JP2015087617A - Device and method for generating guide vocal of karaoke - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a guide vocal generation device which is more human-like and generates a guide vocal that gets a high score in a scoring result.SOLUTION: A reference voice synthesis parameter is corrected according to a setting from an operation part and a display part in parameter correction means 25, and technique data is added to a specified section in which a reference voice synthesis parameter is set by a degree setting, thereby to be set as a correction voice synthesis parameter; scoring means 24 performs a scoring based on reference voice synthesis data 42 generated from text character information and the reference voice synthesis parameter by voice synthesis means 23 and correction voice synthesis data 43 from the text character information and a correction voice synthesis parameter; and the correction voice synthesis parameter is corrected according to a changed setting when a setting is changed by an operation part and a display part based on a scoring result, and the correction voice synthesis data 43 is generated by the voice synthesis means 23, and the scoring is made by scoring means 24.

Description

  The present invention relates to a guide vocal generating apparatus and a guide vocal generating method for generating a guide vocal that guides a singer's song reproduced together with a song in a karaoke system.

  In recent years, karaoke systems have become more sophisticated, and singers can listen to instructional songs that serve as examples of songs that they want to practice. This is a so-called “singing support system”. This system records in advance a guide song by a model singer who can ideally sing a song to be practiced as a guide vocal. You can listen at the timing.

  Various techniques have been conceived in the past in connection with the singing instruction function of such a karaoke system. For example, in order to make it easy for a singer to sing a song, a technique for reproducing a vocal track that is recorded as closely as possible to the original singer's singing along with a karaoke performance, or as disclosed in Patent Document 1, A technique is also known in which a guide vocal is generated by sampling.

  By the way, it has become common to provide a scoring function in a karaoke system, and the pitch of a singer's singing signal is detected and compared with reference pitch information (reference data). Although the scoring is done, the reference data and the above guide vocal are generated for different purposes, so even if you can sing in the same way as the guide vocal with the support of singing by the guide vocal A situation occurs in which the scoring result is not always high.

  On the other hand, a reproduction mode of a large number of speech waveform data (phoneme data) using parameters such as lyric display character information attached to music data, pitch information for singing, timing information, and score information of volume information. There is known a technique for performing voice synthesis by controlling voice and artificially singing. For example, as in Patent Document 2, speech synthesis is performed by obtaining the above parameters for speech synthesis from MIDI (registered trademark) format data including pitch information, timing information, and volume information and lyrics information. is there.

JP 2009-244789 A JP-A-10-240264

  However, the guide vocal generated by the speech synthesis technique as in Patent Document 2 is generated based on the reference data such as pitch information, and the scoring is performed by comparison with the reference data. Singing in the same way as above, a high score can be obtained, but the guide vocal generated by such speech synthesis technology is an artificial (mechanical) mellow singing voice, and the scoring result is a high score However, there is a problem that it is different from the scoring results, unlike the impression that people have heard.

  Accordingly, the present invention has been made in view of the above problems, and an object thereof is to provide a guide vocal generating apparatus and a guide vocal generating method for generating a guide vocal that is more human and has a high scoring result.

  In order to solve the above-mentioned problem, in the invention of claim 1, for each karaoke piece, lyric character information for displaying lyrics to the singer, pitch information, timing information to be sung at least by the singer, A guide vocal generating device that generates guide vocals based on performance data having reference data including volume information, wherein the performance database storing the performance data for each piece of music and various singing techniques are converted into data when singing A singing technique database for storing the technique data, and at least selecting a tune for which guide vocals are to be generated from the performance data, setting information selection in the reference data and its degree, and selecting the technique data and The external input means for setting the degree and the specific section to be added, and the external input means are selected. A data acquisition unit that acquires lyric character information from the performance database based on the performed music, and acquires reference speech synthesis parameters of pitch information, timing information, and volume information from the reference data; and the data acquisition unit The reference speech synthesis parameter acquired from the external input unit is modified according to the setting from the external input unit, and the technique data selected by the external input unit is set to the specific section in which the reference speech synthesis parameter is set according to the degree setting. Parameter correcting means that is added as a modified speech synthesis parameter, and lyric character information obtained by the data obtaining means by generating reference speech synthesis data from the lyric character information obtained by the data obtaining means and the reference speech synthesis parameter And the modified speech synthesis parameters modified by the parameter modification means Speech synthesis means for generating corrected speech synthesis data; and scoring means for scoring by comparing the reference speech synthesis data generated by the speech synthesis means and the modified speech synthesis data, the parameter correction means, When there is a setting change from the external input unit based on the scoring result by the scoring unit, the modified speech synthesis parameter corresponding to the change setting is modified, and the modified speech synthesis data is generated by the speech synthesis unit, and the scoring It is set as the structure which makes a score by means.

