JP2008096567A - Audio encoding device and audio encoding method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve quality of an encoded audio signal. <P>SOLUTION: An audio encoding device includes: a process for creating a sub-band signal of a high frequency section by converting an input signal to the sub-band signal, and by copying the sub-band signal of a low frequency section; and a process for adding an artificial tone signal in order to regenerate a tone nature signal which is included in the input signal, and which is not regenerated by copying the sub-band signal, wherein by newly providing a means for analyzing a harmonics structure of the input signal, it is judged whether or not the input signal has the harmonic structure. When it has the harmonics structure, disturbance of a harmonics interval is prevented and sound quality is improved by not adding the tone signal. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an audio encoding apparatus and method for efficiently encoding an audio signal with a small amount of information.

  An SBR (Spectral Band Replication) system is known as a system that can compress and encode audio signals with high efficiency and achieve high quality even at low bit rates. SBR is an encoding system defined in the international standard MPEG-4 Audio (ISO / IEC 14496-3: 2001 / AMD1: 2003) (see Non-Patent Document 1) of an audio encoding system. SBR is characterized by the fact that it is used in combination with a conventional encoding method, for example, AAC (Advanced Audio Coding), which is also defined by MPEG-4 Audio. The low frequency band is encoded by an encoding scheme such as AAC and the high frequency band is encoded by SBR.

  At this time, on the encoding side, only the energy envelope is encoded without encoding the input signal itself for the high frequency part. On the decoding side, first, the low-frequency part signal is decoded, and then the SBR part replicates the decoded low-frequency signal to the high-frequency part and shapes the signal according to the decoded energy envelope. In this way, in SBR, only the energy envelope of the high frequency band signal is encoded, so that it is possible to encode with a very small amount of information compared to the conventional method of encoding the input signal itself. Become. As a result, a wideband audio signal in the 0-16 kHz band can be encoded with high quality at a low bit rate such as 48 kbps.

  Reconstructing the high frequency signal by duplicating the low frequency signal is usually based on the human auditory characteristics due to the similarity between the low frequency and high frequency of the input signal, for example, the similarity of the harmonic structure. It is not perceived as a major deterioration. However, when the characteristics of the low-frequency part and the high-frequency part of the input signal are significantly different, for example, when a characteristic signal exists only in the high-frequency part, the high-frequency band generated as a replica from the low-frequency part The signal of the part will be greatly different from the input signal, and will be perceived as sound quality degradation. A characteristic signal that exists only in the high frequency part is a tone signal, which is generally present in music played by an electronic musical instrument.

  In order to cope with such a problem, in SBR, in encoding, a tone-like signal included in a high frequency part of an input signal is detected, and its temporal and frequency existing positions are encoded, and in decoding, The tone signal generated artificially can be added to the high frequency band signal copied from the low frequency band (referred to as tone addition). As a specific method of adding a tone signal, for example, there is a method disclosed in Patent Document 1. Hereinafter, the process will be described with reference to the drawings.

  FIG. 2 is a diagram showing a configuration of tone signal addition in the conventional SBR. The input signal 206 is input to the subband filter 201 and converted into a subband signal. The sub-band signal is divided into a high-frequency part 207 and a low-frequency part 208 and input to the high-frequency tone analysis unit 202 and the low-frequency tone analysis unit 203, respectively. As the degree of tonality (degree of tonality), for example, the magnitude of the linear prediction gain for the subband signal is used. The calculated tone characteristic degree 209 of the high frequency signal and the tone characteristic degree 210 of the low frequency signal are input to the tone signal addition determination unit 204. The tone signal addition determination unit 204 compares the tone characteristic degree 209 of the high frequency signal and the tone characteristic degree 210 of the low frequency signal, and outputs the tone addition flag information 212.

