag indicated for each frequency division band (Ms used) 258, composed of Individual channel stream (individual_channel_stream) 260 that includes quantized data.

  Here, (D) shows a case where the common window 252 is 1 and the MS mask present 256 is 10 (binary number) indicating that “encoding by MS stereo is effective for all bands”. In this case, since encoding by MS stereo is used in the entire band, it can be estimated that the correlation of the L / R channel is high. Therefore, in this embodiment, the degree of correlation is set to a maximum value (for example, 5).

  Further, (E) shows a case where the common window 252 is 1 and the MS mask present 256 is 01 (binary number) indicating that “encoding by MS stereo is partially effective”. In this case, MS used 258 is set for each frequency division band, and if MS used 258 is 1, it indicates that encoding by MS stereo is effective in the frequency division band. In the present embodiment, the degree of correlation is derived based on the ratio of frequency division bands in which the MS used 258 is 1 in all frequency bands. The MS used 258 is expressed as MS used [g] [sfb]. Here, g is a window number, and sfb is a number for identifying a frequency division band.

  FIG. 5 is an explanatory diagram showing an example of the flag correlation degree table 262 stored in the storage unit 224. The relationship between the ratio 264 of the MS used 260 and the correlation degree 266 when the MS mask present 256 is 10 (binary number) is shown.

  When the ratio 264 of the frequency division band in which the MS used 260 is 1 is 100 to 80%, the correlation degree 266 is set to 5 which is the maximum value. Similarly, when the frequency division band ratio 264 is 80 to 60%, the correlation degree 266 is 4, when the frequency division band ratio 264 is 60 to 40%, the correlation degree 266 is 3, and the frequency division band ratio 264 is The correlation degree 266 is 2 when 40 to 20%, and the correlation degree 266 is 1 when the frequency division band ratio 264 is 20 to 0%.

  Further, when the MS mask present 256 is other than 01 and 10, it can be estimated that there is no correlation between the channels. Therefore, the correlation degree 266 is 1 indicating no correlation.

  Furthermore, when deriving correlation degree 266 described above, correlation degree deriving unit 222 limits correlation to a predetermined frequency band of interest, or derives a degree of correlation 266 with a greater weight than outside the predetermined frequency band. At this time, the predetermined frequency band is 200 to 4000 Hz.

  FIG. 6 is an explanatory diagram schematically showing an example of the weighting for deriving the frequency band and the degree of correlation. The object of the present embodiment is to clarify the voice when the target voice is a single voice. Therefore, it is considered that the weighting with respect to 200 to 4000 Hz, which is the frequency band of human voice, is made larger than other frequency bands. The frequency band of human voice (sound pressure level) is mostly 200 to 4000 Hz. Therefore, the degree of correlation 266 is derived by multiplying the weighting characteristic values according to the frequency bands as shown in FIG. 6 so as to limit the calculation target of the degree of correlation 266 to about 200 to 4000 Hz.

For example, when the MS mask present 256 is 10, the frequency division bands in which the MS used 260 is 1 are added with the weights shown in FIG. Then, this value is divided by a value obtained by adding the weights shown in FIG. This divided value is expressed as a percentage, and the corresponding degree of correlation 266 is derived with reference to FIG.
By such processing, components other than sound such as background sound can be eliminated or reduced, and the accuracy of deriving the degree of correlation limited to sound can be improved.

The storage unit 224 includes a ROM, a RAM, an E ² PROM, a nonvolatile RAM, a flash memory, an HDD, and the like, and stores a program processed by the control unit, audio data, and the like. In addition, the storage unit 224 includes a flag correlation degree table 262 that associates the proportion 264 of the frequency band in which the MS usage 260 is 1, and the correlation degree 266, and a correlation degree 266 and head transmission described later. A coefficient group table in which the coefficient group of the function is associated is stored.

  The present embodiment not only determines whether or not the voice is a single voice according to the degree of correlation 266, but also increases the clarity in proportion to the probability of being a single voice, that is, in proportion to the value of the degree of correlation 266. Drop. Therefore, the coefficient group of the head related transfer function is changed according to the value of the correlation degree 266. In order to achieve this object, it is desirable that each coefficient of the coefficient group is originally a function having a correlation degree 266. However, an excessive amount of labor is required to derive the function, and only a plurality of steps of association is required. In this case, since the object of the present application can be achieved sufficiently, here, the correlation degree 266 and the coefficient group of the head-related transfer function are associated with each other by a plurality of levels of correspondence using the coefficient group table.

