JP6488767B2 - Singing evaluation device and program - Google Patents

Description

  The present invention relates to a technique for evaluating a song.

  Various singing evaluation techniques for evaluating the skill of singing for each singing part by analyzing the sung sound (hereinafter referred to as “singing sound”) for music having a plurality of singing parts have been proposed. For example, in Patent Document 1, in the technique for evaluating the singing voice of a singer having a plurality of singing parts, evaluation data representing the pitch of the singing voice (singing pitch) and the pitch that the user should sing. The structure which evaluates the skill of a song for every song part is disclosed by comparing with.

JP 2008-268368 A

  However, in the configuration of Patent Document 1, it is necessary to prepare evaluation data for each of a plurality of singing parts, and there is a situation that the burden of creating the evaluation data is large. In view of the above circumstances, an object of the present invention is to appropriately evaluate the skill of a singing voice for a plurality of singing parts without assuming the existence of data for evaluation.

In order to solve the above problems, the singing evaluation apparatus according to the first aspect of the present invention includes a plurality of first pitch stable sections in which the pitch of the first singing signal indicating the singing voice of one piece of music is stable, A section setting unit that sets a plurality of second pitch stable sections in which the pitch of the second singing signal indicating other singing voices of the music is stable, and a plurality of first pitch stability set by the section setting unit For each overlapping section in which each of the sections and each of the plurality of second pitch stable sections overlap on the time axis, a representative value of the pitch of the first singing signal and a representative value of the pitch of the second singing signal And whether the representative value calculation unit for calculating and the representative value of the pitch of the first singing signal and the representative value of the pitch of the second singing signal in each of the overlapping sections have a predetermined pitch relationship. And an evaluation unit that evaluates the degree of harmony between the one singing voice and the other singing voice. In the above configuration, the degree of harmony between the singing voices is evaluated according to whether or not the representative value of the pitch of the first singing signal and the representative value of the pitch of the second singing signal are in a predetermined pitch relationship. Therefore, even when there is no evaluation data for each singing part of the music, it is possible to appropriately evaluate the degree of harmony between the singing voice of one singing part and the singing voice of another singing part. Further, in the above configuration, a plurality of first pitch stable sections where the pitch of the first singing signal is stable and a plurality of second pitch stable sections where the pitch of the second singing signal is stable overlap. Since the relationship between the representative value of the pitch of the first singing signal and the representative value of the pitch of the second singing signal is determined for each section, the singing pitch over the entire song is compared with the configuration to be evaluated. Thus, it is possible to appropriately evaluate the singing pitch while reducing the calculation load.
Here, as the predetermined pitch relationship, for example, the representative value of one pitch stable section and the representative value of another pitch stable section differ from each other by an integral multiple of a semitone (100 cents) in the twelve average temperament. The relationship is illustrated. Examples of the representative value include an average value and a median value.

  In a preferred example of the singing evaluation apparatus according to the first aspect, a pitch analysis unit that sequentially analyzes the pitch of the first singing signal and a receiving unit that receives the pitches sequentially analyzed from the second singing signal. The section setting unit includes a plurality of first pitch stable sections in which the pitch analyzed by the pitch analysis unit of the first singing signal is stable, and the receiving unit of the second singing signal includes the receiving unit. A plurality of second pitch stable sections in which the received pitch is stable are set. In the above configuration, since the pitch analyzed from the second singing signal is received by the receiving unit, it is not necessary to analyze the pitch of the second singing signal by the pitch analyzing unit. Therefore, the processing load by the pitch analysis unit can be reduced as compared with the configuration in which the pitch analysis unit analyzes the pitch of the second singing signal in addition to the pitch of the first singing signal.

  In a preferred example of the singing evaluation apparatus according to the first aspect, the pitch analysis unit sequentially analyzes the pitch of the first singing signal for each first analysis point on the time axis, and the reception unit The pitch of the second singing signal sequentially analyzed for each second analysis point on the axis is received, and the section setting unit is configured to mutually interact between the plurality of first analysis points and the plurality of second analysis points. Are matched with each other on the time axis, and an overlapping section in which the first pitch stable section and the second pitch stable section overlap is set. In the above configuration, since the first analysis point and the second analysis point are matched with each other on the time axis, the pitch of one singing part (for example, one singing voice) at the analysis point that matches on the time axis. There is an advantage that it is possible to appropriately evaluate the degree of harmony with the pitch of other singing parts (for example, other singing voices).

