JP3617603B2 - Audio information encoding method and generation method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a speech information of which reproduction time is extended or contracted and emphasized or attenuated without altering frequency components of original speech information itself in an arbitrary part as a speech for hearing exercise of Japanese English learners. SOLUTION: According to this recording method, a 1st speech information string sampled in a 1st period is divided into plural frequency components, and concerning each frequency component, a sinusoidal wave data, which has been changed in the amplitude and the number of waveforms of a predetermined part to an amplitude information sequence sequentially extracted in a 2nd period, is produced, and a 2nd speech information sequence synthesized by adding a sinusoidal wave data corresponding to these individual frequency components to the former sinusoidal data is recorded in a predetermined recording medium 15. In such a manner, a speech information string which is extended or contracted and emphasized or attenuated without changing the frequency in an arbitrary part is recorded in the obtained recording medium 15.

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an audio information encoding method for generating a new audio information sequence based on a predetermined audio information sequence, and a method for generating the new audio information.
[0002]
[Prior art]
Conventionally, various teaching materials in which audio information is recorded on a recording medium such as a cassette tape have been provided for self-study of languages such as English conversation, practice of poetry, self-study of law, and other purposes. Here, taking a self-study material for English conversation as an example, a conventional main recording medium is, for example, a cassette tape (or record) in which a series of English utterances (voice information) is recorded. This tape material and text were used in combination. There are various levels of such teaching materials from beginner to advanced.
[0003]
In Japanese Patent No. 2581700, a first area in which a speech information sequence (sound generated by natural speed) suitable for advanced learning divided into a plurality of sections is recorded, and these sections correspond to each section. A second area in which a speech information string (sound that is clearly generated and differently derived in the same linguistic sense) suitable for beginner learning is composed of equivalent sections, and the advanced user It has at least a third area in which information indicating the relationship between the corresponding sections of each of the speech information strings for learning and beginner learning is recorded at the recording position in the recording medium of each section of these sound information strings. An information recording medium such as a CD-ROM and a reproducing method including switching reproduction between corresponding sections of the information recording medium having such a structure have been proposed.
[0004]
[Problems to be solved by the invention]
As described above, in the information recording medium of Japanese Patent No. 2581700, the sound generated by a native speaker is recorded in the first area on the medium, and the second area is delayed in the same sense in terms of language. A sound information sequence composed of pronunciation is recorded. Therefore, if the reproduced sound cannot be heard while the audio information sequence recorded in the first area is being reproduced, the audio information sequence having the same content recorded in the second area (during the reproduction of the first audio information sequence) And the correspondence between the section of the second audio information string and the section to be reproduced is recorded in the third area), and the learner can understand the meaning of the voice that could not be heard.
[0005]
However, although the English learner can understand the information recorded in the first area by listening to the information recorded in the second area as described above, it is still recorded in the first area. Information, especially sounds that cannot be heard, cannot be heard just by repeatedly listening. It is well known that Japanese learners of English are not good at listening to phonemes, especially consonants, that are not in Japanese, and hindering conversations with native speakers.
[0006]
The present invention relates to a technique for improving hearing ability of an original voice by letting a voice edited in advance so that a hard-to-hear part is easy to hear for a learner. It is an object of the present invention to provide a method for encoding audio information in which the amplitude and reproduction time of the frequency component are selectively edited without changing the frequency component of the audio information itself, and a method for generating the same.
[0007]
[Means for Solving the Problems]
The present invention emphasizes or attenuates a desired portion of the audio information sequence without changing the frequency component of the captured audio information sequence as hearing practice audio, and partially extends or shortens the reproduction time. The present invention relates to a technique for newly generating a voice information sequence. In the present invention, since the sound quality of the reproduced audio information is not changed, desired editing is performed on each frequency component of the audio information, not on the sampled audio information, and these edited frequency components are changed. A new audio information sequence is obtained by synthesis. With this configuration, it is possible to provide audio information for hearing practice in which portions difficult to hear for Japanese English learners are selectively emphasized and / or extended. On the other hand, when the advanced user desires further improvement of the hearing ability, it is possible to provide audio information in which the audio is selectively attenuated or the reproduction time is shortened.
[0008]
Specifically, in the audio information encoding method according to the present invention, a first audio information sequence sampled in a first period (for example, an audio clock 44.1 KHz of a music CD) is converted into a plurality of frequency components (hereinafter referred to as channels). The amplitude information is obtained for each channel in a second period (e.g., corresponding to the number of data necessary to form one waveform). This amplitude information is given by the amplitude change amount of the waveform corresponding to, for example, 100 data in the first audio information string. If one waveform is not formed in 100 data, the amplitude information is increased to the number of data capable of one waveform ( The second period is extracted). The second cycle may be a regular cycle.
[0009]
Further, a plurality of corrections edited so as to selectively change the amplitude information with respect to the amplitude information sequence of each channel obtained in this way (sequence of amplitude information extracted in the second period for each channel). An amplitude information sequence is generated. The plurality of modified amplitude information sequences are obtained for each channel corresponding to each frequency component. Then, between the modified amplitude information sequences corresponding to each channel, each information component group consisting of amplitude information extracted at the same timing corresponding to each other, and a first cycle prepared for each of these information component groups, V data including control information for instructing to extend or shorten the audio reproduction time with reference to is generated.
[0010]
Subsequently, the audio information generating method according to the present invention has an amplitude (value after correction) given by the V data from the V data generated as the data of the second period and a data interval of the first period. Sine wave data corresponding to each channel and having a wave number corresponding to the reproduction time indicated by the control information. By sequentially adding the sine wave data generated for each channel in this way, audio data (second audio information string) of the first period is generated. The generated audio data is recorded on a predetermined recording medium.
[0011]
In the audio information encoding method according to the present invention, each amplitude information in the amplitude information sequence of each channel extracted in the second period is edited so as to be selectively emphasized or attenuated in an arbitrary portion. Is done. That is, for the amplitude information string of each channel, a corrected amplitude information string is generated by selectively resetting the amplitude value given from the amplitude information of a predetermined portion corresponding to each channel. The In addition, in order to avoid an unnatural amplitude change of the reproduced sound, each amplitude of the generated sine wave data for each channel is based on a value obtained by linear interpolation between adjacent amplitude information in the corrected amplitude information sequence. It is determined.
[0012]
As described above, the audio information encoding method according to the present invention is configured to change the amplitude of an arbitrary portion of the amplitude information sequence generated for each channel, and the reproduction time Since control information for instructing expansion / reduction is prepared for each information component group in which the amplitude information of each channel extracted in the second period is collected, the reproduction is performed in an arbitrary part without changing the frequency component. The audio can be selectively emphasized and attenuated, and the playback time can be partially extended and shortened.