  FIG. 7 is an explanatory diagram showing an example of the coefficient group table 268 stored in the storage unit 224. Here, coefficient group 270 of the head related transfer function is associated with correlation degree 266 in the left column. The head-related transfer functions (1), (2), and (3) shown in the coefficient group table 268 correspond to a direct sound gain, an initial reflected sound gain, and a reverberant sound gain that will be described later with reference to FIG. ing. The coefficient group 270 of the head related transfer function will be described in detail later.

  The table extraction unit 226 extracts the coefficient group 270 of the head related transfer function from the coefficient group table 268 according to the correlation degree 266 derived by the correlation degree deriving unit 222. Thus, the head-related transfer function can be uniquely formed based on the correlation degree 266.

  When the speaker 154 is formed separately from the sound image localization control device 150 and the monitor 152 and the position thereof can be changed, the listener 160 listens to the sound with an arbitrary speaker arrangement as shown in FIG. It will be. The head-related transfer function described above changes depending on the positional relationship between the listener's head (both ears) and the speaker. Therefore, a plurality of head-related transfer functions are prepared for an arbitrary arrangement of a plurality of speakers. And the memory | storage part 224 memorize | stores such several head-related transfer functions corresponding to arrangement | positioning of a speaker, and the coefficient group table 268 corresponding to it.

  The table extraction unit 226 extracts the coefficient group 270 of the head-related transfer function selected and determined according to the selection input from the listener 160 from among the plurality of head-related transfer functions stored in the storage unit 224. The listener 160 can obtain clearer sound more appropriately by selecting an optimal head-related transfer function according to the position of the speaker 154.

  The sound image localization processing unit 228 performs the sound image localization processing using a head-related transfer function that reflects the coefficient group 270 extracted by the table extraction unit 226. Here, in order to output each channel of the 5.1 surround system to a plurality of speakers, for example, two speakers, it is converted into a sound signal of the virtual sound system.

  FIG. 8 is a control block diagram illustrating a configuration example of the sound image localization processing unit 228. In the control block diagram of FIG. 8, an input audio signal representing a 5.1ch surround sound source is transmitted to the head-related transfer function block 272, undergoes convolution processing in the head-related transfer function block 272, and the adder 274 After being synthesized, it is finally output as a 2ch virtual surround sound signal. Here, the head-related transfer functions (1), (2), and (3) correspond to a direct sound gain, an initial reflection sound gain, and a reverberation sound gain, which will be described later with reference to FIG. In the present embodiment, the head-related transfer function block 272 is provided for each of the three objects of the direct sound, the early reflection sound, and the reverberation sound. However, it is also possible to perform processing in one block in which these are mixed. It is.

  FIG. 9 is a control block diagram showing an internal configuration of the head-related transfer function block 272 of FIG. In the control block diagram of FIG. 9, first, an input audio signal representing a 5-channel surround sound source is delayed through a delay 280, and filter processing by the FIR filter 282 is performed on each of the L channel and the R channel. Then, the speech is synthesized by the adder 274 and output as a 2ch virtual surround sound signal.

  The delay 280 in the head related transfer function block 272 corresponds to an initial reflected sound delay amount and a reverberant sound delay amount which will be described later with reference to FIG. Since the delay amount of the direct sound does not occur, the delay 280 of the head related transfer function block 272 regarding the direct sound can be omitted. In this embodiment, an FIR (Finite Impulse Response) filter is used as the head-related transfer function, but an IIR (Infinite Impulse Response) can also be used.

  In order to reproduce a 5.1 surround sound signal with 2ch virtual surround, as shown in FIG. 2, in order to generate sound emitted from two speakers 154 prepared in front of the listener 160, the head transmission is performed. A function (related to head transfer for target) 5ch × 2 and a head related transfer function (reproduction head related transfer function) 4 (both ears × 2) corresponding to the speaker arrangement are required. Here, the reproduction head-related transfer function is a cancel signal for preventing the listener 160 from recognizing the position (sound source) of the speaker.

  Here, the head related transfer function can adjust the direct sound and the reflected sound of the sound signal independently by the coefficient group 270. Therefore, the sound image localization processing unit 228 can control the direct sound and the reflected sound independently by the coefficient group 270 extracted by the table extraction unit 226.

  FIG. 10 is an audio signal time waveform diagram showing an impulse response of the head-related transfer function. Such impulse responses are measured in concert halls, with the first half peak being the direct sound part, followed by the early reflections (reflections) that bounce off the walls, ceiling, floor, etc., and then called reverberation. Secondary reflected sound is distributed. The sense of reality is felt mainly by the effects of early reflections and reverberation.