  The singing evaluation apparatus according to the second aspect of the present invention shows a plurality of first pitch stable sections in which the pitch of the first singing signal indicating one singing voice of the music is stable, and other singing voices of the music. A section setting unit for setting a plurality of second pitch stable sections in which the pitch of the second singing signal is stabilized, and calculating a representative value of the pitch in each of the plurality of first pitch stable sections, A representative value calculation unit that calculates a representative value of the pitch in each of the plurality of second pitch stable sections, and a first frequency distribution that indicates a frequency distribution of the representative values of the pitches over the plurality of first pitch stable sections. A pitch distribution generation unit that generates a second frequency distribution indicating a frequency distribution of representative values of pitches over the plurality of second pitch stable sections, and the first frequency distribution as a sound of a scale tone. Divide into multiple unit ranges centered on the height, and overlap each unit range with each other. The first evaluation distribution is created by summing the frequency of each frequency over a plurality of unit ranges for each pitch, while the second frequency distribution is divided into a plurality of unit ranges centered on the pitch of the scale sound. An analysis processing unit that creates a second evaluation distribution by overlapping ranges with each other and summing the frequency of the distribution of each unit range over a plurality of unit ranges for each pitch, the first evaluation distribution, and the second And an evaluation unit that evaluates the degree of similarity in singing pitch between the one singing voice and the other singing voice based on the evaluation distribution. In the above configuration, the first frequency distribution of the representative value of the pitch of the first singing signal over the plurality of first pitch stable sections and the pitch of the second singing signal over the plurality of second pitch stable sections. Since the degree of similarity in the singing pitch between one singing voice and the other singing voice is evaluated according to the second frequency distribution of the representative value, there is no evaluation data to be compared with the singing voice. Even in this case, it is possible to appropriately evaluate the degree of similarity in tendency in the singing pitch between one singing voice and another singing voice. Moreover, compared with the structure by which all the sections of a 1st song signal and a 2nd song signal are made into evaluation object, it becomes possible to evaluate a song pitch appropriately, reducing the processing load of a calculation. There is.

  The singing evaluation apparatus according to each of the above aspects is realized not only by dedicated hardware (electronic circuit) but also by cooperation of a general-purpose arithmetic processing apparatus such as a CPU (Central Processing Unit) and a program. The program of the present invention can be provided in a form stored in a computer-readable recording medium and installed in the computer. The recording medium is, for example, a non-transitory recording medium, and an optical recording medium (optical disk) such as a CD-ROM is a good example, but a known arbitrary one such as a semiconductor recording medium or a magnetic recording medium This type of recording medium can be included. For example, the program of the present invention can be provided in the form of distribution via a communication network and installed in a computer. Moreover, this invention is specified also as the operation | movement method (singing evaluation method) of the song evaluation apparatus which concerns on each aspect demonstrated above.

It is the schematic of the song evaluation system 1 which concerns on 1st Embodiment. It is a graph of the pitch PA of the 1st song signal V1, and the pitch PB of the 2nd song signal V2. It is explanatory drawing about the setting of the analysis area TA and the 1st pitch stable area TSA. It is a flowchart of operation | movement of the song evaluation part 22 (pitch analysis part 222, section setting part 224). It is a flowchart of operation | movement of the process of the song evaluation part 22 (evaluation part 228). It is explanatory drawing which shows an example of distribution of the evaluation value with respect to the difference value C. It is the schematic of the song evaluation system 1 which concerns on 2nd Embodiment. It is a graph of the pitch PA of the 1st song signal V1 which the pitch analysis part 222 analyzed, and the pitch PB of the 2nd song signal V2 which the communication apparatus 15 received. It is a graph of the pitch PA of the 1st song signal V1 and the pitch PB of the 2nd song signal V2 after the section setting part 224 matched the 1st analysis point KA and the 2nd analysis point KB. It is the schematic of the song evaluation system 1 which concerns on 4th Embodiment. It is explanatory drawing about the process of the pitch distribution generation part 225 of 4th Embodiment. It is explanatory drawing about the process of the analysis process part 227 of 4th Embodiment. It is a flowchart of operation | movement of the process of the song evaluation part 22 of 4th Embodiment. It is the schematic of the song evaluation system 1 which concerns on 5th Embodiment.

<First Embodiment>
FIG. 1 is a schematic diagram of a singing evaluation system 1 according to the first embodiment. As illustrated in FIG. 1, the singing evaluation system 1 includes a terminal device D1 used by the user U1 and a terminal device D2 used by the user U2. In 1st Embodiment, the structure which utilizes terminal device D1 as a song evaluation apparatus is illustrated. The user U1 sings toward the terminal device D1. The user U2 sings toward the terminal device D2.