[0013]
This is mainly because even if Japanese people simply listen to natural speed English by listening slowly, it is not enough to simply extend or shorten the audio playback time for each frequency component. This is because, depending on the type of generated sound, the temporal change in the spectrum of the consonant part may mean another sound as a linguistic sound. For example, BA (PA) and PA (PA) pronunciations have a fast spectrum change, the latter is slow, and the spectrum itself has almost the same shape. Therefore, if the time is extended including the consonant part of the pronunciation of BA, it will be heard as PA. In order to prevent this, if the extension level of the consonant part is limited to the limit that can be heard as BA, and only the vowel part is extended or shortened to the desired voice reproduction time, it can be heard as BA. . On the other hand, the vowel part can be set to the desired length (desired playback time) since it can be heard as it is, no matter how much it is expanded or shortened. On the other hand, it is also necessary to selectively emphasize only the small consonant sounds that are too weak to be heard by the Japanese and to emphasize them twice or three times. The emphasis including the vowel part is too large to be effective. It must be emphasized selectively. For the reasons described above, the amplitude information sequence for each channel is also edited at the same timing in the modified amplitude information sequence for each channel, with the portion of the amplitude information sequence that is particularly difficult to hear for the beginner being selectively emphasized. The control information for instructing the extension of the reproduction time is sequentially recorded together with the V data composed of the amplitude information. On the other hand, in the case of an advanced user, even if the audio information string is selectively edited so that the reproduced audio is attenuated at a desired portion or the reproduction time is shortened in consideration of the characteristics of each uttered sound described above. Good.
[0014]
Furthermore, in the audio information encoding method according to the present invention, when male audio is recorded on a predetermined recording medium, the frequency spectrum of the output audio remains unchanged when the audio is reproduced while extending the audio reproduction time. However, there is a possibility of causing the illusion of shifting to a lower sound. On the other hand, when the playback is performed while shortening the voice playback time, there is a possibility of causing an illusion that the sound is shifted to a lower level. Therefore, it is preferable that the control information includes frequency shift instruction information for shifting the entire frequency component in the high pitch direction or the low pitch direction for about a semitone or one tone so as to be reproduced.
[0015]
In addition, the present invention can be expected to have a dramatic learning effect when combined with the technique disclosed in the above-mentioned Japanese Patent No. 2581700. In other words, corresponding to variable-length sections obtained by dividing the utterance sound of the native speaker by utterance milestones, by separately preparing audio information in which the audio of an arbitrary part is expanded and / or emphasized, the audio that could not be heard is repeated. By listening to the sound that can be reproduced and heard and the emphasized / extended portion of the sound that is difficult to hear can be expected to improve the hearing ability of the original sound. In addition, for advanced users, in order to improve the learning effect more positively, by separately preparing voice information in which the voice of an arbitrary part is shortened and / or attenuated together with the voice of the native speaker divided into sections, It is possible to play back with a shortened playback time, or to make it difficult to hear the consonant part (reduce the amplitude), and learning combined with the uttered sound of the native speaker becomes possible.
[0016]
Specifically, the first audio information sequence is an audio information sequence corresponding to one or more sentences composed of a word sequence to be reproduced and output by a predetermined audio reproducing means, and is divided at each pronunciation node. Each piece of information is recorded on a recording medium in a state of being divided into variable length sections. Thereby, the second audio information sequence is recorded on a predetermined recording medium for each section divided corresponding to the section of the first audio information sequence, and further, the first audio information sequence and the second audio information sequence are recorded on the recording medium. In order to switch and reproduce the second audio information string by a predetermined audio reproducing means, recording position identification information indicating each section that can be switched is indicated by the recording position of each section on the predetermined recording medium. In this manner, by recording in advance the correspondence relationship between each segment of the first audio information sequence and each segment of the second audio information sequence, one or more desired segments are repeatedly reproduced. In addition, it is possible to reproduce the natural speed reproduction sound and the same generated sound prepared according to the level of each learner while switching in real time.
[0017]
Therefore, the audio information recording medium in which predetermined audio information (not the waveform data but the amplitude information sequence in which each frequency component is corrected) is recorded by the encoding method according to the present invention is obtained.
[0018]
As such a sound information recording medium, for example, a disk-shaped recording medium such as a CD-ROM, MD, or MO, or a tape-shaped recording medium such as a DAT can be applied. The first audio information sequence sampled in the first period is divided into a plurality of frequency components, and for each of these frequency components, the amplitude of a predetermined portion and a predetermined value with respect to the amplitude information sequence sequentially extracted in the second period At least a second audio information sequence synthesized by generating sine wave data having a changed number of waveforms and adding the sine wave data corresponding to each frequency component is recorded. That is, the second audio information sequence recorded in the audio recording medium is a portion corresponding to each other between the frequency components for each frequency component constituting the first audio information sequence sampled at a predetermined cycle, This is the second audio information sequence in which the amplitude and the reproduction time are selectively edited by changing at least the amplitude or the number of waveforms.
[0019]
Further, in the audio information recording medium, the first information sequence, which is an audio information sequence corresponding to one or more sentences composed of a word sequence to be reproduced and output by a predetermined audio reproduction means, is a pronunciation node. By recording each divided information in a state of being divided into variable-length sections, it is possible to combine with the technique disclosed in the above-mentioned Japanese Patent No. 2581700.
[0020]
In the audio information recording medium configured as described above, the second audio information sequence is recorded for each of the sections divided corresponding to the sections of the first audio information sequence, together with the first audio information sequence. Recording position identification information indicating each switchable section as a recording position of each section on the predetermined recording medium so that the first sound information string and the second sound information string are switched and reproduced by a predetermined sound reproducing means. Therefore, by preparing such a sound information recording medium, the sound information in the corresponding section of the other sound information sequence can be reproduced even when one sound information sequence is being reproduced. Real-time switching playback is possible for the columns.
[0021]
In the above-described embodiment, the recording software (a program capable of executing the above-described recording method on a personal computer or the like, or a recording medium on which the program is recorded), a dedicated recording device, a use manual, or a combination thereof In addition to sales by such a voice information recording medium, the voice information recording medium, voice information playback software (including a program that can be executed by a personal computer, or a recording medium on which the program is recorded), dedicated playback The sale by an apparatus, a use manual, or these combination is considered.
[0022]
DETAILED DESCRIPTION OF THE INVENTION
An embodiment of the present invention will be described below with reference to FIGS. In addition, the same code | symbol is attached | subjected to the same part in a figure, and the overlapping description is abbreviate | omitted.