  Thus, when the impulse response through the head-related transfer function is viewed in the time domain, the response waveform can be distinguished into direct sound and reflected sound. In this embodiment, reverberation due to reflection is suppressed by adjusting the coefficient of the direct sound and the reflected sound in the head-related transfer function and lowering the ratio of the reflected sound to the direct sound. Specifically, the direct sound and the early reflection sound (reflection sound) are distinguished, and the direct sound gain, the initial reflection sound gain and the delay amount with respect to the direct sound, the reverberation sound gain after the initial reflection sound and the initial reflection sound are distinguished. The delay amount is adjusted independently.

  Adjustment of the gain of the early reflection sound and reverberation sound described above is affected by the measurement conditions of the head related transfer function and the size of the space to be reproduced. Thus, when the ratio of the reflected sound is lowered, the sound becomes clearer, and when it is increased, the presence of the sound is enhanced. As an example of a specific processing method, the direct sound component maintains its value, and the ratio 264 of the frequency division band in which the MS used 260 is 1 with respect to the entire frequency band used in the derivation of the correlation 266 described above is used. A value obtained by subtracting from 100% is multiplied as a gain coefficient by the early reflection sound and reverberation sound which are reflection components. In this way, in the case of a single voice in which the ratio 264 of the frequency division band in which the MS usage 260 is 1 for all the frequency bands is close to 100%, the reflection component is suppressed, and an audio signal with direct sound-oriented clarity can be obtained. Can do.

  Here, the initial reflection sound and the reverberation sound are distinguished and processed, but they can be processed as a series of waveforms, and the reverberation sound portion can be omitted to reduce the processing amount. Sound image localization processing includes the method of convolving the time axis signal as it is and the method of convolution on the frequency axis. Generally, the convolution operation on the frequency axis can suppress the processing amount compared to the time axis, but either method can be used. Even if it exists, the objective of this embodiment can be achieved and it can be decided arbitrarily which is used. Further, as described above, even if the coefficient group 270 corresponding to the degree of correlation 266 is selected from the coefficient group table 268, the MS transfer 260 for all frequency bands is 1 as the head related transfer function used for convolution. A coefficient group obtained by multiplying a value obtained by subtracting the frequency division band ratio 264 from 100% by multiplying the initial reflected sound and the reverberant sound, which are reflection components, as gain coefficients may be used. The former can reduce the processing load for deriving the coefficient group 270 by referring to the coefficient group table 268, and the latter does not store the coefficient group table 268, so that the memory area to be used can be saved.

  The amplifying unit 230 amplifies the virtual sound system audio signal converted by the sound image localization processing unit 228 to a signal level that can be output to the speaker 154.

  The sound image localization control device 150 described above improves the audibility with clear sound regardless of the type of content, and even when a single sound such as speech or announcement is included in the L / R channel. Is possible.

(Sound image localization control method)
Next, a sound image localization control method for performing sound image localization processing on a television broadcast that supports surround using the above-described sound image localization control device 150 will be specifically described.

  FIG. 11 is a flowchart for explaining the processing flow of the sound image localization control method. When the broadcast receiving unit 210 of the sound image localization control device 150 receives a television broadcast (YES in S300), the surround determination unit 218 determines whether the audio signal of the television broadcast received by the broadcast receiving unit 210 is compatible with surround. (S302), if it is determined that the signal corresponds to surround (YES in S302), the correlation degree deriving unit 222 extracts the L / R channel of the audio signal, and further, from the auxiliary information. A flag indicating whether or not encoding by the MS stereo method is performed is extracted (S304). When the encoding by the MS stereo method is not performed in any frequency band (S306 YES), the correlation degree 266 is set to 1 (S308). In addition, when the encoding by the MS stereo method has been performed (NO in S306), and the encoding target is the entire frequency band (S310 YES), the correlation degree 266 is set to 5 (S312).

  Then, when the encoding by the MS stereo method is performed or not performed for each frequency band (NO in S310), the ratio 264 of the frequency band subjected to the correlation encoding for all the frequency bands, and the storage unit The correlation degree 266 is determined based on the flag correlation degree table 262 stored in the H.224 (S314).

  Then, the table extraction unit 226 extracts the coefficient group 270 of the head related transfer function from the coefficient group table 268 stored in the storage unit 224 according to the correlation degree 266 derived by the correlation degree deriving unit 222 (S316). The sound image localization processing unit 228 executes sound image localization processing using a head-related transfer function that reflects the extracted coefficient group 270 (S318). Then, the sound signal subjected to the sound image localization processing is provided to the listener 160 through the speaker 154 (S320).

  With this sound image localization control method, it is possible to improve the audibility with clear audio regardless of the type of content, and even when single speech such as dialogue or announcement is included in the L / R channel. It becomes.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this embodiment. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Understood.

