JP4220108B2 - Acoustic signal coding system - Google Patents

Description

[0001]
[Industrial application fields]
The present invention includes broadcast media (radio, television), communication media (CS video / audio distribution, Internet music distribution, communication karaoke), package media (CD, MD, cassette, video, LD, CD-ROM, game cassette, mobile phone). Production of various audio contents provided by a solid-state memory medium for music players), music content including vocals for dedicated mobile music players, mobile phones, PHS, pagers, literary arts such as Kabuki, Noh, Reading, Poetry The present invention relates to an audio signal encoding technique suitable for use in MIDI transmission of audio material of a work or audio teaching material for language education.
[0002]
[Prior art]
A time-series signal represented by an acoustic signal includes a plurality of periodic signals as its constituent elements. For this reason, a method for analyzing what kind of periodic signal is included in a given time-series signal has been known for a long time. For example, Fourier analysis is widely used as a method for analyzing frequency components included in a given time series signal.
[0003]
By using such a time-series signal analysis method, an acoustic signal can be encoded. With the spread of computers, it has become easy to sample an analog audio signal as the original sound at a predetermined sampling frequency, quantize the signal intensity at each sampling, and capture it as digital data. If a method such as Fourier analysis is applied to the data and the frequency components included in the original sound signal are extracted, the original sound signal can be encoded by a code indicating each frequency component.
[0004]
On the other hand, the MIDI (Musical Instrument Digital Interface) standard, which was born from the idea of encoding musical instrument sounds by electronic musical instruments, has been actively used with the spread of personal computers. The code data according to the MIDI standard (hereinafter referred to as MIDI data) is basically data that describes the operation of the musical instrument performance such as which keyboard key of the instrument is played with what strength. The data itself does not include the actual sound waveform. Therefore, in order to reproduce the actual sound, a MIDI sound source that stores the waveform of the instrument sound (distortion waveform pattern) is required separately, but its high encoding efficiency has attracted attention, and is based on the MIDI standard. Encoding and decoding techniques are currently widely used in software that performs musical instrument performance, musical instrument practice, composition, etc. using a personal computer.
[0005]
Therefore, by analyzing a time-series signal represented by an acoustic signal by a predetermined method, a periodic signal as a constituent element is extracted, and the extracted periodic signal is encoded using MIDI data. Proposals have been made. For example, JP-A-10-247099, JP-A-11-73199, JP-A-11-73200, JP-A-11-95753, JP-A-12-99009, JP-A-12-99093, In Japanese Patent Application No. 11-58431, Japanese Patent Application No. 11-177875, and Japanese Patent Application No. 11-329297, the frequency as a component is analyzed for an arbitrary time series signal, and the analysis is performed. Various methods for creating MIDI data from the results have been proposed.
[0006]
[Problems to be solved by the invention]
The MIDI encoding method proposed in each of the above publications or specifications has made it possible to reproduce music including vocals, but in order to apply to audio encoding in general, a multitrack that can reproduce a plurality of instrument sounds Formal encoding needs to be realized. In order to realize this, there are two problems of “sound source separation technology” and “MIDI encoding of a single musical instrument”, and an ideal MIDI code that recognizes a musical note played from an instrument sound and reproduces it with a MIDI instrument. The realization of the conversion format is extremely difficult at present.
[0007]
As for sound source separation technology, it is necessary to separate sound sources from mixed performance recordings for each instrument part and vocal part and perform MIDI encoding for each part, but this is technically impossible at present and can be realized to some extent. There is a problem that imperfection remains even if it becomes. As for MIDI encoding of single music instruments, stringed instruments such as piano, guitar, violin have a low ratio of harmonics and are easy to perform MIDI encoding, but brass instruments such as brass containing many overtones and percussion instruments that are noise sources However, there is a problem that MIDI encoding is difficult.
[0008]
On the other hand, an analysis / synthesis coding method for spectrally decomposing an acoustic signal into a sine wave and coding spectrum information has already been put to practical use in MP3 (MPEG-1 layer 3) and the like. The amount of information is smaller than that of waveform coding methods such as PCM and ADPCM, and the reproduction speed and pitch can be changed as in the case of MIDI. However, this method has the following three problems.