[0023]
The present invention is a technique that makes it possible to provide voice information that is selectively emphasized or attenuated in advance, or that has been extended or shortened in reproduction time, for example, in hearing practice of an English learner. Therefore, the learner who has listened to the voice information edited in advance can be expected to improve the hearing ability with respect to the original voice.
[0024]
FIG. 1 is a conceptual diagram for schematically explaining the generation and recording operations of audio information according to the present invention. First, the natural speed sound (first sound information) of the native speaker sampled at, for example, the audio clock 44.1 KHz (first period) of the music CD is taken into the PC 1 main body by the microphone 11 or the like and once recorded on the hard disk or the like. Is done. Then, the captured audio information is filtered to divide into each channel (frequency component) divided as shown in the table of FIG. The frequency range of the audio information to be captured is 75 Hz to 10,000 Hz, and the sampling frequency is 44.1 kHz (22.68 μs) in accordance with the audio clock of the music CD. The number of channels to be divided is 85 (7 octaves + 1 sound), and the center frequency (center f) of each channel # 1 to # 85 is set to be a semitone string of equal temperament (12 equal temperament per octave). (77.78 Hz (D #) to 9,960 Hz (D #)).
[0025]
As described above, the data divided into the respective channels # 1 to # 85 has amplitude information of 2.268 ms (corresponding to 100 data of 44.1 kHz sampling, provided that one waveform cannot be formed with 100 data). To increase the number of data). Therefore, in this embodiment, the sampling rate (second period) of amplitude information in each channel # 1 to # 85 is 441 samples / s (2.268 ms). Note that this sampling rate may be a regular cycle. For example, after sampling 100 data, and then processing 120 data, the processing is repeated alternately at these different rates. May be.
[0026]
Further, the control system 10 of the PC 1 performs various editing (display 12 and input device 13 such as a keyboard and a mouse) on the amplitude information of each channel # 1 to # 85 sampled every 2.268 ms. To generate a modified amplitude information group every 2.268 ms. Then, the modified amplitude information (elements constituting the modified amplitude information group) of each channel # 1 to # 85 is expressed by 1 byte (8 bits), and further 2 bytes of control information is added to 87 bytes ( 85 channels × 1 byte + 2 bytes) V data 19 is generated.
[0027]
The corrected amplitude information is edited so that each amplitude information in the amplitude information string (amplitude information sampled at 2.268 ms) of each channel # 1 to # 85 is selectively emphasized or attenuated at an arbitrary portion. This is the information obtained. That is, with respect to the amplitude information string of each channel # 1 to # 85, the corrected amplitude is selectively set to be larger or smaller by selectively resetting the amplitude value given from the amplitude information of a predetermined portion corresponding to each other between the channels. An information sequence is generated. Further, the control information includes expansion instruction information (1 byte) instructed to extend or shorten the time to reproduce the frequency components of the channels # 1 to # 85, which are instructed by the editing operation, and each channel # 1. It is composed of frequency shift instruction information (1 byte) for instructing whether or not the frequency component corresponding to .about. # 85 is to be reproduced by shifting the entire half tone or one sound in the low or high pitch direction.
[0028]
The decompression instruction information is expressed by the number of reproduction clocks indicating how many ms one data is reproduced. For example, if this decompression instruction information is expressed by a half of the number of clocks to be reproduced, 50 is the same as the original reproduction time, and if this information is set to 100, it is reproduced at 200 clocks as a 44.1 kHz clock. Thus, it is possible to extend the reproduction time by a factor of 2 (since this information is expressed in 1 byte, the reproduction time can be expanded up to 256 ÷ 50 = 5.12 times). On the contrary, when this information is set to 25, reproduction is performed with 50 clocks as a 44.1 kHz clock, and the reproduction time can be reduced to ½. In the frequency shift instruction information, ON “1” is set when all frequency components are shifted in the bass direction or treble direction, and OFF “0” is set when no shift is necessary.
[0029]
As described above, a new audio information sequence is generated based on the V data 19 generated by the control system 10 performing desired editing on the amplitude information of each channel # 1 to # 85 sampled at 2.268 ms. Is done.
[0030]
In order to generate a new audio information sequence from the generated V data 19, sine wave generation circuits 16-1 to 16-85 that generate sine waves having wavelengths corresponding to the channels # 1 to # 85. The external device 16 having the above is required. In each of the generation circuits 16-1 to 16-85, a ROM in which basic data of a sine wave having a frequency corresponding to each of the channels # 1 to # 85 is recorded, and a RAM # in which the generated sine wave data is temporarily recorded. 1 to # 85. Each of these circuits is shaped based on the corrected amplitude information of the V data 19 sent from the control system 10, and has the number of waveforms indicated by the expansion instruction information of the control information. The sine wave data is written into each RAM # 1 to # 85. Note that the data interval constituting the sine wave data is a data interval of 22.68 μs with a sampling frequency of 44.1 kHz.
[0031]
The sine wave data written in the RAMs # 1 to # 85 in each of the generating circuits 16-1 to 16-85 are sequentially read out at the timing of 44.1 kHz and added to each other to add audio data ( Audio information sequence) is generated. This audio data is sent to the control system 10 and output from the control system 10 to the input / output device 14 such as a CD-ROM writing device via the I / O. The input / output device 14 records the 44.1 kHz audio data sent from the control system 10 on a predetermined audio information recording medium 15 such as a CD-ROM.
[0032]
In the generation of the sine wave data performed by each of the generation circuits 16-1 to 16-85, in order to avoid an unnatural amplitude change of the reproduced sound, each amplitude of the sine wave data for each channel # 1 to # 85 is It is determined by a value obtained by linear interpolation between adjacent amplitude information in the corrected amplitude information sequence. Further, the audio data generated by the external device 16 may be output as audio from the speaker 18 via the DAC 17 and AMP as it is. Furthermore, as the audio information recording medium 15, for example, a disk-shaped recording medium such as CD-ROM, MD, or MO, or a tape-shaped recording medium such as DAT can be applied.
[0033]
On the other hand, the present invention mainly relates to a technology that allows Japanese people to simply reproduce and listen to natural speed English, but it has not been possible to simply and uniformly extend or shorten the audio playback time for each frequency component. Insufficient. That is, FIG. 3 is a diagram showing the basic shape of the speech spectrum, but depending on the type of generated sound, the temporal change in the spectrum of the consonant part may mean another sound as a linguistic sound. . For example, BA (PA) and PA (PA) pronunciations have a fast spectrum change, the latter is slow, and the spectrum itself has almost the same shape. Therefore, if the time is extended including the consonant part of the pronunciation of BA, it will be heard as PA. In order to prevent this, if the extension level of the consonant part is limited to the limit that can be heard as BA, and only the vowel part is extended or shortened to the desired voice reproduction time, it can be heard as BA. . On the other hand, the vowel part can be set to the desired length (desired playback time) because it can be heard as it is, no matter how much it is expanded or shortened. On the other hand, it is also necessary to selectively emphasize only small consonant sounds that are too weak for the Japanese to listen to, with a double or triple enhancement (with an increased amplitude). The emphasis including the vowel part is too large to be effective. It must be emphasized selectively. For the reasons described above, the amplitude information sequence for each channel is also edited at the same timing in the modified amplitude information sequence for each channel, with the portion of the amplitude information sequence that is particularly difficult to hear for the beginner being selectively emphasized. Control information for instructing the extension of the reproduction time is generated as V data together with the information component group constituted by the amplitude information. On the other hand, in the case of an advanced user, even if the audio information string is selectively edited so that the reproduction sound is attenuated at a desired portion or the reproduction time is shortened in consideration of the characteristics of each uttered sound described above. Good.