  The terminal device D2 is a communication terminal such as a mobile phone or a smartphone, and includes a sound collection device 34 and a communication device 35. The sound collection device 34 collects the singing voice of the user U2 singing the music and generates the second singing signal V2. The communication device 35 is a communication device for communicating with the terminal device D1, and transmits the second singing signal V2 generated by the sound collection device 34 to the terminal device D1.

  The terminal device D1 of the first embodiment has a singing voice (first singing voice) in which the user U1 sings a song and a singing voice (second singing voice) in which the user U2 of the terminal device D2 sings the same piece of music. A computer system comprising an arithmetic processing unit 10, a storage unit 12, an input unit 13, a sound collection unit 14, a communication unit 15, a display unit 16, and a sound emission unit 18. It is realized with. For example, a portable communication terminal (mobile phone or smartphone) carried by a user and an information processing device such as a personal computer are used as the terminal device D1. The sound collection device 14 is a device (microphone) that collects ambient sounds. The sound collection device 14 according to the first embodiment collects a singing voice in which the user U1 of the terminal device D1 sang a song and generates a first singing signal V1. The input device 13 is a device (for example, a touch panel) that accepts an instruction from a user.

  The storage device 12 stores a program PGM executed by the arithmetic processing device 10 and various data used by the arithmetic processing device 10. A known recording medium such as a semiconductor recording medium or a magnetic recording medium or a combination of a plurality of types of recording media is arbitrarily employed as the storage device 12. Specifically, the storage device 12 stores music data L. The music data L includes accompaniment data B that defines the accompaniment sound of the music in time series, and lyric data Q that indicates lyrics. The accompaniment data B is a music file in a format such as MP3, for example. When the user selects a desired music piece, accompaniment data B of the music piece is reproduced. The display device 16 (for example, a liquid crystal display panel) displays an image instructed from the arithmetic processing device 10. For example, the lyrics data Q of the music selected by the user and the evaluation result (evaluation value S) of the singing voice are displayed on the display device 16. The communication device 15 is a communication device for communicating with the terminal device D2. Specifically, the communication device 15 wirelessly receives the second singing signal V2 transmitted by the communication device 35 of the terminal device D2. The communication between the communication device 15 and the terminal device D2 may be communication via a communication network 400 such as a mobile communication network or the Internet, or a short distance such as a wireless LAN or Bluetooth (registered trademark) standard. Wireless communication may be used.

  1 executes the program PGM stored in the storage device 12 to centrally control each element of the terminal device D1. Specifically, the arithmetic processing device 10 indicates the first singing signal V1 generated by the sound collecting device 14 of the terminal device D1 and the second singing signal V2 generated by the sound collecting device 34 of the terminal device D2. A plurality of functions (singing evaluation unit 22, reproduction processing unit 26, display processing unit 28) for evaluating a song are realized. A configuration in which each function of the arithmetic processing device 10 is distributed to a plurality of devices, or a configuration in which a dedicated electronic circuit (for example, DSP) realizes a part of the functions of the arithmetic processing device 10 may be employed.

  The reproduction processing unit 26 mixes the first singing signal V 1 generated by the sound collection device 14 and the accompaniment data B read from the storage device 12, converts it into an analog signal, and supplies it to the sound emission device 18. The sound emitting device 18 (for example, a speaker or headphones) emits sound corresponding to the signal supplied from the reproduction processing unit 26. That is, a mixed sound of the musical accompaniment sound represented by the accompaniment data B and the singing voice of the user U1 is emitted from the sound emitting device 18. The display processing unit 28 displays various images on the display device 16. Specifically, the display processing unit 28 causes the display device 16 to display the lyrics data Q included in the song selected as the singing target and the evaluation value S indicating the result of the singing evaluation generated by the singing evaluation unit 22. .

  The singing evaluation unit 22 analyzes the first singing signal V1 and the second singing signal V2, thereby harmonizing the pitch of the singing voice by the user U1 (singing pitch) and the singing pitch by the user U2. It is a means for evaluating the degree, and includes a pitch analysis unit 222, a section setting unit 224, a representative value calculation unit 226, and an evaluation unit 228.

  The pitch analysis unit 222 of the first embodiment sequentially analyzes the pitch PA of the first singing signal V1 generated by the sound collecting device 14 at predetermined intervals. In addition, the pitch analysis unit 222 of the first embodiment sequentially analyzes the pitch PB of the second singing signal V2 received by the communication device 15 every predetermined period.