[0009]
First, there is a problem that the compression ratio cannot be increased so much and the practical compression ratio is about 1/10 even with MP3. The reason for this is that musical instrument sounds and the like need to faithfully encode overtone components that continue indefinitely to the maximum audible frequency, which is a disadvantage of the idea of decomposing signals into sine waves. Second, there is a problem of the combined processing load of the decomposed sine wave. On the decoding side, it is necessary to synthesize 100 or more sine waves subjected to spectral decomposition in real time. This technique has already been established, but the decoding process is complicated. Third, there is a problem of sound reproduction in the hypersonic region. The maximum human audible frequency is 22 kHz, so the sampling frequency of the CD is determined to be 44.1 kHz. Recently, however, the theory that a hypersonic region of 22 kHz or higher can be perceived to some extent has become rich, and the next-generation DVD-audio or the like tends to increase the sampling frequency to 96 kHz. (Analog LP records also reproduce more than 22 kHz.) Then, the amount of spectral code is quadrupled.
[0010]
On the other hand, 128 types of GM (General MIDI) standard are stored in the MIDI sound source, and thousands or more types of distortion waveforms are stored in the extended sound source, and 16 to 32 types of distortion waveforms can be synthesized in real time. Therefore, in view of the above points, the present invention obtains encoded data corresponding to a musical instrument part by extracting and encoding from a sound signal a distortion waveform close to a MIDI sound source prepared in advance. It is an object of the present invention to provide an acoustic signal encoding system that can perform the above-described process.
[0011]
[Means for Solving the Problems]
In order to solve the above-described problems, in the present invention, an acoustic signal encoding system defines a plurality of distortion waveform pattern storage means for storing a plurality of distortion waveform patterns and a plurality of predetermined analysis sections for a given acoustic signal. a section defining means for, with respect to the strain waveform pattern and the hand stage from the storage unit to enter the distorted waveform pattern corresponding to the acoustic signal as a basic distorted waveform pattern, one signal sections waveform of the analysis zones, While performing frequency analysis to decompose into a set of sine waves, divide the decomposed sine waves into a plurality of groups according to a predetermined rule, and obtain a distorted wave of each group by synthesizing the sine waves belonging to each group the fundamental distortion frequency and strain waveform variation with varying amplitude of the waveform patterns create multiple waveform distortion shape close to the strain wave of each group And signal analyzing means for selecting a shape variation as a distorted waveform variation of each group, the signal frequency and amplitude of the distortion waveform variation of each selected by the analyzing means, and start and end time information given of the analysis zones It is characterized by having a code forming means for creating acoustic encoded data by encoding with a code set.
[0012]
According to the present invention, a structure as described above, the basic distorted waveform pattern corresponding to the acoustic signal to enter, to create a plurality of strain waveform variations of varying the frequency and amplitude of the fundamental distortion waveform pattern, the Among the distortion waveform variations, the signal intercept waveform obtained from the acoustic signal is frequency-analyzed and the group closest to the distortion wave of each group is selected and encoded. It is possible to perform encoding for each component close to a musical instrument part.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
First, the basic concept in the acoustic signal encoding system according to the present invention will be described. An acoustic signal digitized by a technique such as PCM can be decomposed into a sine wave as shown in the following (Equation 1) by short-time Fourier transform, generalized harmonic analysis, and the like, which are well-known techniques.
[0014]
(Formula 1)
g (t) ≈Σ _i {A _i sin (2πf _i t) + B _i cos (2πf _i t)}
[0015]
Reproducing g (t) by [f _i , A _i , B _i ] in (Equation 1) is the principle of a conventional analysis / synthesis coding method. When the above (Formula 1) is further arranged, it is transformed as the following (Formula 2).
[0016]
(Formula 2)
g (t) ≈Σ _n α _n [Σ _j {P _nj sin (2πjf _n t) + Q _nj cos (2πjf _n t)}]
[0017]
In this (Formula 2), j is an integer indicating that the frequency is multiplied by an integer, and indicates a state in which sine waves having a frequency jf _n that is an integral multiple of a certain frequency f _n are grouped. When the above (Formula 2) is further arranged, it is transformed as the following (Formula 3).