  Here, the degree of tone to be compared is determined by which subband signal in the low frequency band is copied to which subband in the high frequency band in decoding. FIG. 3 is a diagram showing the relationship between the subband signals on the encoding side and the decoding side, and is divided into a high frequency part and a low frequency part by a boundary 303. In the subband signal 302 on the decoding side, the signal 307 in the low band subband m is duplicated as the signal 306 in the high band subband n. In the subband signal 301, a signal 305 in the subband m and a signal 304 in the subband n are obtained. In the subband signal on the encoding side, when the tone property of the signal 304 of the subband n is high and the tone property of the signal 305 of the subband m is low, the signal 304 of the input subband n is high in tone. The sub-band n signal 306 generated in the decoding has a low tone property, and the sound quality is deteriorated due to the difference in characteristics between the input signal and the decoded signal. Therefore, it can be determined that, under such conditions, it is necessary to add an artificial tone signal to the sub-band n signal 306 generated by decoding to enhance the tone characteristics of the generated signal. The presence / absence of tone addition for each subband is encoded and output as on / off flag information.

In SBR defined by MPEG-4, encoding is performed in units called scale factor bands, which are a group of a plurality of adjacent subbands. In this case, since the flag indicating the presence / absence of tone addition is also set for each scale factor band, it is necessary to determine tone characteristics for each scale factor band. For example, the subband signal having the highest tone characteristic is selected from the subband signals included in the high-frequency part scale factor band to be compared, and the tone characteristic degree is set as a representative value of the high-frequency part scale factor band. . Similarly, a subband having the highest tone characteristic is selected from the subband signals included in the corresponding low-frequency scale factor band, and the tone characteristic degree is selected as a representative value of the low-frequency scale factor band. This can be realized by comparing with the representative value of the scale factor band.
Japanese translation of PCT publication No. 2005-51077 International standard ISO / IEC 14496-3: 2001 / AMD1: 2003

  When the tone addition flag is set for a scale factor band that is a group of multiple adjacent subbands, one subband (usually centered as a frequency position) among the multiple subbands included in the scale factor band. A tone signal is added only to the existing subband. The frequency position of the additional tone signal is the center frequency of the subband. Therefore, when the frequency position of the tone signal included in the input signal is at the end of the scale factor band, the frequency shift from the additional tone signal becomes large.

  In a signal of a relatively high frequency band handled by SBR (for example, about 10 kHz or more), the frequency resolution of human auditory characteristics is reduced. Therefore, a slight frequency position shift in a single tone signal leads to a large deterioration in sound quality. Absent. However, the frequency position shift of a plurality of tone signals having a harmonic structure may greatly affect sound quality. In general, a signal having a frequency spectrum structure called a harmonic structure or a harmonic structure is a collection of a primary signal serving as a fundamental tone and harmonics from secondary to n-order having a frequency that is an integral multiple of the fundamental tone.

  FIG. 4 is a diagram illustrating a relationship between a signal duplicated from a low frequency region and an additional tone signal in a signal having a harmonic structure. Here, it is assumed that tone addition is performed on the scale factor band k, 401 is a tone addition signal, 402 is a signal copied from the low frequency part, and 403 is a signal obtained by combining 401 and 402. Reference numeral 404 denotes a subband width, and reference numeral 405 denotes a scale factor band width. The added tone signal 406 is arranged at the center frequency position of the center subband of the scale factor band k. In addition, since the signal 402 replicated from the low frequency part has a harmonic structure, it includes tone signals represented by 407, 408, 409, and 410. At this time, for example, an interval 415 between 407 and 408 is equal to an interval 416 between 408 and 409.

  When tone addition by SBR is performed on a signal having such a harmonic structure, an interval between harmonics is disturbed due to a shift in frequency position of the added tone signal. The synthesized signal 403 is obtained by replacing the signal 402 of the scale factor band k obtained by performing tone addition with respect to the signal 402 copied from the low frequency portion with the tone addition signal 401. The tone signal included is 411. 412, 413, and 414. The interval 417 between 411 and 412 is equal to the harmonic interval 415 of the signal replicated from the low band, whereas the interval 418 between 412 and 413 is equal to the harmonic interval 416 of the signal replicated from the low band. Will be different.