  FIG. 2 is a graph of the pitch PA of the first singing signal V1 and the pitch PB of the second singing signal V2 analyzed by the pitch analysis unit 222. The pitch analysis unit 222 has a pitch PA of the first singing signal V1 and a pitch PB of the second singing signal V2 with a sufficiently short period (for example, every 10 ms) as compared with the time length of notes that can be assumed in the music. Are sequentially identified at predetermined intervals. A known pitch detection technique can be arbitrarily employed to specify the pitch of the singing voice.

  The section setting unit 224 in FIG. 1 sets a plurality of first pitch stable sections TSA in which the pitch PA of the first singing signal V1 analyzed by the pitch analysis section 222 is stabilized in parallel with the progress of the singing. Specifically, as illustrated in FIG. 2, the section setting unit 224 has a predetermined variation amount of the pitch PA (that is, the difference between the highest value and the lowest value) of the first singing signal V1. Is set as the first pitch stable section TSA (TSA1, TSA2, TSA3...). Similarly, the section setting unit 224 sets a plurality of second pitch stable sections TSB in which the pitch PB of the second singing signal V2 analyzed by the pitch analysis unit 222 is stabilized in parallel with the progress of the singing. Specifically, the section setting unit 224 uses the same method as the specification of the first pitch stable section TSA to change the pitch PB of the second singing signal V2 (that is, the highest and lowest pitches PB). A plurality of sections in which the difference between the two values is maintained within a predetermined range are set as second pitch stable sections TSB (TSB1, TSB2, TSB3...).

  In addition, the section setting unit 224 identifies an overlapping section RTS in which each of the plurality of first pitch stable sections TSA and each of the plurality of second pitch stable sections TSB overlap on the time axis. As illustrated in FIG. 2, the overlapping section RTS includes a plurality of first pitch stable sections TSA (TSA1, TSA2, TSA3...) And a plurality of second pitch stable sections TSB (TSB1, TSB2, TSB3... ) Are ranges that overlap on the time axis.

  In the representative value calculation unit 226 of FIG. 1, the overlapping section RTS set by the section setting unit 224, that is, each of the plurality of first pitch stable sections TSA and each of the plurality of second pitch stable sections TSB is timed. The representative value RPA of the pitch PA of the first singing signal V1 (the representative value of the pitch of the first singing voice) and the representative value RPB of the pitch PB of the second singing signal V2 in the overlapping section RTS overlapping on the axis (first) 2) (representative value of the pitch of the two singing voices). As illustrated in FIG. 2, the representative value calculation unit 226, for each overlapping section RTS, represents the representative value RPA (RPA1, RPA2, RPA3...) Of the pitch PA and the representative value RPB (RPB1, RPB2,. RPB3 ...) is calculated. Specifically, an average value of a plurality of pitches PA (PB) specified in each of the first pitch stable sections TSA (second pitch stable sections TSB) is calculated as a representative value RPA (RPB).

  In the evaluation unit 228 of FIG. 1, the representative value RPA of the pitch PA of the first singing signal V1 and the representative value RPB of the pitch PB of the second singing signal V2 calculated by the representative value calculating unit 226 for each overlapping section RTS are predetermined. The degree of harmony between the first singing voice and the second singing voice is evaluated according to whether or not the pitch relationship is satisfied. The evaluation unit 228 of the present embodiment pays attention to whether or not the first singing voice and the second singing voice have an audibly harmonized pitch relationship, and the first singing voice and the second singing voice, An evaluation value S that evaluates the degree of harmony is output.

  FIG. 4 is a flowchart of processing operations of the pitch analysis unit 222 and the section setting unit 224. For example, when the reproduction of the music data L is started, the process of FIG. 4 is started. FIG. 4 illustrates the processing for the first singing signal V1 by the pitch analysis unit 222 and the section setting unit 224. Since the process for the second singing signal V2 is the same as the process for the first singing signal V1, detailed description thereof is omitted.

  When the pitch analysis unit 222 specifies the pitch PA for one time point on the time axis (hereinafter referred to as “analysis point”) in the first singing signal V1, (SA1), the section setting unit 224 is illustrated in FIG. As described above, an analysis section TA having a predetermined length with the analysis point KA on the time axis on which the pitch PA is specified as an end point is set (SA2). The analysis interval TA is a time interval that is an object of analysis defined by the time window function, and is set to a time length (for example, 200 ms) that is sufficiently longer than the cycle (10 ms) in which the pitch analysis is performed. The Therefore, a plurality of pitches PA including the pitch PA newly specified for the analysis point KA and the pitch PA before the analysis point KA are included in the analysis section TA.