  The invention of claim 2 is provided with display means, and the scoring means is configured to display scoring information of each section on the display means when the scoring result is not full.

  In the invention of claim 3, in order to score the lyric character information for displaying lyrics to the singer corresponding to an arbitrary karaoke piece and the singing of the karaoke piece by the singer for each predetermined singing section A guide vocal generating device for generating a guide vocal for supporting the singing of the karaoke music based on reference data including pitch information, timing information, and volume information of the lyric character information and the reference data A plurality of speech synthesis parameters including pitch information, timing information, and volume information included in the reference data input by the data input means, the lyric character information input by the data input means, and the lyric character information input by the data input means Is generated by a speech synthesis unit that generates speech synthesis data and a speech synthesis unit. Based on the plurality of speech synthesis parameters including playback means for reproducibly reproducing the speech synthesis data, parameter correction means for correcting the speech synthesis parameters, and the speech synthesis parameters corrected by the parameter correction means The modified speech synthesis data generated by the speech synthesis means includes scoring means for scoring for each predetermined singing section based on the reference data input by the data input means and a predetermined scoring rule.

  In the invention of claim 4, comprising a display means, the scoring means, when the scoring result is less than a predetermined value, the scoring information of the singing section related to the speech synthesis parameter corrected by the parameter correction means, It is the structure displayed on a display means.

  In the invention of claim 5, in order to score lyric character information for displaying lyrics to a singer corresponding to an arbitrary karaoke piece and the singing of the karaoke piece by the singer for each predetermined singing section A guide vocal generating method for generating a guide vocal for supporting the singing of the karaoke music on the basis of reference data including pitch information, timing information and volume information of the lyrics character information and the reference data A speech input step for generating speech synthesis data based on a plurality of speech synthesis parameters including the lyric character information, the pitch information, the timing information, and the volume information, A reproduction step for reproducing the voice synthesis data, a parameter correction step for correcting the voice synthesis parameter, A modified speech synthesis step for generating modified speech synthesis data based on the plurality of speech synthesis parameters including the modified speech synthesis parameters, and the modified speech synthesis data based on the reference data and a predetermined scoring rule A scoring step for scoring for each predetermined singing section; and a display step for displaying scoring information of the singing section related to the modified speech synthesis parameter when the scoring result is less than a predetermined value; To do.

  According to the first aspect of the present invention, the parameter correcting unit corrects the reference speech synthesis parameter in accordance with the setting from the external input unit, and the technique data is set to the specific section in which the reference speech synthesis parameter is set by the degree setting. A scoring means is added based on the reference speech synthesis data generated from the lyric character information and the reference speech synthesis parameter by the speech synthesizer and the modified speech synthesis data from the lyric character information and the modified speech synthesis parameter. If the setting is changed from the external input unit based on the scoring result, the modified speech synthesis parameter corresponding to the changed setting is corrected, the modified speech synthesis data is generated by the speech synthesis unit, and the scoring unit By adopting a scoring configuration, the speech synthesis parameters are corrected and generated within the range where the scoring results are high. Those that can be generated to allow humanly listening than the modified speech synthesis data as a guide vocal.

  According to the invention of claim 2, the display means is provided, and when the scoring result of the scoring means does not reach a perfect score, the scoring information of each section is displayed on the display means, thereby correcting the scoring by the parameter correcting means. It can be made easy.