[0034]
Furthermore, when a male voice is recorded on a predetermined recording medium by the above-described recording method, the frequency spectrum of the output voice remains unchanged if the voice playback time is extended and / or the desired portion of the voice is emphasized. There is a possibility that the illusion of shifting to a lower sound will be caused. On the other hand, when playback is performed while reducing the audio playback time and / or attenuating the audio of a desired portion, there is a possibility that an illusion that the sound has been shifted to a lower sound sensuously. Therefore, the control information includes frequency shift instruction information for making the entire frequency component reproducible by shifting the entire frequency component in the bass direction or the treble direction for about one semitone or one tone.
[0035]
Next, as disclosed in the above-mentioned Japanese Patent No. 2581700, the present invention is suitable for a technique for reproducing a recording medium in which the sound of a native speaker is recorded. Hereinafter, a configuration in which the present invention is applied to the technology will be described.
[0036]
This invention can be expected to have a dramatic learning effect when combined with the technique disclosed in the above-mentioned Japanese Patent No. 2581700. In other words, corresponding to variable-length sections obtained by dividing the utterances of native speakers by utterance milestones, learning is performed by separately preparing voice information in which arbitrary portions are selectively expanded or reduced, or emphasized or attenuated. The user can repeatedly hear and listen to the sound that he could not hear, and to improve the listening ability, it is also possible to hear the part of the reproduced sound that is difficult to hear as the sound is expanded, shortened, emphasized or attenuated. Become.
[0037]
FIG. 4 is a diagram for conceptually explaining various types of information including an audio information string to be recorded on a predetermined audio recording medium.
[0038]
First, the first audio information sequence (the audio information sequence sampled at 44.1 kHz) recorded on the audio information recording medium 15 has a length such as a conversation of a performer in a movie, a conversation in a daily living environment, and the like. It is composed of a plurality of sentences (sentences) with different voices, and between each sentence (voice information of each talker), a situation where no voice is played, a situation where only noise is played, and music (BGM) only This is a series of audio information sequences in which there may be a silent period that occurs at random, such as a situation where is being played. Therefore, the first audio information sequence is an audio information sequence corresponding to one or more sentences composed of a plurality of word sequences to be reproduced and output by a predetermined audio reproducing means, and the audio information recording medium 15 In the first area, as shown in FIG. 4, each piece of audio information divided at each pronunciation node is recorded in a state of being divided into variable-length sections (hereinafter referred to as segments).
[0039]
In general, in a native speaker's English conversation, one sentence is uttered in about 3 seconds. Therefore, by setting a pronunciation node that determines a segment constituting a speech information string to be recorded between each sentence, FIG. As shown in (a), (b), or (d), it is appropriate to configure the variable length segments 621, 622, and 799 constituting the audio information sequence, respectively. Note that, as shown in FIG. 4C, an extremely short sentence is included in the sentence in conversation, but this sentence 701 also constitutes one segment. On the other hand, as shown in FIG. 4 (e), in the case of an extremely long sentence, the sentence before the conjunction or relative is a pronunciation milestone, so the sentence as shown in FIG. 4 (e). Therefore, it is reasonable to configure the segment by two continuous segments 801 and 802. Therefore, the segment of the speech information sequence to be recorded means a recording unit of speech information divided based on a speech break (breathing position) or some linguistic (grammatical) break.
[0040]
In the present invention, first, an arbitrary portion is selectively edited (change of amplitude of each frequency component, change of reproduction time) for each segment obtained by dividing the first information string as described above. A second audio information sequence is generated. Specifically, as shown in FIG. 5, the second audio information sequence is generated by the PC 1 main body that edits each frequency component and the external device 16 that generates the edited audio data (second audio information sequence). The information is recorded on a predetermined audio information recording medium 15 by the configured device.
[0041]
In particular, as shown in FIG. 5, the external device 16 is a slave having a master board 165 for generating audio data and sine wave generation circuits 16-1 to 16-85 provided corresponding to the respective channels. The board 166 is configured. The master board 165 includes a timing controller 171 and a FIFO 172 to supply the V data from the PC 1 to each of the generation circuits 16-1 to 16-85 according to the control signal, and from each of the generation circuits 16-1 to 16-85. An adder 173 that sequentially adds the sine wave data (16 bits) sent to generate audio data (16 bits), and a RAM 174 as a buffer that temporarily stores the generated audio data transmitted to the PC 1. Prepare. Note that the master board 165 shown in FIG. 5 uses the first voice information sequence and the newly edited second voice information sequence as sounds in response to an instruction from the PC 1 and outputs them to the speaker again and again. The DAC 175 and the AMP 176 are provided so that the generated audio data can be directly reproduced and output by the speaker 177 (the structure for audio reproduction is provided on the PC 1 side as shown in FIG. 1). May be good). On the other hand, the slave board 166 includes sine wave generation circuits 16-1 to 16-85 that respectively generate sine waves of a predetermined frequency corresponding to each channel, and these generation circuits 16-1 to 16-85 are provided. , A ROM in which data for generating a sine wave is recorded, and RAMs # 1 to # 85 as buffers for storing the sine wave data once generated.
[0042]
Note that the master board 165 and the slave board 166 are connected by 30 signal buses and a total of 32 buses of GND and Vcc. In FIG. A bus group related to V data for supplying V data to 16-85, and a bus group indicated by 168 is related to audio data for sending sine wave data for generating audio data from each generating circuit to the master board 165. A group of buses.
[0043]
Next, an embodiment of the audio information generation method according to the present invention applied to the technique disclosed in Japanese Patent No. 2581700 will be described with reference to FIG. 5 and the flowcharts of FIGS. explain.
[0044]
First, generation of V data is performed on the PC 1 side. That is, in PC1, a series of audio information sequences (first audio information sequence) samples 44.1 kHz (16 bits / data), and sampling data corresponding to the first audio information sequence is temporarily stored in the hard disk (step ST1). 4), as shown in FIG. 4, it is decomposed into a plurality of segments (step ST2).