[0018]
(Formula 3)
g (t) ≈Σ _n α _n u _n (2πf _n t)
[0019]
In this (Equation 3), u _n (2πf _n t) is a distorted wave having the fundamental frequency f _n , and if the one having the closest shape is selected from those prepared in the MIDI sound source as this distorted wave, the theory G (t) can be reproduced in the same manner as the above analysis / synthesis coding method. An advantage of the present invention, compared to the perform calculations Σ in i in (Equation 1), who performs Σ calculated by n in (Equation 3) is overwhelmingly small number of calculations, the parameters to be encoded [f _n , Α _n ] also decreases. The problem here is a method of determining the strain wave u _n (2πf _n t), it is difficult to determine the input audio signal g (t) directly from the strain wave u _n (2πf _n t). Therefore, the present invention adopts a method of selecting one corresponding to the input acoustic signal from the distortion waveform patterns stored in advance in the distortion waveform pattern storage means such as the MID sound source.
[0020]
Subsequently, a specific configuration of the present system will be described. FIG. 1 is a functional block diagram showing the configuration of an acoustic signal encoding system according to the present invention. In FIG. 1, an acoustic signal input unit 1 has a function of inputting a digital acoustic signal obtained by digitizing an acoustic signal that is an analog signal by a technique such as PCM. The section definition unit 2 has a function of dividing a time-series acoustic signal into predetermined sections in order to analyze the acoustic signal. The signal analysis unit 3 has a function of analyzing an acoustic signal in units of sections defined by the section definition unit 2. The distortion waveform pattern accumulation unit 4 accumulates a plurality of distortion waveform patterns, and is realized by, for example, a MIDI sound source having 128 types of distortion waveforms. The distortion waveform pattern search unit 5 searches the distortion waveform pattern accumulated in the distortion waveform pattern accumulation unit 4 for the one corresponding to the input acoustic signal and inputs it as a basic distortion waveform pattern. . The code forming unit 6 has a function of encoding the acoustic signal analyzed by the signal analyzing unit 3 into data such as MIDI.
[0021]
Next, the processing operation of the acoustic signal encoding system shown in FIG. 1 will be described. First, an acoustic signal digitized by PCM is input from the acoustic signal input unit 1. Here, for example, an acoustic signal represented by a waveform as shown in FIG. 2 is input. In the example of FIG. 2, this acoustic signal is shown with time t on the horizontal axis and amplitude (intensity) on the vertical axis.
[0022]
Subsequently, the section definition unit 2 sets a plurality of analysis sections on the time axis of the acoustic signal to be analyzed. In the example shown in FIG. 2, six times t1 to t6 are defined at equal intervals on the time axis t, and five analysis sections d1 to d5 having these times as a start point and an end point are set. In the example of FIG. 2, analysis sections having the same section length are set, but the section length may be changed for each analysis section. Alternatively, section settings may be made such that adjacent analysis sections partially overlap on the time axis. When the analysis section is set in this way, the signal analysis unit 3 analyzes the acoustic signal for each analysis section (hereinafter referred to as section signal).
[0023]
In the signal analysis unit 3, first, generalized harmonic analysis is performed for each analysis section, and the section signal is decomposed into a plurality of sine waves. Here, 1280 different sine waves are obtained by decomposition. The amplitude of each sine wave is determined based on the correlation between the interval signal and the sine function and cosine function having each frequency. Generalized harmonic analysis is a well-known frequency analysis method, and is described in the above-mentioned publications and specifications, so detailed description thereof is omitted.
[0024]
Subsequently, one having a large amplitude value is selected from the decomposed sine waves. The number of sine waves to be selected can be set in advance, or the threshold value of the amplitude value can be set and the number exceeding the threshold value. Here, the number of sine waves to be selected is the number of single notes that finally make up a chord. When a sine wave is selected, the frequency of each sine wave is set as a fundamental frequency, and sine waves having an integer multiple of the fundamental frequency are grouped. Next, sine waves belonging to the same group are synthesized to create a distorted wave. FIG. 3 shows an example of a decomposed sine wave and a combined distorted wave. In FIG. 3, six sine waves having different frequencies are shown on the left side. Of these, the third and fourth sine waves from the top have a frequency that is an integral multiple of twice or three times the top sine wave, respectively. For this reason, these three sine waves are combined as one group and become a distorted wave as shown in the upper right in FIG. Similarly, the fifth and sixth sine waves from the top have a frequency that is an integer multiple of three and five times that of the second sine wave from the top. , It becomes a distorted wave like the lower right. By performing these processes for each group, as many distortion waves as the number of groups are created. The method of grouping is not limited to the method of grouping with a sine wave having an integer multiple of a sine wave having a specific frequency, but other methods may be used. By grouping sine waves having an integer multiple frequency, it is possible to reduce the processing load as shown in the above (Equation 2) and (Equation 3). In addition, when grouping by the above method, a sine wave that should be a fundamental frequency may not be included in the group. This is because the amplitude of the sine wave that should be the fundamental frequency is small, so that it is not the first choice. In such a case, sine waves having a frequency that is an integral multiple of the frequency of an appropriate sine wave belonging to another group are grouped.