  According to the third and fifth aspects of the present invention, there are a plurality of lyrics character information input by the data input means and a plurality of pitch information, timing information, and volume information included in the reference data input by the data input means. Based on the speech synthesis parameters, the speech synthesis data generated by the speech synthesis means is audibly reproduced based on the speech synthesis parameters, and based on a plurality of speech synthesis parameters including the speech synthesis parameters obtained by correcting the speech synthesis parameters by the parameter correction means. The lyric character information is formed by scoring the corrected speech synthesis data generated by the speech synthesizer for each predetermined singing section based on the reference data input by the data input unit and a predetermined scoring rule. Range in which the scoring result is high even if the reference data is entered Are those which can be generated as a modified speech synthesis data generated by modifying the speech synthesis parameters can be humanly listening than as a guide vocal.

  According to the invention of claim 4, the display means is provided, and when the scoring result of the scoring means is less than a predetermined value, the scoring information of the singing section related to the speech synthesis parameter corrected by the parameter correcting means is displayed on the display means. By displaying it, it is possible to facilitate correction with a high score by the parameter correcting means.

It is a block block diagram of 1st Embodiment in the guide vocal production | generation apparatus of the karaoke which concerns on this invention. It is explanatory drawing of the performance data and singing technique data of FIG. It is a flowchart of operation | movement description in the guide vocal production | generation apparatus of this invention. It is processing explanatory drawing of the speech synthesis means in FIG. It is processing explanatory drawing of the scoring means in FIG. It is a block block diagram of 2nd Embodiment in the guide vocal production | generation apparatus of the karaoke which concerns on this invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 shows a block diagram of the first embodiment of the karaoke guide vocal generating apparatus according to the present invention, and FIG. 2 shows an explanatory diagram of the performance data and singing technique data of FIG. A guide vocal generating apparatus 11 shown in FIG. 1 includes an operation unit 13 and a display unit 14 that are externally connected to the apparatus main body 12 in a wired or wireless manner. The operation unit 13 and the display unit 14 constitute an external input unit.

  The operation unit 13 is used to input various signals to the apparatus main body 12 and includes a keyboard and a mouse. Here, at least in cooperation with the display unit 14, a musical piece for which a guide vocal is to be generated is selected from the performance data, the selection of information in reference data described later and the degree thereof are set, and the singing technique data described later is selected. It becomes an input means for setting the selection, its degree, and a specific section to be added.

  The display unit 14 displays a display screen necessary for generating guide vocals, and sends various setting values input from the operation unit 13 to the apparatus main body 12. In addition, the operation unit 13 and the display unit 14 are integrated, and a liquid crystal display (LCD) and a touch sensor are stacked for input / output, and data such as selection of music by the touch sensor corresponding to displayed icons and the like. It is possible to have a GUI user interface function.

  The apparatus body 12 includes a bus 21, a control unit 22, a speech synthesis unit 23, a scoring unit 24, a data acquisition unit 25, a parameter correction unit 26, a display control unit 27, a performance data database (performance DB) 28, and a singing technique database. A (DB) 29 and a guide vocal database (DB) 30 are provided as appropriate. In addition, a storage area of the element storage unit 41 is formed in the speech synthesizer 23. In the scoring means 24, an area for storing the reference speech synthesis data 42 to be scored and an area for storing the modified speech synthesis data 43 are formed.

  The control unit 22 is a physical CPU that performs overall process control of the guide vocal generation device, and performs algorithm processing based on a program stored in a ROM (not shown). The speech synthesis means 23 is a program that performs speech synthesis based on speech synthesis parameters (corrected speech synthesis parameters), lyrics character data, and phoneme data stored in the element storage unit 41, and specific speech synthesis processing is as follows. The method described in Patent Document 2 described above can be used.

  The scoring unit 24 is a program that performs scoring processing by comparing speech synthesis data, and compares the reference speech synthesis data 42 generated by the speech synthesis unit 23 with the modified speech synthesis data 43 for scoring. Specifically, for example, the method described in Japanese Patent No. 4222915 can be used. Basically, the correction of the speech synthesis parameter is a deduction evaluation, but the expression power is increased by adding the singing technique. It becomes a score evaluation. In addition, when the scoring result is not full, the scoring information of each section may be displayed on the display unit 14 via the display control unit 27.