[0045]
Subsequently, the waveform information of the band width (75.57 kHz to 80.06 kHz) of the first channel # 1 is first developed in the memory for one of the divided segments by the digital band bass filter program. . Also at this time, the data interval corresponding to the 44.1 kHz rate is developed. Then, average amplitude information (8 bits) is extracted every 100 data (step ST3). As described above, when one waveform cannot be formed with 100 data with respect to the frequency component of the first channel # 1, the number of data that can be formed with one waveform is increased to obtain the distribution information. The same operation is repeated by shifting by 100 data until the sampling data of the target segment is completed. With this operation, the amplitude information string (441 pieces of amplitude information per second) with a data interval of 2.268 ms is obtained for the first channel # 1, which is the target channel. When the amplitude information extraction operation of the first channel # 1 as the target channel is completed (step ST5), the frequency of the second channel # 2 is subsequently divided by the digital band bus filter, and the operations of the above-described steps ST3 to ST5 are performed. Is repeated and the target channel is changed (step ST7), and the amplitude information string of the target segment is generated for the first channel # 1 to the 85th channel # 85.
[0046]
The above operation is performed until all segments constituting the first audio information sequence sampled in step ST1 are completed while changing the target segment (step ST10) (step ST9).
[0047]
Next, on the PC1 side, the following editing is performed on the amplitude information string for 85 channels for each segment obtained by executing the above steps ST1 to ST9, and V data is generated. (Step ST11).
[0048]
First, the 85-channel amplitude information sequence group generated for each divided segment is called from the storage destination hard disk, and the amplitude waveform is sequentially displayed on the monitor 12.
[0049]
In actual editing work, a desired portion of the displayed amplitude waveform is designated and a reproduction time is designated (clock 50 is a reference). Further, if necessary, the part to be changed is specified to change the amplitude (set with the magnification based on the displayed amplitude), or to specify a frequency shift instruction in the bass direction or the treble direction. For example, the consonant part in the segment doubles the amplitude and the reproduction time 1.5 times, while the vowel part keeps the amplitude as it is and the reproduction time only 2.5 times. The amplitude information sequence obtained for the portion is edited, and a modified amplitude information sequence in which each amplitude information is newly corrected is generated.
[0050]
Then, among the obtained corrected amplitude information strings for 85 channels, the above-described reproduction time change is instructed for each information component group in which information components of the same timing corresponding to each other between the corrected amplitude information strings are collected. By adding control information including information to be transmitted and information for instructing frequency shift, V data having a data interval of 2.268 ms can be obtained.
[0051]
Next, the V data (87 bytes / data) prepared on the PC 1 side as described above is sent to the master board 165 of the external device 16, and further from the master board 165 via the data bus, each data on the slave board 166. It is sent to the sine wave generation circuits 16-1 to 16-85. The slave board 166 is actually composed of 11 boards on which 8 circuits are mounted (only 5 of the 8 circuits are used for the 11th board), and each circuit corresponds. Channel sine wave data is generated (step ST12). Each circuit has the same configuration except that the ROM storing the sine wave waveform data is different and the 7-bit DIP / SW setting for specifying the corresponding channel is different.
[0052]
In each circuit that handles each channel, first, the header (2 bytes) of the 87-byte V data sent from the master board 165 is commonly received, while the corrected amplitude information in the V data corresponds. Only the modified amplitude information (1 byte) for the channel is received. Each circuit can check the reproduction time of the received header information in order to determine the number of clocks of 44.1 kHz for shaping and outputting the waveform. For example, when the instructed playback time is given as 50, 100 clock playback is performed (the playback time does not change), and when 110 is played, 220 clock playback is performed (the playback time is doubled). Each circuit stores the sine wave data of the frequency it is responsible for in the ROM at the data interval when it is output at 44.1 kHz (the sine wave wave of that frequency is accurately detected from address 0 to address N of the ROM address). M-waves are stored (M and N are natural numbers) The processor in each circuit increments the ROM address every time one sine wave data is created (every 22.68 μs). Next to the address, it returns to zero, so that an accurate sine wave can be created without discontinuities, but the above-mentioned sine wave data is converted into the basic data stored in the ROM. Multiply each piece to generate one sine wave data, and each amplitude information is a value obtained by linear interpolation between the current amplitude information and the previous amplitude information.
[0053]
As described above, the sine wave data generated by each circuit is an output buffer by referring to the ROM at a cycle of 44.1 kHz (22.68 μs) and multiplying the referenced data by the coefficient obtained by the above interpolation. Store in each of the RAMs # 1 to # 85.
[0054]
Then, the control signal from the master board 165 obtains the timing (22.68 μs cycle) at which the sine wave data stored in each of the RAMs # 1 to # 85 is sent to the output bus (16 bits), and is sent to the bus only at that time. To do. The time width given to one circuit is 226 ns (22.68 μs ÷ 85). On the other hand, the capture timing on the master board 165 side is given by a clock and a synchronization signal. The number of clocks from the synchronization signal is the same as the channel number specified by the DIP switch. Further, when the frequency shift instruction information included in the header information of the V data is ON, the sine wave data of the frequency shifted by a semitone (or full tone) can be output from each circuit to the ROM provided in each circuit. Stores various types of waveform data so that one can be selected.
[0055]
On the other hand, the master board 165 receives the sine wave data generated by the circuits 16-1 to 16-85 at a ratio of 85 data out of 22.68 μs (data interval is 22.68 μs ÷ 85 = 266 ns). In practice, the adder 173 adds the sine wave data from each circuit while receiving it to generate 44.1 kHz audio data (second audio information string) (step ST13). The generated audio data is sequentially stored in the RAM 174 which is a buffer and sent to the PC 1.
[0056]
The PC 1 records the audio data on the predetermined recording medium 15 while controlling the input / output device 14 with the sent audio data (step ST14).
[0057]
Next, each embodiment to which the technique disclosed in the above-mentioned Japanese Patent No. 2581700 is applied will be described.
[0058]
First, in the first embodiment, at least two types of audio information sequences and recording position identification information are recorded. That is, the first audio information sequence is composed of, for example, English audio information that is spoken by a native speaker at a natural speed, and this audio information sequence is generated as described above (the end of a sentence or a break in a sentence) Divided into a plurality of variable length segments. The second audio information sequence is an audio information sequence obtained by selectively editing the first information sequence as described above, and corresponds to each segment of the first audio information sequence. It is divided into multiple variable length segments. The recording position identification information is information indicating at which position on the audio recording medium at least each segment in the first and second audio information strings is recorded. Therefore, for example, the segment “It's ... not ... much ·· of ··· a ··· corresponding to the t-th segment"It's not much of a problem. " The position where “problem.” Is recorded can be recognized from this recording position identification information.