[0025]
On the other hand, a distortion waveform pattern corresponding to the acoustic signal input from the acoustic signal input unit 1 is searched by the distortion waveform pattern search unit 5 from the distortion waveform pattern storage unit 4 realized by the MIDI sound source, and the basic distortion waveform pattern is searched. Is entered as In this distortion waveform pattern search, an acoustic signal to be analyzed is output as a separate voice, the searcher listens to it, decides which distortion waveform pattern to use, and uses the information for identifying the distortion waveform pattern as a search key. This is done by inputting. For example, when the searcher listens to the sound signal by voice and feels that it is optimal to express the sound with a piano sound, a search is made for a distortion waveform pattern expressing the tone of the piano.
[0026]
The signal analysis unit 3 uses the distortion waveform pattern searched and input by the distortion waveform pattern search unit 5 as a basic distortion waveform pattern, and creates a distortion waveform variation in which the frequency and amplitude of the basic distortion waveform pattern are changed. The number of distortion waveform variations to be created can be changed by setting. However, if 128 frequencies and amplitudes are set, 128 × 128 distortion waveform variations are created.
[0027]
From these 128 × 128 distortion waveform variations, one having a waveform shape close to each distortion wave obtained by synthesizing the sine waves grouped as described above is selected one by one. As a result, a predetermined number of distortion waveform variations are selected.
[0028]
Next, the code forming unit 6 outputs the frequency and amplitude of each selected distortion waveform variation and information on the start time and end time of the analysis section in a predetermined code set. Here, the frequency f _n and the amplitude α _n to be encoded correspond to the above (Formula 3). As for the analysis section, for example, in the case of the analysis section d1 shown in FIG. 2, the start time is t1 and the end time is t2.
[0029]
The above processing is performed for all analysis sections of the acoustic signal, and a code set is output for each analysis section. At this time, the end time of the previous analysis interval is close to the start time of the subsequent analysis interval, and the frequency and amplitude of the code set in the previous analysis interval are changed to the frequency and amplitude of the code set in the subsequent analysis interval. If they are similar, the code sets in the two analysis intervals are integrated. Specifically, the code set end time in the previous analysis interval is changed to the code set end time in the subsequent analysis interval, and the code set of the later end time is deleted. This is performed so that when sounds in adjacent analysis sections are similar, they are regarded as the same sound and are combined into one sound at the time of encoding.
[0030]
The encoded data need not be particularly limited, but it is preferable to use the MIDI standard. When the MIDI standard is used, the frequency information of the code set is selected from 128 types of frequencies corresponding to the note numbers defined in the MIDI standard, and the amplitude information is a 128-step velocity defined in the MIDI standard. The start time and end time are described in SMF standard delta time, and the code set is described as a pair of note-on event and note-off event.
[0031]
The acoustic encoded data obtained in this way is decoded by an acoustic data decoding system via a storage means or a transmission means, and is emitted as an acoustic signal. FIG. 4 shows a functional block diagram of the acoustic data decoding system. In FIG. 4, the distortion waveform pattern accumulating unit 4 has the same function as that shown in FIG. The encoded data input unit 7 is for inputting acoustic encoded data. The distorted waveform variation reproducing unit 8 has a function of reproducing the distorted waveform variation by changing the basic distorted waveform pattern obtained from the distorted waveform pattern searching unit 9 based on the input encoded sound data. The distortion waveform pattern search unit 9 has a function of searching for a distortion waveform pattern from the distortion waveform pattern storage unit 4 in accordance with an instruction from the distortion waveform variation reproduction unit 8. The signal synthesis unit 10 has a function of reproducing the acoustic signal by synthesizing the distortion waveform variations reproduced for each code set.