  The section setting unit 224 identifies a maximum value PA-MAX and a minimum value PA-MIN of a plurality of pitches PA in the analysis section TA, and a difference value R () between the maximum value PA-MAX and the minimum value PA-MIN. It is determined whether (absolute value) is below a predetermined threshold PATH (SA3). It can be evaluated that the pitch PA of the first singing signal V1 is more stable as the difference value R (that is, the distribution width of the pitch PA in the analysis section TA) is narrower. For example, the threshold value PATH can be set to 50 cent in the twelve average temperament.

  When the difference value R falls below the threshold PATH (SA3: YES), the section setting unit 224 includes the analysis section TA in the first pitch stable section TSA (SA4). In the analysis section TAn in FIG. 3, since the difference value R between the maximum value PA-MAX and the minimum value PA-MIN of the pitch PA is below the threshold value PATH, the analysis section TAn including the analysis point KAn is the first pitch stable section TSA. Is included. The section setting unit 224 stores the pitch PA at the analysis point KA in the storage device 12 (SA5).

  The interval setting unit 224 uses the analysis point KA of the pitch PA as an end point every time the pitch PA is specified by the pitch analysis unit 222 (SA1) until the music ends (SA9: NO). An analysis section TA is set (SA2), and it is determined whether or not the difference value R of the pitch PA in the analysis section TA is below the threshold PATH (SA3). That is, as understood from FIG. 3, the analysis interval TA is moved to the first sound while the analysis interval TA is sequentially moved on the time axis for every period (10 ms) in which the pitch analysis unit 222 specifies the pitch PA. It is determined whether or not it is included in the highly stable section TSA. Accordingly, the first pitch stable section TSA is sequentially extended on the time axis every time the pitch analysis unit 222 specifies the pitch PA within the range of the distribution width below the threshold PATH.

  On the other hand, when the difference value R of the pitch PA in the analysis section TA is equal to or greater than the threshold PATH (SA3: NO), the section setting unit 224 does not include the analysis point KA in the first pitch stable section TSA (SA6). ). The section setting unit 224 determines whether or not the analysis section TA immediately before the current analysis section TA exists in the first pitch stable section TSA (SA7). If the determination result is affirmative, the section setting unit 224 determines the end point (analysis point KA) of the immediately preceding analysis section TA as the end point of one first pitch stable section TSA (SA8). That is, the section setting unit 224 sequentially sets the first pitch stable section TSA in parallel with the progress of singing. According to the above configuration, it is possible to set only the first pitch stable section TSA with a small fluctuation amount of the pitch PA in the entire music as the object of the singing evaluation. When the first pitch stable section TSA and the second pitch stable section TSB are sequentially set by the above procedure, the storage device 12 stores the pitch PA included in the first pitch stable section TSA, The pitch PB included in the second pitch stable section TSB is sequentially accumulated. Next, as illustrated in FIG. 2, the section setting unit 224 identifies an overlapping section RTS in which the first pitch stable section TSA and the second pitch stable section TSB overlap on the time axis.

  Next, singing evaluation by the evaluation unit 228 will be described. FIG. 5 is a flowchart of a process (evaluation process) in which the evaluation unit 228 evaluates the degree of harmony between the pitch of the first singing voice and the pitch of the second singing voice. The section setting unit 224 sets the overlapping section RTSn (n = 1, 2, 3, 4,...) In the music, and the representative value calculating section 226 uses the representative value RPA and the pitch of the pitch PA in the overlapping section RTS. Each time the representative value RPB of PB is calculated, the evaluation process of FIG. 5 is executed.

  The evaluation unit 228 calculates a difference value C between the representative value RPA of the pitch PA and the representative value RPB of the pitch PB in the overlapping section RTS (SB1). Specifically, the difference value C between the representative value RPA and the representative value RPB of the overlapping section RTS newly set by the section setting unit 224 is calculated. When the difference value between the representative value RPA and the representative value RPB exceeds 1200 cents (one octave), a difference value C of 1200 cents or less is calculated by subtracting an integer multiple of 1200 cents from the difference.