  The data acquisition unit 25 acquires lyric character information from a performance DB 28 described later based on the music selected from the operation unit 13 and the display unit 14 described later, and includes pitch information, timing information, and timing information from reference data described later. This is a program for acquiring a reference speech synthesis parameter for volume information.

  The parameter correction unit 26 corrects the reference speech synthesis parameter acquired from the data acquisition unit 25 according to the settings from the operation unit 13 and the display unit 14 and details the selected technique data to the extent that will be described in detail later. When the setting is changed from the operation unit 13 and the display unit 14 based on the scoring result by the scoring means 24, it is added to the specific section of the reference speech synthesis parameter according to the setting, and the modified speech synthesis parameter is used. This is a program for correcting a modified speech synthesis parameter corresponding to a change setting. The display control unit 27 is a program or an electronic circuit that creates display data for a display screen to be displayed on the display unit 14, sends the display data to the display unit 14, and inputs input data from the display unit 14.

  As shown in FIG. 2A, the performance DB 28 is a content database that stores performance data such as note information and lyrics character information for each karaoke piece. Specifically, it has a song table in which a song ID, song name, and artist ID (artist name) are associated, and each song has a predetermined data format (for example, MIDI (registered trademark) format) managed by the song ID. The musical note information and the like of each musical piece are stored as karaoke performance data, and at least reference data and lyrics character data including pitch information, timing information and volume information for scoring a song are stored.

  The singing technique DB 29 is a database that stores technique data obtained by converting various singing techniques into data at the time of singing. As shown in FIG. 2B, vibrato information, intonation information, long tone information, and running information are stored. Singing technique data such as information, fist information, shackle information, fall information, and shout information.

  The guide vocal DB 30 is a database that stores guide vocals (corrected guide vocals) that are voice synthesis data (corrected voice synthesized data) for each piece of music finally generated by the voice synthesizer 23 in association with the music ID. It is. The database of the performance DB 28, singing technique DB 29, and guide vocal DB 30 may be constructed in, for example, a hard disk included in the apparatus main body 12, or may be constructed in an external storage device connected to the apparatus main body 12.

  Therefore, FIG. 3 shows a flowchart for explaining the operation of the guide vocal generating apparatus of the present invention, FIG. 4 shows a process explanation diagram of the speech synthesis means in FIG. 1, and FIG. 5 shows a process explanation diagram of the scoring means in FIG. In FIG. 3, first, an input screen for inputting an object for generating a guide vocal is displayed on the display unit 14, and the creator inputs a song for creating a guide vocal from the operation unit 13, for example, a song ID, and the apparatus main body. 12 (step (S) 1).

  When the music ID is input from the operation unit 13 and the display unit 14, the data acquisition unit 25 specifies the performance data of the music ID specified with reference to the performance DB 28 and acquires the lyric character data in the performance data. The voice synthesis parameters (pitch information, timing information, volume information, section information) are acquired from the reference data in the performance data and sent to the voice synthesis unit 23, and the voice synthesis parameters are sent to the parameter correction unit 26. (S2).

  As shown in FIG. 4A, the speech synthesizer 23 uses reference speech synthesis data (reference guide) as a reference based on phoneme data, speech synthesis parameters and lyrics character information sent from the data acquisition unit 25. Vocal) is generated and sent to the scoring means 23 (S3). The scoring means 23 stores the reference speech synthesis data 42 sent from the speech synthesis means 23.

  On the other hand, the parameter correction means 26 sends a display screen for setting the type of correction target of the speech synthesis parameter, the song section to be corrected, and the type of technique data to the display unit 14 via the display control unit 27. Display (S4). On this display screen, the creator sets the type of speech synthesis parameter correction target, sets the correction singing section, sets the type of technique data, and sends it to the apparatus main body 12 (S5). Here, for example, the speech synthesis parameter to be corrected is pitch information and the technique data is vibrato. However, the other parameters of the speech synthesis parameter and the other technique data of the singing technique data are corrected. With respect to, it is only necessary to change the set values according to these, and it can be dealt with in the same manner as pitch information and vibrato.