  Disturbances in the harmonic spacing cause interference between adjacent tone signals, and as a result, a phenomenon that greatly impairs sound quality, such as a sense of distortion or a sense of undulation, may occur. In such a case, it is better to generate the high frequency band by duplicating from the low frequency band without adding the tone signal, because the spacing between the harmonics is maintained at least in the high frequency band. There is little decrease in However, in the conventional configuration, since the tone signal addition is determined in units of subbands or scale factor bands, it cannot be determined whether the input signal has a harmonic structure, and the harmonic interval is disturbed. A tone signal may be added at such a position.

  An object of the present invention is to solve such problems, and by newly providing means for analyzing the harmonic structure of an input signal, it is determined whether or not the input signal has a harmonic structure. In the case of having a harmonic structure, by avoiding the addition of a tone, disturbance of harmonic intervals is prevented and sound quality is improved.

  According to the configuration of the present invention, encoding can be performed without disturbing the harmonic interval of the input signal, and the quality of the encoded audio signal can be improved.

(Embodiment 1)
FIG. 1 is a diagram illustrating an example of a configuration according to Embodiment 1 of the present invention. Compared to the conventional configuration shown in FIG. 2, means 105 for analyzing the harmonic structure of the input signal is provided, and the analysis result 111 of the harmonic structure is output and input to the tone signal adding means. Different. Hereinafter, the operation will be described in detail.

  The input signal 106 is divided into a high-frequency subband signal 107 and a low-frequency subband signal 108 by a subband filter 101, and a high-frequency tone analysis unit 102 and a low-frequency tone analysis, respectively. Input to means 103. Each tone property analyzing means calculates the degree of tone properties 109 and 110 in units of subbands for the input subband signal. As the degree of tonality (degree of tonality), for example, the magnitude of the linear prediction gain for the subband signal is used. The harmonic structure analysis means 105 analyzes the high band sub-band signal, particularly encoded by SBR, to determine whether a harmonic structure exists.

  As a specific determination method, for example, an average value of tone characteristics 109 of subbands included in the high frequency part is calculated, and compared with a predetermined threshold value. Check whether it shows gender. Since a signal having a harmonic structure is an aggregate of many tone characteristics signals, it exhibits high tone characteristics on average, so if the average value of the calculated tone characteristics indicates high tone characteristics, it is input. It can be determined that the signal has a harmonic structure.

  In addition, in the calculation of the average value of the tone characteristics, a part of the tone characteristics of the low frequency subbands may be added in addition to the tone characteristics of the high frequency subbands. The boundary between the high and low frequencies is often set to about 6 to 10 kHz, but the frequency band near this is a band where harmonic structures are strongly observed in many music pieces. If the frequency is set high (for example, 10 kHz), it is better to analyze the presence or absence of the harmonic structure by analyzing the high band from a frequency slightly lower than the boundary (for example, 8 kHz), which is classified as a low frequency part. Judgment can be made.

  The result 111 of the harmonic structure analysis is output as binary information indicating whether or not it has a harmonic structure. The tone signal addition determination unit 104 switches whether to add a tone according to the presence or absence of a harmonic structure, and outputs tone addition flag information 112. As a preferable example of switching, first, the tone characteristic degree 110 of the low-frequency sub-band signal is compared with the tone characteristic degree 109 of the high-frequency sub-band signal to determine the sub-band to which the tone signal needs to be added. If there is no harmonic structure, this determination is the final determination as it is. In the case of having a harmonic structure, subsequently, in the peripheral subbands of the high-frequency subband determined to be added with the tone signal first, for example, in the subbands on both sides, the low-frequency subband to be compared Investigate the degree of tone. If the tone characteristic of the low frequency band subband is large, the tone signal included in the subband signal replicated from the low frequency band is likely to interfere with the additional tone signal. judge.

  In addition, since switching based on binary information increases the possibility of sound quality degradation due to a determination error, the result of harmonic structure analysis is output as the degree of intensity of the harmonic structure, and the tone signal adding means is more detailed. A determination can also be made. As a determination method based on the intensity level of the harmonic structure, for example, as the intensity level of the harmonic structure increases, a threshold value used as a reference for tone addition determination is controlled in a direction in which tone addition is not performed. You can do it.