[0032]
Next, the processing operation of the acoustic data decoding system shown in FIG. 4 will be described. When the acoustic encoded data obtained by the acoustic signal encoding system is input from the encoded data input unit 7, the distortion waveform variation reproducing unit 8 extracts code data in units of code sets from the acoustic encoded data. Performs waveform variation reproduction processing. As a distortion waveform variation reproduction process, first, information for identifying a basic distortion waveform pattern is extracted from a code set. Subsequently, using this identification information, the distortion waveform pattern search unit 9 extracts a corresponding distortion waveform pattern from the distortion waveform pattern storage unit 4. The distortion waveform variation reproduction unit 8 obtains the extracted distortion waveform pattern as a basic distortion waveform pattern, and performs deformation so that the basic distortion waveform pattern has the frequency and amplitude of the code set. Create This distortion waveform variation is created for each code set by the number of single notes constituting the chord.
[0033]
The processing by the distortion waveform variation reproduction unit 8 is performed on all the code sets, and when all the distortion waveform variations are reproduced, the signal synthesis unit 10 synthesizes the reproduced distortion waveform variation. This synthesis process is performed on the time axis based on the information on the start time and end time of each code set. Since the process of synthesizing the distorted wave corresponding to the code set from the start time and the end time of such a code set is a method conventionally performed when reproducing MIDI code data using a MIDI sound source, Detailed description is omitted. The acoustic signal reproduced by being synthesized by the signal synthesizing unit 10 can be output in the PCM format, for example.
[0034]
The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments, and various modifications can be made. For example, in the above embodiment, the case where the encoded sound data is configured by one track has been described, but the encoded sound data may be configured by a plurality of tracks. Since one track can correspond to one instrument part, by preparing a plurality of tracks, it is possible to reproduce music played by a plurality of instruments. When creating encoded audio data having a plurality of tracks, a plurality of distortion waveform patterns may be retrieved from the distortion waveform pattern storage unit 4 shown in FIG. good. As a result, the signal analysis unit 3 recognizes a plurality of basic distortion waveform patterns and creates a track for each. A code set to be assigned to each track is created for each basic distortion waveform pattern. In the MIDI standard, a track corresponds to a channel, and information for identifying a basic distortion pattern corresponds to a program number indicating a musical instrument tone color and a bank number.
[0035]
In the above embodiment, a distortion waveform pattern prepared in advance in the distortion waveform pattern storage unit 4 is searched and input as a basic distortion waveform pattern. However, it seems to be most suitable for the sound input by the user. It is also possible to add a distorted waveform pattern to be added to the distorted waveform pattern storage unit 4 as needed. In that case, if the basic distortion waveform pattern undefined area of the bank number defined by the MIDI standard is defined as a basic distortion waveform pattern added by the user, encoding and decoding can be performed without departing from the MIDI standard. Such waveform addition can be realized using a MIDI standard sound source having a sampler function.
[0036]
【The invention's effect】
As described above, according to the present invention, the distortion waveform pattern storage means for storing a plurality of distortion waveform patterns, and the section definition means for defining a plurality of predetermined analysis sections for a given acoustic signal, , and hand stage to enter the distorted waveform pattern corresponding to the acoustic signal from the distorted waveform pattern storage means as a basic distorted waveform pattern, for one of the signal sections waveform of the analysis zones, sine performs frequency analysis While disassembling into a set of waves, dividing the decomposed sine waves into a plurality of groups according to a predetermined rule, and combining the sine waves belonging to each group to obtain a distorted wave of each group , the basic distortion waveform pattern frequency and distortion waveform variations with varying amplitudes create multiple, each group distortion waveform variation of the waveform shape close to the strain wave of each group And signal analyzing means for selecting as a distorted waveform variations flops, numerals frequency and amplitude of the distortion waveform variation of each selected by the signal analyzing means, and the information of the start and end times of the analysis zones in a predetermined code set Therefore, there is an effect that it is possible to perform coding for each component close to a musical instrument part prepared as a sound source in advance.
[Brief description of the drawings]
FIG. 1 is a functional block diagram showing a configuration of an acoustic signal encoding system according to the present invention.
FIG. 2 is a diagram for explaining analysis section definition by the section definition section 2;
FIG. 3 is a diagram showing a decomposed sine wave and a distorted wave obtained by synthesizing the sine wave.