  Assuming that the music is composed of twelve types of twelve scales (discrete pitches constituting the scale of the scale), the voice (first singing voice) that the user U1 sang the music and When the pitch of the user U2 is harmonized with the voice of the song sung (second song voice), the difference value between the pitch of the first song voice and the pitch of the second song voice is: It is expected to approximate or match an integer multiple of the interval between two adjacent musical notes (100 cents corresponding to a semitone). Therefore, when the difference value C between the representative value RPA and the representative value RPB of one overlapping section RTS corresponds to an integral multiple of 100 cents, in the overlapping section RTS, the singing pitch of the user U1 and the singing of the user U2 It can be evaluated that the pitch is in harmony. Therefore, the evaluation unit 228 of the first embodiment, when the difference value C between the representative value RPA and the representative value RPB in the overlapping section RTS corresponds to an integer multiple of 100 cent, the pitch of the first singing voice and the second singing voice. As a result, the evaluation value S is increased. On the other hand, when the difference value C deviates from an integer multiple of 100 cents, it is estimated that the pitch of the first singing voice and the pitch of the second singing voice are not in harmony. Therefore, when the difference value C deviates from an integer multiple of 100 cents, the evaluation unit 228 decreases the evaluation value S, assuming that the pitch of the first singing voice and the pitch of the second singing voice do not match.

  For example, when the difference value C between the representative value RPA and the representative value RPB is 300 cents (short 3 degrees) or 400 cent (long 3 degrees) which is an integral multiple of 100 cent, the pitch of the first singing voice and the second singing voice Harmonizes with the pitch and gives a pleasant impression. Further, when the difference value is 700 cent (completely 5 degrees), 1200 cent (1 octave, that is, completely 8 degrees), or 0 cent (completely 1 degree), that is, the pitch of the first singing voice and the pitch of the second singing voice. Are in harmony with each other, the pitch of the first singing voice and the pitch of the second singing voice are harmonized, giving an impression of being audibly audible. Considering the above circumstances, the difference value C between the pitch of the first singing voice and the pitch of the second singing voice is 300 cent and 400 cent, or the pitch of the first singing voice and the second singing voice When the pitch is in the form of a consonant tone (that is, the difference value C is 700 cent, 1200 cent, 0 cent), the evaluation value S may be increased.

  The evaluation unit 228 determines whether or not the difference value C between the representative value RPA and the representative value RPB in the overlapping section RTS approximates (including a match) to an integer multiple of 100 cent (SB2). Specifically, it is determined whether or not the difference value C is within a predetermined range (eg, ± 10%) including an integer multiple of 100 cent. When the difference value C approximates an integer multiple of 100 cent (SB2: YES), the evaluation unit 228 evaluates the overlapping section RTSn from the start of the music until the overlapping section RTSn-1 (score: An added value of (Score) Sn-1 and a predetermined value Δ (positive integer) is set as an evaluation value (Score) Sn acquired until the overlapping section RTSn (Sn = Sn-1 + Δ) (SB3).

  On the other hand, when the difference value C deviates from an integer multiple of 100 cent (SB2: NO), the evaluation value Sn-1 acquired from the start of the music to the overlapping section RTSn-1 and the predetermined value Δ (positive) The subtraction value with (integer) is set as the evaluation value Sn acquired up to the n-th overlapping section RTSn (Sn = Sn-1−Δ) (SB4). The evaluation unit 228 outputs the evaluation value Sn set in SB3 and SB4 to the display processing unit 28 of the terminal device D1. The display processing unit 28 displays the evaluation value Sn of the singing on the display device 16 (SB5). That is, the evaluation value Sn changes every moment in parallel with the reproduction of the music (song progress). Further, the evaluation value Sn calculated by the evaluation unit 228 is transmitted by the communication device 15 to the terminal device D2. In the terminal device D2, the evaluation value Sn is displayed as in the terminal device D1.

  In addition, the calculation method of the evaluation value S according to the difference value C is appropriately changed. For example, the evaluation value S can be set so that the difference value C has a relationship as shown in FIG. As illustrated in FIG. 6, the distribution of the evaluation value S is set in advance for each of the plurality of ranges of the difference value C, and according to the difference value C between the representative value RPA and the representative value RPB among the plurality of evaluation values S. The evaluated value S is selected. The difference value C of the minor third (3 semitones) and the major third (four semitones) takes into account the tendency that the degree of harmony perceived by the listener is high, and the difference value C corresponding to the minor third (C = 300) A high evaluation score S is set for the difference value C (C = 400) corresponding to 3 degrees.