  The configuration used in the above description performs tone addition determination in units of subbands. However, in the configuration in which tone addition determination is performed in units of scale factor bands, the tone characteristic degree to be compared is determined by the scale factor. By replacing the band with a representative value, the same configuration can be realized. As a value representative of the scale factor band, the maximum value or the average value of the tone characteristics of the subband signals included in the scale factor band can be used.

  The configuration used in the above description is realized using a subband signal generated by a subband filter. However, any known time-frequency conversion such as Fourier transform or cosine transform can be used instead of the subband filter. It is obvious to those skilled in the art that a similar configuration can be realized even when using.

  Although the present invention has been described based on the above embodiment, it is needless to say that the present invention is not limited to the above embodiment. The following cases are also included in the present invention.

  (1) Each of the above devices is specifically a computer system including a microprocessor, a ROM, a RAM, a hard disk unit, a display unit, a keyboard, a mouse, and the like. A computer program is stored in the RAM or hard disk unit. Each device achieves its functions by the microprocessor operating according to the computer program. Here, the computer program is configured by combining a plurality of instruction codes indicating instructions for the computer in order to achieve a predetermined function.

  (2) A part or all of the components constituting each of the above devices may be configured by one system LSI (Large Scale Integration). The system LSI is a super multifunctional LSI manufactured by integrating a plurality of components on one chip, and specifically, a computer system including a microprocessor, a ROM, a RAM, and the like. . A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating according to the computer program.

  (3) Part or all of the constituent elements constituting each of the above devices may be configured from an IC card that can be attached to and detached from each device or a single module. The IC card or the module is a computer system including a microprocessor, a ROM, a RAM, and the like. The IC card or the module may include the super multifunctional LSI described above. The IC card or the module achieves its function by the microprocessor operating according to the computer program. This IC card or this module may have tamper resistance.

  (4) The present invention may be the method described above. Further, the present invention may be a computer program that realizes these methods by a computer, or may be a digital signal composed of the computer program.

  The present invention also provides a computer-readable recording medium such as a flexible disk, a hard disk, a CD-ROM, an MO, a DVD, a DVD-ROM, a DVD-RAM, a BD (Blu-ray Disc). ), Recorded in a semiconductor memory or the like. The digital signal may be recorded on these recording media.

  In the present invention, the computer program or the digital signal may be transmitted via an electric communication line, a wireless or wired communication line, a network represented by the Internet, a data broadcast, or the like.

  The present invention may be a computer system including a microprocessor and a memory, wherein the memory stores the computer program, and the microprocessor operates according to the computer program.

  In addition, the program or the digital signal is recorded on the recording medium and transferred, or the program or the digital signal is transferred via the network or the like, and executed by another independent computer system. It is good.

  (5) The above embodiment and the above modifications may be combined.

  A device that compresses and encodes a voice or audio signal with a small amount of information, stores it in a storage medium, extracts it from the storage medium directly or via a transmission line, and decodes the original voice or audio signal, for example, a mobile phone It can also be applied to general devices such as music players. Specifically, it is suitable for use in audio / music players using optical disks, magnetic disks, semiconductor memories, etc. as storage media, broadcasting of audio / music / video, on-demand distribution, and the like.

The block diagram which shows the structure of the encoding apparatus of this invention Block diagram showing the configuration of a conventional encoding device Conceptual diagram showing the relationship between the low-frequency and high-frequency sub-band signals Conceptual diagram showing the relationship between the duplicate signal from the low-frequency region and the tone addition signal

Explanation of symbols

101, 201 Subband filter 102, 202 High frequency tone analysis means 103, 203 Low frequency tone analysis means 104, 204 Tone signal addition availability determination means 105 Harmonic structure analysis means 106, 206 Input signal 107, 207 High frequency sub Band signal 108, 208 Low frequency sub-band signal 109, 209 High frequency signal tone characteristic 110, 210 Low frequency signal tone characteristic 111 Harmonic structure analysis result 112, 212 Tone addition flag information 301 Encoding side subband Signal 302 Decoding side subband signal 303 Low band and high band boundary 304 Coding side subband n
305 Encoding side subband m
306 Decoding side subband n
307 Decoding side subband m
401 Tone added signal 402 Duplicated signal from low frequency 403 Combined signal 404 Subband 405 Scale factor band 406 Added tone signal 407, 408, 409, 410 Tone signals 411, 412 included in duplicated signal from low frequency 413, 414 Tone signal included in synthesized signal 415, 416 Harmonic spacing of duplicate signal from low frequency 417, 418 Harmonic spacing of synthesized signal