  In the above-described embodiment, the determination as to whether or not the program requires clarity is performed only by the similarity deriving unit 222. However, the determination is not limited to such a case. For example, content type information is received as data. However, when the received type is a program that requires clarity, such as a news program, the processing content of the sound image localization process may be changed regardless of the similarity derived by the similarity deriving unit 222.

  Note that each step in the sound image localization control method of the present specification does not necessarily have to be processed in time series in the order described in the flowchart, and may include parallel or subroutine processing.

  INDUSTRIAL APPLICABILITY The present invention can be used for a sound image localization control device and a sound image localization control method that perform sound image localization processing on a television broadcast that supports surround.

It is the schematic diagram which showed the structure of the surround system by a 5.1 surround system. It is the schematic diagram which showed the structure of the virtual surround system. It is the functional block diagram which showed the hardware constitutions of the sound image localization control apparatus. It is explanatory drawing which shows the frame structure of AAC-ADTS. It is explanatory drawing which showed an example of the flag correlation degree table memorize | stored in the memory | storage part. It is explanatory drawing which represented typically an example of the weighting of a frequency band and correlation degree derivation. It is explanatory drawing which showed an example of the coefficient group table memorize | stored in the memory | storage part. It is the control block diagram which showed the structural example of the sound image localization process part. FIG. 9 is a control block diagram illustrating an internal configuration of the head-related transfer function block of FIG. 8. It is an audio | voice signal time waveform figure which showed the impulse response of the head-related transfer function. It is the flowchart explaining the flow of processing of the sound image localization control method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 150 ... Sound image localization control apparatus 152 ... Monitor 154 ... Speaker 210 ... Broadcast receiving part 216 ... Audio coding signal extraction part 218 ... Surround determination part 222 ... Correlation degree derivation part 224 ... Storage part 226 ... Table extraction part 228 ... Sound image localization process Unit 266 ... degree of correlation 268 ... coefficient group table 270 ... coefficient group

Claims (8)

A broadcast receiver capable of receiving television broadcasts;
An audio encoded signal extracting unit for extracting audio data of the television broadcast received by the broadcast receiving unit;
A surround determination unit that determines whether or not the audio encoded signal is compatible with surround;
When it is determined that the audio signal corresponds to surround, correlation encoding, which is encoding using correlation between L / R channels, is applied to the audio encoded signal from auxiliary information of the audio encoded signal. A correlation degree deriving unit that extracts a flag indicating whether there is a correlation between the L / R channels based on the extracted flag;
A storage unit for storing a coefficient group table in which the degree of correlation and the coefficient group of the head related transfer function are associated;
A table extracting unit that extracts a coefficient group of a head related transfer function from the coefficient group table according to the degree of correlation derived by the correlation degree deriving unit;
A sound image localization processing unit that performs sound image localization processing by a head-related transfer function reflecting the extracted coefficient group;
A sound image localization control device comprising:   When the correlation encoding is performed for each of a plurality of frequency bands of the audio encoded signal, the correlation degree deriving unit determines the L based on the ratio of the frequency bands subjected to the correlation encoding with respect to the entire frequency band. The sound image localization control apparatus according to claim 1, wherein a degree of correlation between the / R channels is derived.   The sound image localization control apparatus according to claim 1 or 2, wherein the audio encoded signal is encoded by an AAC method, and the correlation encoding method is an MS stereo method.   The sound image localization control apparatus according to any one of claims 1 to 3, wherein the correlation degree deriving unit derives the correlation degree by weighting a predetermined frequency band larger than other frequency bands. .   The sound image localization control apparatus according to claim 4, wherein the predetermined frequency band is 200 to 4000 Hz.   6. The sound image localization control device according to claim 1, wherein the head-related transfer function can independently adjust a direct sound and a reflected sound of the audio signal by the coefficient group. . The storage unit includes a coefficient group table corresponding to a plurality of head related transfer functions corresponding to the arrangement of speakers,
The table extraction unit extracts a coefficient group of a head-related transfer function selected and determined according to a speaker arrangement from the plurality of head-related transfer functions. The sound image localization control apparatus according to the item. Receive television broadcasts,
Extracting an audio encoded signal that is audio data of the received television broadcast,
Determining whether the audio encoded signal is compatible with surround;
When it is determined that the audio signal corresponds to surround, correlation encoding, which is encoding using correlation between L / R channels, is performed on the audio encoded signal from the auxiliary information of the audio encoded signal. A flag indicating whether or not there is, and based on the extracted flag, a correlation degree between the L / R channels is derived,
According to the derived degree of correlation, the coefficient group of the head related transfer function is extracted from the coefficient group table that associates the degree of correlation with the coefficient group of the head related transfer function,
A sound image localization control method, wherein sound image localization processing is executed by a head-related transfer function reflecting the extracted coefficient group.

Images