  As described above, in the configuration of the first embodiment, the representative value RPA and the representative value RPB in the overlapping section RTS in which the first pitch stable section TSA and the second pitch stable section TSB overlap on the time axis are Depending on whether or not the difference value C is in a predetermined pitch relationship (specifically, an integral multiple of the interval between each tone adjacent to each other under a specific temperament such as twelve equal temperament). Singing is evaluated. Therefore, even when there is no musical score data for each of the plurality of singing parts, it is possible to appropriately evaluate the singing.

  In the first embodiment, the first pitch stable section TSA in which the pitch PA is stable in the first singing signal V1 and the second pitch stable section TSB in which the pitch PB is stable in the second singing signal V2 overlap. The degree of harmony between the pitch of the first singing voice and the pitch of the second singing voice is evaluated according to the representative value RPA and the representative value RPB in each overlapping section RTS. Here, the composition which evaluates the singing over the whole music including the section where the pitch continuously changes between notes and the section where the pitch fluctuates as a singing expression (for example, vibrato section) (hereinafter referred to as “comparative”) ), Since the singing of the entire music including the section in which the pitch is unstablely shifted is the target of the singing evaluation, the appropriate evaluation may not always be performed. On the other hand, according to the configuration of the first embodiment, the representative value RPA and the representative value RPB calculated in the overlapping section RTS in which the first pitch stable section TSA and the second pitch stable section TSB in which the pitch is stabilized overlap. Accordingly, since the degree of harmony between the pitch of the first singing voice and the pitch of the second singing voice is evaluated, an appropriate evaluation using the pitch of the stable section as compared with the comparative configuration. Can be realized.

Second Embodiment
In the first embodiment, the second song signal V2 is transmitted from the terminal device D2 to the terminal device D1, and the pitch PB of the second song signal V2 is analyzed by the terminal device D1. In the second embodiment, the pitch PB of the second singing signal V2 analyzed by the terminal device D2 is sequentially transmitted from the terminal device D2 to the terminal device D1.

  FIG. 7 is a schematic diagram of the singing evaluation system 1 of the second embodiment. In the second embodiment, a pitch analysis unit 36 is added to the terminal device D2 used by the user U2. The pitch analysis unit 36 sequentially analyzes the pitch PB of the second singing signal V2 generated by the sound collecting device 34. The analysis method of the pitch PB by the pitch analysis unit 36 is the same as the analysis method of the pitch PB by the pitch analysis unit 222 of the first embodiment. The communication apparatus 35 of 2nd Embodiment transmits the pitch PB of the 2nd song signal V2 analyzed by the pitch analysis part 36 to the terminal device D1, as illustrated by FIG. The communication device 15 of the terminal device D1 sequentially receives the pitch PB transmitted from the terminal device D2.

  The pitch analysis unit 222 of the terminal device D1 sequentially analyzes the pitch PA of the first singing signal V1 and notifies the section setting unit 224 of the analyzed pitch PA. The section setting unit 224 sets a plurality of first pitch stable sections TSA in which the pitch PA analyzed by the pitch analysis unit 222 is stable, and the second pitch in which the pitch PB received by the communication device 15 is stable. A plurality of stable sections TSB are set. Since the subsequent description is the same as that of the first embodiment, the description is omitted.

  As understood from the above description, the same effects as those of the first embodiment are realized by the configuration of the second embodiment. Further, according to the configuration of the second embodiment, the pitch analysis unit 222 does not need to analyze the pitch PB of the second singing signal V2, and thus it is possible to reduce the load of calculation processing by the terminal device D1. There is also an advantage of becoming.

<Third Embodiment>
In the second embodiment, the section setting unit 224 acquires the pitch PA from the pitch analysis unit 222 in the terminal device D1, while acquiring the pitch PB received by the communication device 15. Since the pitch PB is transmitted from the terminal device D2 to the terminal device D1, even when the user U1 and the user U2 sing the same music in synchronism with each other, the section setting unit 224 has a specific time point in the music. The time point when the pitch PB is acquired may be delayed from the time point when the section setting unit 224 acquires the pitch PA at that time, for example, by the communication delay between the terminal device D1 and the terminal device D2. Specifically, as illustrated in FIG. 8, the time axis at which the section setting unit 224 acquires the pitch PA at each time point in the music (for example, the analysis point KAn [n = 1, 2, 3...] And the time point when the section setting unit 224 acquires the pitch PB at the same time in the music (for example, the position on the time axis of the analysis point KBn [n = 1,2,3...]). May cause a delay (Δt) required for communication. Therefore, in the third embodiment, a process for compensating for the delay time is executed.