Claims (8)

An input signal is divided into a low-frequency signal that is a low-frequency band signal and a high-frequency signal that is a high-frequency signal, and the high-frequency signal is divided from the low-frequency signal. An audio encoding device for restoring using a copy of
Means for converting an input signal into a sub-band signal in a low-frequency part and a sub-band signal in a high-frequency part; first tone characteristic analyzing means for calculating a tone characteristic of the sub-band signal in the low-frequency part; and A second tone characteristic analyzing means for calculating a tone characteristic degree of the sub-band signal in the high frequency part; a harmonic structure analyzing means for analyzing a harmonic structure of the input signal; and an artificial signal for the sub-band signal in the high frequency part. Tone signal addition availability determination means for determining whether or not to add a typical tone signal,
The tone signal addition availability determination means compares the calculated tone characteristics of the low-frequency subband signal with the calculated tone characteristics of the high-frequency subband signal, and the harmonics An audio encoding device that determines whether or not to add a tone signal based on a result of harmonic structure analysis by the structure analysis means. The harmonic structure analysis means calculates a degree indicating the tone characteristic of the entire high band part from the tone characteristic degree of each sub-band signal of the high band part, and compares it with a predetermined threshold value to thereby calculate the input signal. The audio encoding device according to claim 1, wherein it is determined whether or not has a harmonic structure. The degree of the tone characteristic of the entire high band calculated by the harmonic structure analyzing means is a value obtained by averaging the tone characteristics of each subband signal of the high band in the entire high band. The audio encoding device according to claim 2. The harmonic structure analysis means calculates a degree indicating the tone characteristic of the entire included range from the degree of tone characteristic of each subband signal in the high frequency part and the tone characteristic degree of some subbands in the low frequency part. The audio encoding apparatus according to claim 1, wherein it is determined whether or not the input signal has a harmonic structure by comparing with a predetermined threshold value. When the input signal has a harmonic structure, the tone signal addition enable / disable determining unit is configured to add a tone signal to a low frequency subband corresponding to a peripheral subband of the high frequency subband to which a tone signal is to be added. 3. The audio encoding apparatus according to claim 2, wherein a degree of tone is investigated, and when a tone degree of a low-frequency subband is high, it is determined not to add a tone signal. The harmonic structure analysis means outputs the analysis result of the harmonic structure as a degree of intensity of the harmonic structure, and the tone signal addition determination means determines whether the tone signal increases as the degree of intensity of the harmonic structure increases. The audio encoding device according to claim 1, wherein the audio encoding device is determined so as to be difficult to add. An input signal is divided into a low-frequency signal that is a low-frequency band signal and a high-frequency signal that is a high-frequency signal, and the high-frequency signal is divided from the low-frequency signal. An audio encoding method for reconstructing using a copy of
Converting an input signal into a low-frequency subband signal and a high-frequency subband signal; a first tone analysis step of calculating a tone characteristic of the low-frequency subband signal; and A second tone characteristic analyzing step for calculating a tone characteristic degree of the high-frequency subband signal; a harmonic structure analyzing step for analyzing a harmonic structure of the input signal; and an artificial signal for the high-frequency subband signal. A tone signal addition determination step for determining whether or not to add a typical tone signal,
The step of determining whether or not to add a tone signal includes comparing the calculated tone characteristic of the low-frequency subband signal and the calculated tone characteristic of the high-frequency subband signal with the harmonics. An audio encoding method comprising: determining whether to add a tone signal based on a result of harmonic structure analysis in the wave structure analysis step. A program for causing a computer to execute each step included in the audio encoding method according to claim 7.

Images