  In the apparatus main body 12, the parameter (pitch information) selection is input, and the selection of the singing section and the vibrato selection in the technique data are input (S6). In the parameter correction means 26, the pitch information is changed and set, the vibrato information is acquired from the singing technique DB 29, and the data of the display screen for setting the degree of vibrato numerically is displayed on the display unit 14 via the display control unit 27. Display (S7).

  On this display screen, the creator is, for example, a mechanical pitch information in which the pitch changes at an accurate timing in a section in which notes of different pitches are continuous. The pitch information is set so that the pitch changes gently between notes so that the actual pitch changes as if singing. Since the mechanical pitch information is such that the pitch does not change, the degree of vibrato is set numerically and sent to the apparatus main body 12 (S8). Although not particularly illustrated, when setting is made as an object candidate in the above setting, the setting operation may be performed while listening to the creator by temporarily performing speech synthesis.

  Here, vibrato will be described with reference to FIG. FIG. 5 is a graph showing the pitch of the audio frequency sampled every 30 ms in time series. In the scoring means, as shown in FIG. 5 (A), the cycle and the pitch width are constant, and this is repeated. Therefore, it is evaluated as vibrato and added. On the other hand, for example, even when the period and the pitch width are constant, if one period is equal to or greater than the predetermined number of samplings or less than the predetermined number of samplings, it is not evaluated as vibrato, and as shown in FIG. Even if the width is constant, evaluation is made as vibrato even when the variation of each cycle is greater than or equal to the predetermined number of samplings, or even when the cycle is constant and the variation of each pitch is greater than or equal to the predetermined pitch as shown in FIG. Because it does not, it is not eligible for points.

  Since the vibrato shown in FIG. 5A is heard as an extremely artificial song, in view of being heard as a song with a more human taste by shifting this, the worker is more human. For example, as shown in FIG. 5D, within the range of the number of samplings in one cycle evaluated as vibrato, and the variation in each cycle and each pitch width as vibrato Each period and pitch width can be set differently so as to be within the evaluated range.

  Returning to FIG. 4, when the pitch information correction setting and vibrato setting are input (S8), the parameter correction means 26 uses the voice synthesis parameters acquired by the data acquisition means 25 in each singing section designated. Then, it is corrected with the numerical value of the input pitch information, added with the vibrato data of the vibrato numerical value, and sent to the speech synthesizing means 23 as a corrected speech synthesis parameter (S9).

  As shown in FIG. 4B, the speech synthesizer 23 generates modified speech synthesis data based on the modified speech synthesis parameters, lyric character information, and phoneme data, and sends the modified speech synthesis data to the scoring unit 24. (S10). The scoring unit 24 acquires and stores the modified speech synthesis data 43 from the speech synthesis unit 23, compares the stored reference speech synthesis data 42 with the modified speech synthesis data 43, performs a scoring process, and performs scoring results and A display screen for confirming or resetting is sent to the display unit 14 via the display control unit 27 (S11).

  For example, when the scoring result does not reach a perfect score, the creator selects resetting (S12), and the process returns to step (S) 4 to reset. In the scoring unit 24, when the scoring result is not full, scoring information for each section may be displayed on the display unit 14. As a result, it is possible to facilitate the correction to be a perfect score by the parameter correcting means 26.

  On the other hand, when the scoring result is full, the creator is satisfied and sends a confirmation signal to the apparatus main body 12 (S12), so that the speech synthesis means 23 uses the modified speech synthesis data as guide vocal data of the music. Then, it is stored in the guide vocal DB 30 in association with the music ID (S13). In addition, it is possible to listen to the modified guide vocal that has reached the full score to the creator, so even if it is full, it can be reset to give a little more human taste, and other singing sections and other voice synthesis Parameters and other singing techniques can be set.

  In this way, by correcting the speech synthesis parameters that are one element of guide vocal generation and adding singing techniques, it is generated so that it can be heard more as a guide vocal within the range where high scoring scores can be obtained. Is something that can be done.