[0059]
As a result, the first and second audio information sequences and the recording position identification information are not recorded independently of each other but are recorded with a certain relationship, and each audio information sequence is organically combined in units of segments. . That is, the first and second audio information strings are paired with each other, and the recording position identification information is associated with each segment. In this embodiment, the recording / recording position identification information is recorded in the directory area of the audio information recording medium and includes at least information on the head position of each segment.
[0060]
In the reproduction of the audio information recording medium having the above-described structure, audio reproduction is performed in order for each recorded segment. In particular, in this reproduction method, the first audio recorded on the audio information recording medium is performed. It is characterized in that reproduction switching from the information sequence to the second audio information sequence (or reproduction switching from the second audio information sequence to the first audio information sequence) is possible. This reproduction switching operation is performed in units of segments. For example, when a reproduction instruction for the second audio information sequence is input while the t-th segment of the first audio information sequence is being reproduced (generation of an interrupt request), the second audio information is based on the recording position identification information. The corresponding t-th segment in the column is read, and the sound reproduction of the corresponding segment is executed. Conversely, playback switching from the second audio information sequence to the first audio information sequence is also performed in units of segments in the same manner as the above-described playback switching operation.
[0061]
In this reproduction, in addition to the reproduction switching operation described above, various modifications such as repeat reproduction are possible. A typical example is a so-called return command. That is, when a return command is input after temporary playback is interrupted by a stop command during playback, the voice information can be reproduced more appropriately to the operator's wish by returning the read position of the voice information by the commanded amount. Done.
[0062]
This second embodiment has basically the same structure as the first embodiment described above, but is equivalent to the contents of the first audio information sequence in addition to the first audio information sequence and the second audio information sequence. Although it is meaning content, it is another voice information, for example, it is characterized by having a third voice information sequence which is English voice information of slow speed speaking with words separated. The third audio information sequence is also composed of a plurality of variable length segments, and the recording position identification information manages the recording positions between the segments in the first to third audio information sequences. . Therefore, the reproducing operation in the second embodiment is the same as that in the first embodiment.
[0063]
In the second embodiment, the important thing is that the first audio information sequence and the third audio information sequence are each divided into a plurality of variable length segments. Is. For example, when the t-th segment (621 in FIG. 4A) of the first audio information sequence is “It's not much of a problem.” Spoken by the native speaker, the t-th segment of the third audio information sequence is The segment is "It is not much of a problem." However, the content corresponding to the second audio information string and consisting of different audio information indicates that the utterances are different in the same meaning in the language.
[0064]
Furthermore, in the third embodiment, in addition to the first and second audio information sequences, a fourth audio information sequence that is an audio information sequence such as a grammar commentary is recorded on the audio information recording medium. Different from the first embodiment.
[0065]
What is important here is that the third audio information sequence is divided into a group of one or more variable length segments of the first and second audio information sequences. In other words, one segment group of the fourth audio information sequence includes one or more segments of the first and second audio information sequences, and thus one segment group of the fourth audio information sequence is It is paired with one or more segments of the first and second audio information sequences. In particular, this configuration assumes a case where one sentence is divided into a plurality of segments as shown in FIG.
[0066]
In the third embodiment, the recording position identification information recorded in the predetermined area includes information indicating the recording position of the content of the fourth audio information sequence for each segment group. Therefore, the first, second and fourth audio information sequences and the recording position identification information are recorded on the medium with a certain relationship with each other, and each audio information sequence is organically combined in units of segments or segment groups. Also in the third embodiment, the recording position identification information is recorded in the directory area of the audio information recording medium, and includes information on the start position of the segment in each audio information string. Also in this embodiment, a third speech information sequence that is equivalent to the speech information of the first speech information sequence and has a slow speed may be recorded.
[0067]
The reproduction of the audio information recording medium having the structure as described above is basically the same as in the case of the first embodiment described above, but in addition to reproduction switching between the first and second audio information strings, The difference is that the playback switching operation is performed between the first and second audio information sequences and the fourth audio information sequence.
[0068]
For example, “It's not much of the native speaker during playback of the first audio information sequence”
If “problem” cannot be heard, the playback is switched from the first audio information sequence being reproduced to the second audio information sequence, so that the audio that has been selectively edited such as expanded
You can listen to "It's ... not ... much ... of ... a ... problem." And if you want to know the meaning and grammar of this Japanese language,
What is necessary is just to switch reproduction | regeneration to an audio | voice information sequence. Of course, it is needless to say that this reproduction method can also be applied so as to be used in combination with the return command and the stop command described in the reproduction operation in the first embodiment. Also in the playback in the third embodiment, switching playback and repeat playback are possible.
[0069]
The fourth embodiment is basically the same as the case of the first embodiment described above, except that a character information string is recorded in addition to the first and second audio information strings. This character information string corresponds to character information having contents corresponding to the first or second sound information string, and corresponds to, for example, character information corresponding to English (speech) spoken by a native speaker.
[0070]
This character information sequence is also divided into segments corresponding to the segments of the first and second audio information sequences. Also in the fourth embodiment, the recording position identification information includes information regarding the recording position of the character information string for each segment of each audio information string, and the audio information recording medium. Is recorded in the directory area. Therefore, the first and second audio information strings and the character information strings correspond to each other in segment units.
[0071]
In the fourth embodiment, when the fourth audio information sequence in the third embodiment is added as recording information, one or more segments of the first and second audio information sequences and the character information sequence are the first and second segments. This also corresponds to one segment group of the three audio information strings. Also in this configuration, the recording position identification information includes the head position of each segment and is recorded in the directory area of the audio recording medium. As in the third embodiment described above, the fourth embodiment also has a third voice information sequence that is equivalent to the voice information of the first voice information sequence and has a slow speed for speaking words. Further recording may be performed.
[0072]
The reproduction of the audio information recording medium having the above structure is basically the same as that of the second embodiment described above, but the character information string is displayed during the reproduction of the first or second audio information string. Different points are displayed.
[0073]
For example, when the segment “It's not much of a problem.” Of the first audio information sequence is being reproduced, “It's not much of a problem.” Or “It is not much of a problem.” "Is displayed on the display. Note that this display does not need to be completely synchronized with the audio information string being reproduced in time, and the characters may be displayed little by little or may be displayed little by little. In the reproduction in the fourth embodiment, switching reproduction and repeat reproduction are possible.
[0074]
Next, the specific structure of the audio recording medium will be described in detail below with reference to FIGS.