FIG. 4 is a functional block diagram showing a configuration of an acoustic data decoding system.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Acoustic signal input part 2 ... Section definition part 3 ... Signal analysis part 4 ... Distortion waveform pattern storage part 5 ... Distortion waveform pattern search part 6 ... Code formation part 7 ... Encoded data input unit 8 ... distortion waveform variation reproduction unit 9 ... distortion waveform pattern search unit 10 ... signal synthesis unit

Claims (8)

Distortion waveform pattern storage means for storing a plurality of distortion waveform patterns, section definition means for defining a plurality of predetermined analysis sections for a given acoustic signal, and from the distortion waveform pattern storage means to the acoustic signal hand stage to enter the corresponding distorted waveform pattern as basic distorted waveform pattern,
One signal intercept waveform in the analysis interval is subjected to frequency analysis to be decomposed into a set of sine waves, the decomposed sine waves are divided into a plurality of groups according to a predetermined rule, and the sine waves belonging to each group The distortion wave of each group is obtained by synthesizing, while creating multiple distortion waveform variations in which the frequency and amplitude of the basic distortion waveform pattern are changed, and the distortion waveform variation of the waveform shape close to the distortion wave of each group is created. Signal analysis means to select as a distortion waveform variation of each group ;
Code formation for generating acoustic encoded data by encoding the frequency and amplitude of each distortion waveform variation selected by the signal analysis means, and information on the start time and end time of the analysis section with a predetermined code set An acoustic signal encoding system comprising: means. With the previous SL basic distorted waveform pattern is a multiple-input, the sound encoded data is composed of a plurality of tracks, said isolated in any of the tracks relative to the basic distorted waveform patterns each code set is used, each The acoustic signal encoding system according to claim 1, wherein identification information of a basic distortion waveform pattern selected for each track is added. The code forming means is configured to perform the operation in the case where the end time of the previous code set is close to the start time of the subsequent code set and the frequency and amplitude of the previous code set are similar to the frequency and amplitude of the subsequent code set. The acoustic signal encoding system according to claim 1 or 2 , wherein the end time of the code set is changed to the end time of the subsequent code set and the subsequent code set is deleted. The grouping in the signal analyzing means is performed such that the frequency of the sine wave belonging to each group is an integer multiple of the frequency of one or more specific sine waves in the group. The acoustic signal encoding system according to any one of claims 1 to 3 . The grouping in the signal analyzing means is performed such that the frequency of the sine wave belonging to each group is an integer multiple of the frequency of one or more specific sine waves belonging to another group. The acoustic signal encoding system according to claim 4 . The frequency information of the code set is selected from 128 types of frequencies corresponding to the note numbers defined in the MIDI standard, the amplitude information is described in 128 levels of velocity defined in the MIDI standard, and the start time and end time Are described in SMF standard delta time, the code set is described as a pair of MIDI standard note-on and note-off events, and the distortion waveform pattern is realized by a MIDI sound source. The acoustic signal encoding system according to any one of claims 1 to 5 .   The frequency information of the code set is selected from 128 types of frequencies corresponding to the note numbers defined in the MIDI standard, the amplitude information is described in 128 levels of velocity defined in the MIDI standard, and the start time and end time Is described in SMF standard delta time, the code set is described as a pair of MIDI standard note-on event and note-off event, the distortion waveform pattern is realized by a MIDI sound source, and the track is 3. The acoustic signal encoding system according to claim 2, wherein the identification signal is a MIDI standard channel, and identification information of a basic distortion waveform pattern is described by a program number and a bank number indicating a musical instrument tone color according to the MIDI standard. An acoustic data decoding system for reproducing the acoustic signal by using acoustic encoded data encoded by a predetermined method in advance for a given acoustic signal,
A distorted waveform pattern storing means for storing a plurality of strain waveform pattern, the encoded data input means for inputting the sound encoded data, a distorted waveform pattern corresponding to the input audio coded data, the sound coded data The distortion waveform pattern search means for searching from the distortion waveform pattern storage means based on the identification information of the code set of the code, and the searched basics based on the frequency and amplitude information of the code set of the acoustic encoded data Distortion waveform variation reproduction means for creating a distortion waveform variation by changing the frequency and amplitude of the distortion waveform pattern, and each of the code sets with respect to a plurality of distortion waveform variations obtained by creating each of the code sets wherein by performing synthesis on the time axis based on the information of the start and end times Sound data decoding system characterized in that it comprises a signal combining means for reproducing a sound signal.

Images