Next, FIG. 6 shows a block diagram of a second embodiment of the karaoke guide vocal generating apparatus according to the present invention. In FIG. 6, in the first embodiment described above, the guide vocal generating device includes the performance DB 28, the singing technique DB 29, and the guide vocal DB 30, and data is acquired from these databases by the data acquisition unit 25. Instead, the data input means 31 may be configured to input only data necessary for generating a guide vocal for an arbitrary karaoke piece from the outside. More specifically, a data file may be downloaded from a server via a transmission path or acquired via a storage medium.

Furthermore, using the lyric character information input by the data input means 31, a plurality of speech synthesis parameters including pitch information, timing information, and volume information included in the reference data input in the same manner, and phoneme data, The speech synthesizing data is once generated by the synthesizing means 23, and the speech synthesizing data is corrected by the parameter correcting means 26 so that the generated speech synthesized data can be heard by the human being while being reproduced by the reproducing means 32. The modified speech synthesis data 46 may be generated by the synthesis means 23.

The reproduction means 32 converts the voice synthesis data into an analog signal so that the operator of the apparatus can audition the voice synthesis data, amplifies it appropriately, and emits the sound from the sound emission means 34 such as a speaker.

Furthermore, in the first embodiment, the scoring unit 24 compares the reference speech synthesis data 42 and the modified speech synthesis data 43. In the second embodiment, the scoring unit 33 is input by the scoring rule 44 and the data input unit 31. On the basis of the reference data 45, the modified speech synthesis data 46 may be scored for each predetermined singing section, and a scoring value obtained by counting the grading results of each singing section may be notified.

Here, the scoring rules 44 are the same as those installed in a conventional karaoke apparatus having a singing scoring function. More specifically, the singing voice signal of the singer is analyzed, and, for example, if the pitch (pitch) change in the singing section to be scored satisfies a predetermined condition, it is added as a high-quality vibrato expression, etc. The scoring standard is defined. And it is the scoring rule 44 with each scoring standard about various singing expressions. In the present invention, instead of a singing voice signal by a singer, it is used for analyzing and scoring modified voice synthesis data that has been corrected so that it can be heard more humanly.

In this way, even if the speech synthesis parameters are modified so that the mechanical guide vocal generated based on the lyric character information and the reference data can be heard by humans, the modified speech synthesis data is scored. It is possible to confirm that the singing does not deviate from the range determined to be an appropriate singing, so that a high-scoring score can be obtained and a guide vocal that can be heard more humanly can be generated.

In the second embodiment, a display unit is further provided, and when the corrected speech synthesis data 46 is scored by the scoring unit 33, if the score value is less than a predetermined value (for example, a perfect score), the corrected speech synthesis parameter is displayed. You may comprise so that the scoring information of the song area which concerns may be displayed. The scoring information here may include the scoring item name of the singing expression that has been deducted, the degree of deviation from the appropriate value, etc., in addition to the scoring value of the singing section.

As a result, when the modification of the speech synthesis parameters performed for the purpose of being able to be heard by humans deviates from the range determined as appropriate singing based on the reference data and the scoring rules, the deduction is made. The speech synthesis parameter that caused the problem can be corrected again based on the scoring information, and the score of the modified speech synthesis data can be increased. The display unit 14 and the display control unit 27 correspond to display means.

  Further, when controlling the data input means 31, the voice synthesis means 23, the reproduction means 32, the parameter correction means 26, the scoring means 33, and the display means, the operation unit 13 and the display unit 14 are the same as in the first embodiment. The control unit 22 functions as a user interface and controls the entire apparatus.

  The guide vocal generating device of the present invention can be used in the industrial field of karaoke devices having basic karaoke functions.