[0075]
FIG. 8 is a diagram for explaining each voice information sequence A, B, C when the third embodiment described above is applied for English conversation self-study as an example of a voice information recording medium, and the recorded contents thereof. In this figure, an audio information sequence A is an English information sequence (first audio information sequence) spoken by a native speaker, and is composed of a plurality of segments 621 and 622. As described with reference to the flowcharts shown in FIGS. 6 and 7, the audio information sequence B is an information sequence (second audio information sequence) edited so that a predetermined portion of the first information sequence is selectively expanded. ). The audio information sequence C is an information sequence (third audio information sequence) for explaining Japanese, and the segment group included in the audio information sequence C is the segments 621 and 622 of the audio information sequences A and B. It corresponds to each.
[0076]
FIG. 9 is a table for explaining the relationship between the time per segment and the capacity in the aspect shown in FIG. In this table, one second corresponds to a capacity of 6 kilobytes. For example, in the segment 621 of the audio information sequence A, the utterance time of “It's” is 0.2 seconds, the capacity is 1.2 KB (kilobytes), the utterance time of “not” is 0.1 seconds, and the capacity is 0.6 KB. (Kilobytes), “much of a” vocalization time is 0.4 seconds, its capacity is 2.4 KB (kilobytes), and “problem” vocalization time is 0.3 seconds, its capacity is 1.8 KB (kilobytes) The utterance time of the entire segment 621 is 2.0 seconds, and its capacity is 12 KB (kilobytes).
[0077]
Further, FIG. 10 is a table for explaining the recorded contents of the directory area in the form shown in FIGS. In this table, the directory area is composed of 9 × 3 = 27 bytes (B) per segment. The audio information sequences A, B, and C correspond to the audio information sequences A, B, and C in FIG. Further, C of 1 byte indicates an attribute, C = 0 means a voice information string A, and C = 64 means a voice information string B. Also, C = 128, 129 means the audio information sequence C, and in particular when C = 129, that is, when “10000001” in the bit representation (8 bits (bits)), it is the same explanation object as the previous segment. (Indicating that they belong to the same segment group to be explained in the audio information sequence C, for example, corresponding to the segments 801 and 802 in FIG. 4E).
[0078]
The position information M, S, and B (one byte each) are parameters representing the position on the CD-ROM that is standard in the industry. That is, M indicates minutes, S indicates seconds, and B indicates blocks. One block is 2,048 bytes, and 75 blocks constitute one second. Therefore, the maximum numbers are M = 59, S = 59, and B = 74. The next 2-byte SB indicates a start byte, and the next 3-byte LLL indicates the length of each segment. The reason why minutes and seconds are used for the parameter indicating the position is that the CD-ROM was originally developed for music, and the recording position is expressed as the time from the beginning. Therefore, when a CD-ROM is used as the audio information recording medium, the minutes and seconds are completely independent of the reproduction time, and are merely information indicating the recording position on the recording medium. Become.
[0079]
As a result, for example, “It's not much of a problem.” Of the segment 621 in the audio information sequence A is an English voice spoken by a natives beaker with a length of 6,000 bytes from the 826th byte of O block 11 seconds 3 blocks. The information is recorded, and the corresponding segment in the audio information sequence B is recorded in English of the native speaker, which is selectively extended with a length of 17,400 bytes from the 2,026 bytes of 0 minutes 11 seconds 3 blocks, In the segment group of the audio information string C, Japanese commentary is recorded with a length of 72,000 bytes from the 1st 282nd byte of 0 minute 11 seconds 6 blocks. Note that segment numbers such as 621 and 622 do not exist in the memory but correspond to their addresses. Further, the recording position identification information indicating the relationship between the segments is included in this directory area.
[0080]
More specifically, from the recorded contents of the directory area shown in FIG. 10 from the 826th byte to the 826 + 6,000-1 = 6,825th byte in the 0 minute 11 second 3 block of the audio information recording medium. In the area, an audio information string having a segment of 621 and an attribute C of 0, that is, information corresponding to “It's not much of a problem” spoken by a native speaker is recorded. Also, in the area from the 2,026th byte to the 2,026 + 17,400-1 = 19,425th byte in the 0 minute, 11 second, 3 block of the audio information recording medium, the audio of segment 621 and attribute C of 64 An information string, that is, selectively expanded audio information is recorded. Furthermore, in the area from the 1st 282nd byte to the 1st 282 + 72,000-1 = 73,281th byte in the 0 minute 11 second 6 block of the audio information recording medium, the audio of which the segment is 621 and the attribute C is 128 An information string, that is, information corresponding to a Japanese commentary is recorded.
[0081]
As described above, if the directory area shown in FIG. 10 is provided, each audio information sequence shown in FIG. 8 can be recorded with the reproduction time and capacity shown in FIG.
[0082]
Next, the information regarding each segment 621 and 622 is recorded on the header part of the variable-length segment shown by Fig.11 (a), for example. As shown in FIG. 11B, this header portion is a 1-byte area (1B) for indicating the presence / absence of character information and image information from the beginning, and an area prepared for the audio information string A. A 5-byte area (5B) composed of 1-byte data indicating information string type (information for distinguishing audio information strings A, B, etc.), 3-byte data indicating the data length, and spare 1-byte data A 5-byte area (5B) composed of 1-byte data indicating the information string type, 3-byte data indicating the data length, and spare 1-byte data, which is an area prepared for the audio information string B An area prepared for the information string C, a 4-byte area (4B) composed of 1-byte data indicating the information string type and 3-byte data indicating the data length, and an area prepared for the character information string D In 4-byte area (4B) composed of 1-byte data indicating the information string type and 3-byte data indicating the data length, and similarly 3-byte data indicating the address, which is an area prepared for the character information string D And a 6-byte area (6B) composed of 3-byte data indicating the data length, a 4-byte area (4B) prepared for another information sequence (type E) such as the third audio information sequence, and This is a 32-byte area consisting of a spare 3-byte area (3B).
[0083]
Next, the operation of reproducing audio information and the apparatus configuration will be described with reference to FIGS.
[0084]
First, FIG. 12 is a perspective view showing the overall configuration of a playback apparatus for realizing audio playback. As can be seen from this figure, the audio recording medium is, for example, a CD-ROM that can be played back by a portable CD player (playback apparatus main body 200). The playback apparatus main body 200 is remotely controlled by a corded handset 80. The This handset 80 is provided with at least a display unit 210 such as a liquid crystal display (LCD) for displaying a segment number being reproduced, and various control buttons 240. Further, the operator can listen to the audio information reproduced by the reproduction apparatus main body 200 via the earphone 130.