DESCRIPTION OF SYMBOLS 11 Guide vocal production | generation apparatus 12 Apparatus main body 13 Operation part 14 Display part 21 Bus 22 Control part 23 Speech synthesis means 24,33 Scoring means 25 Data acquisition means 26 Parameter correction means 27 Display control part 28 Performance data 29 Technical data 30 Guide vocal database (DB)
31 Data input means 32 Playback means 41 Element storage section 42 Reference speech synthesis data 43, 46 Modified speech synthesis data 44 Scoring rule 45 Reference data

Claims (5)

Lyric character information for displaying lyrics to the singer for each karaoke song, and pitch information, timing information, and volume information for scoring the singing of the karaoke song by the singer for each predetermined singing section A guide vocal generating device that generates a guide vocal for supporting the singing of the karaoke music based on performance data having reference data including:
A performance database storing the performance data for each of the music pieces;
A singing technique database storing technique data obtained by converting various singing techniques into data when singing,
At least the guide vocal generation target music is selected from the performance data, the selection of the information in the reference data and the degree thereof are set, and the selection of the technique data and the degree thereof and the specific section to be added are set externally Input means;
Data acquisition means for acquiring lyric character information from the performance database based on the music selected by the external input means, and for acquiring reference speech synthesis parameters for pitch information, timing information and volume information from the reference data When,
The reference speech synthesis parameter acquired from the data acquisition unit is modified according to the setting from the external input unit, and the technique data selected by the external input unit is set according to the degree setting. Parameter correction means for adding to the specified specific section and making it a modified speech synthesis parameter;
Reference speech synthesis data is generated from the lyric character information acquired by the data acquisition unit and the reference speech synthesis parameter, and corrected from the lyric character information acquired by the data acquisition unit and the modified speech synthesis parameter corrected by the parameter correction unit. Speech synthesis means for generating speech synthesis data;
Scoring means for scoring by comparing the reference speech synthesis data generated by the speech synthesis means and the modified speech synthesis data;
Have
The parameter correcting means corrects a modified speech synthesis parameter corresponding to the change setting when the setting is changed from the external input means based on a scoring result by the scoring means, and the speech synthesis means corrects the synthesized speech Generating data and scoring by the scoring means,
A guide vocal generator characterized by the above. A guide vocal generating device according to claim 1,
A display means,
The scoring means displays the scoring information of each section on the display means when the scoring result is not full, the guide vocal generating device. Lyric character information for displaying lyrics to a singer corresponding to an arbitrary karaoke piece, pitch information and timing information for scoring the singing of the karaoke piece by the singer for each predetermined singing section And a guide vocal generating device for generating a guide vocal for supporting the singing of the karaoke music based on reference data including volume information,
Data input means for inputting the lyric character information and the reference data;
Speech synthesis data based on lyrics character information input by the data input means and a plurality of speech synthesis parameters including pitch information, timing information, and volume information included in the reference data input by the data input means Speech synthesis means for generating
Reproducing means for reproducing the voice synthesis data generated by the voice synthesizing means in an audible manner;
Parameter correcting means for correcting the speech synthesis parameter;
The modified speech synthesis data generated by the speech synthesis unit based on the plurality of speech synthesis parameters including the speech synthesis parameter modified by the parameter modification unit, the reference data input by the data input unit and a predetermined scoring Scoring means for scoring for each predetermined singing section based on the rules;
A guide vocal generating device characterized by comprising: A guide vocal generating device according to claim 3,
A display means,
The scoring means, when the scoring result is less than a predetermined value, displays on the display means the scoring information of the singing section related to the speech synthesis parameter corrected by the parameter correcting means. Generator. Lyric character information for displaying lyrics to a singer corresponding to an arbitrary karaoke piece, pitch information and timing information for scoring the singing of the karaoke piece by the singer for each predetermined singing section And a guide vocal generating method for generating a guide vocal for supporting the singing of the karaoke song based on reference data including volume information,
A data input step for inputting the lyric character information and the reference data;
A speech synthesis step of generating speech synthesis data based on the lyrics character information and a plurality of speech synthesis parameters including the pitch information, the timing information, and the volume information;
A reproduction step of reproducing the speech synthesis data;
A parameter correcting step for correcting the speech synthesis parameter;
A modified speech synthesis step for generating modified speech synthesis data based on the plurality of speech synthesis parameters including the modified speech synthesis parameters;
A scoring step of scoring the modified speech synthesis data for each predetermined singing section based on the reference data and a predetermined scoring rule;
When the scoring result is less than a predetermined value, a display step for displaying scoring information of a singing section related to the modified speech synthesis parameter;
A guide vocal generation method characterized by comprising:

Images