[0085]
FIG. 13 is a block diagram showing the configuration of the playback apparatus shown in FIG. As shown in this figure, the CD-ROM which is the audio information recording medium 15 is set in the reproduction mechanism 205. The playback mechanism 205 is connected to the CPU 50 via the disk interface (I / F) 30 and the bus 40. Also connected to the bus 40 are, for example, a 32 kilobyte (KB) ROM 60 for storing programs and a 256 kilobyte RAM 70 for temporarily storing directories and audio information strings. The bus 40 further includes a handset interface (I / F) 90 for transferring information to and from the handset 80 for manual operation, and an external terminal 110 and the handset 80 via an audio output amplifier (AMP) 100. Is connected to the D / A converter 12 connected to. Note that the earphone 130 is connected to the handset 80 as described above.
[0086]
FIGS. 14A and 14B are diagrams for explaining the memory allocation status of the ROM 60 and the RAM 70, respectively. As shown in FIG. 14A, the program is stored in the ROM 60 of 32 kilobytes. On the other hand, as shown in FIG. 14B, the RAM 70 includes a buffer of (50 + 50) = 100 kilobytes (corresponding to 50 blocks), a directory of (75 + 75) = 150 kilobytes, and a system of 6 kilobytes. Area is allocated. Therefore, the RAM 70 always holds an audio information string for 50 blocks, and holds a directory for 150 kilobytes ÷ 27≈5,555 segments (corresponding to about 30 minutes only for the audio information string A portion).
[0087]
In the above specific example, a CD-ROM is used as the audio information recording medium, but the typical capacity is 552 megabytes (MB). The CD-ROM uses units of minutes, seconds, and blocks to represent addresses. Since one block is 2,048 bytes, 75 blocks are 1 second, and 60 seconds is 1 minute, the maximum value of the address of the CD-ROM is 59 minutes 59 seconds 74 blocks. Conversely, the capacity of this CD-ROM is 2,048 × 75 × 60 × 60 = 552.96 megabytes. Among these, since the user cannot use the CD-ROM format for 2 seconds from the beginning, the maximum capacity is accurately set to 552.6528 MB. Furthermore, if a directory area is allocated up to 20 seconds from the beginning, a directory capacity of 3 megabytes can be secured in the CD-ROM.
[0088]
In the above-described embodiment, audio information recording software (a program capable of executing the above-described recording method on a personal computer or the like, or a recording medium on which the program is recorded), a dedicated recording device, a use manual, or these In addition to sales in combination, sales of the audio information recording medium alone, the audio information recording medium, playback software (including a program executable on a personal computer or the like, or a recording medium recording the program), a dedicated playback device , Sales manuals, or a combination of these.
[0089]
【The invention's effect】
As described above, according to the present invention, for a plurality of frequency components divided from the first audio information sequence sampled in the first period, the amplitude is changed (emphasized or attenuated) to a desired portion or the wave number is changed (reproduction time). The sine wave data is generated by adding the sine wave data of each of the frequency components, and is newly synthesized. A second audio information sequence is obtained. The desired audio information sequence generated in this way has an effect that the reproduction time can be extended or shortened at an arbitrary portion without changing the frequency, or the audio at the arbitrary portion can be reproduced as an enhanced or attenuated audio.
[0090]
In addition, the present invention can be combined with the technology disclosed in Japanese Patent No. 2581700, and an audio of an arbitrary part corresponding to a variable length section obtained by dividing the utterance sound of a native speaker by a utterance node. By separately preparing voice information that has been expanded and / or emphasized, beginner learners can repeatedly reproduce and listen to sounds that they have not been able to hear. It also has the effect of being able to hear. In addition, advanced learners can learn more actively in combination with the playback of utterances of native speakers by separately preparing audio information in which the audio of any part is shortened and / or attenuated. There is an effect.
[Brief description of the drawings]
FIG. 1 is a conceptual diagram for explaining generation and recording operations of audio information according to the present invention.
FIG. 2 is a table showing an example of each frequency component (channel) divided from sampled input voice information.
FIG. 3 is a diagram for explaining a basic shape of a speech spectrum.
FIG. 4 is a diagram for conceptually explaining various types of information including an audio information sequence to be recorded on an audio recording medium.
FIG. 5 is a diagram showing an overall configuration of a peripheral device for realizing a method for generating audio information according to the present invention.
FIG. 6 is a flowchart (No. 1) for explaining a method of generating audio information according to the present invention.
FIG. 7 is a flowchart (No. 2) for explaining a method of generating audio information according to the present invention.
FIG. 8 is a diagram for explaining each voice information sequence of a voice recording medium applied for English conversation self-study and the recorded contents thereof.
FIG. 9 is a table for explaining the relationship between time and capacity per segment for each audio information sequence shown in FIG. 8;
10 is a table for explaining recorded contents (including recording position identification information) of a directory area in the audio recording medium shown in FIGS. 8 and 9. FIG.
FIG. 11 is a diagram showing a configuration of variable length segments to be recorded on an audio recording medium.
FIG. 12 is a perspective view showing an overall configuration of a playback apparatus that realizes a playback method of an audio recording medium.
FIG. 13 is a block diagram showing a configuration of the playback apparatus shown in FIG.
14 is a diagram for explaining a memory allocation situation of the ROM and RAM shown in FIG. 13; FIG.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... PC, 10 ... Control system, 14 ... Input / output device, 15 ... Audio information recording medium, 19 ... V data, 16-1 to 16-85 ... Sine wave data generation circuit, 17, 175 ... DAC, 18, 177 ... speaker, 173 ... adder.

Claims (4)

Dividing the first audio information sequence sampled in the first period into a plurality of frequency components;
Correction in which each of the plurality of frequency components is selectively edited in one or more predetermined portions in an amplitude information sequence including amplitude information extracted in a second period corresponding to at least one period of sine wave data Generate an amplitude information sequence,
Each of the plurality of frequency components is prepared for each information component group composed of amplitude information extracted at the same timing corresponding to each other among the corrected amplitude information sequences of each of the frequency components, and each information component group. An audio information encoding method for generating V data comprising control information for instructing to extend or shorten the audio reproduction time with reference to the first period. 2. The audio information encoding method according to claim 1, wherein the control information includes frequency shift instruction information for reproducing the plurality of frequency components in a state shifted in a high pitch direction or a low pitch direction. A sine wave data corresponding to each of the plurality of frequency components, having amplitude given by V data generated by the speech information encoding method according to claim 1 and having a data interval of the first period, Generating sine wave data having a wave number corresponding to the reproduction time indicated by the control information included in the V data;
A speech information generation method for generating the second speech information sequence of the first period by sequentially adding the generated sine wave data. 4. Each amplitude of sine wave data corresponding to each of the plurality of frequency components is determined by a value obtained by linear interpolation between adjacent amplitude information in the modified amplitude information sequence. A method for generating the described audio information.

Images