  The configuration of the singing evaluation system 1 of the third embodiment is the same as that of the second embodiment. When the user U1 selects a song to be sung, a playback instruction for the song is transmitted from the terminal device D1 to the terminal device D2. Since the terminal device D1 and the terminal device D2 have the same music start time, the user U1 and the user U2 sing the same music synchronously. That is, the time point when the terminal device D1 analyzes the pitch (analysis point KA) coincides with the time point when the terminal device D2 analyzes the pitch (analysis point KB). In 3rd Embodiment, the pitch analysis part 222 of the terminal device D1 is the information (following description) which shows the position on the time-axis of the said analysis point KA whenever it analyzes the pitch PA among the 1st song signals V1. Then, it is referred to as “first time information”) to the section setting unit 224. Further, whenever the pitch analysis unit 36 of the terminal device D2 analyzes the pitch PB of the second singing signal V2, information indicating the position of the analyzed pitch PB and the analysis point KB on the time axis (hereinafter referred to as the pitch PB). Is referred to as “second time information”) via the communication device 35 to the terminal device D1. The communication device 15 of the terminal device D1 receives the pitch PB and the second time information transmitted by the communication device 35. As an example of the time information, for example, an elapsed time and time from the start of music are exemplified.

  FIG. 9 is an explanatory diagram of processing by the section setting unit 224 of the third embodiment. The section setting unit 224 of the third embodiment is notified by the pitch analysis unit 36 via the first time information notified by the pitch analysis unit 222 and communication between the communication device 35 and the communication device 15. By associating the second time information with each other, a plurality of analysis points KA (first analysis point KA) indicated by the first time information and a plurality of analysis points KB (second analysis point KB) indicated by the second time information. That correspond to each other (corresponding to the same point in the music) are matched with each other on the time axis.

  The section setting unit 224 matches the analysis point KA and the analysis point KB corresponding to each other on the time axis, and then performs the same procedure as in the first embodiment by using the same procedure as in the first embodiment. A two pitch stable section TSB is set, and an overlapping section RTS in which the first pitch stable section TSA and the second pitch stable section TSB overlap on the time axis is set.

  As described above, the same effects as those of the above-described embodiments can be obtained by the configuration of the third embodiment. In the configuration of the third embodiment, the first analysis point KA at which the pitch PA of the first singing signal V1 is analyzed and the second analysis point KB at which the pitch PB of the second singing signal V2 is analyzed are time axes. Match above. Therefore, even when a delay occurs in communication between the terminal device D1 and the terminal device D2, it is possible to appropriately evaluate the degree of harmony between the first singing voice and the second singing voice.

<Fourth embodiment>
In each above-mentioned form, according to whether the 1st singing voice and the 2nd singing voice have a predetermined pitch relation, the pitch of the 1st singing voice and the pitch of the 2nd singing voice are in harmony. The degree was evaluated. In the fourth embodiment, the degree of similarity in singing pitch between the first singing voice and the second singing voice is evaluated.

  FIG. 10 is a schematic diagram of the singing evaluation system 1 of the fourth embodiment. As understood from FIG. 10, in the configuration of the terminal device D1 of the fourth embodiment, a pitch distribution generation unit 225 and an analysis processing unit 227 are added to the configuration of the terminal device D1 of the first embodiment. . Since the function and operation of the pitch analysis unit 222 are the same as those in the first embodiment, a detailed description thereof will be omitted.

  The section setting unit 224 of the fourth embodiment includes a first pitch stable section TSA in which the pitch PA of the first singing signal V1 is stable and a second pitch stable section in which the pitch PB of the second singing signal V2 is stable. A plurality of TSBs are set. The section setting unit 224 of the fourth embodiment does not set an overlapping section RTS in which the first pitch stable section TSA and the second pitch stable section TSB overlap. When the setting of the first pitch stable section TSA and the second pitch stable section TSB is completed over the entire musical piece, the representative value calculating unit 226 uses the same method as in the first embodiment to perform a plurality of first pitch stable sections. A representative value RPA of the pitch PA in each of the sections TSA and a representative value RPB of the pitch PB in each of the plurality of second pitch stable sections TSB are calculated. The representative value calculation unit 226 sequentially stores the calculated representative value RPA and the representative value RPB in the storage device 12. Therefore, the number of representative values RPA corresponding to the total number of first pitch stable sections TSA and the number of representative values RPB corresponding to the total number of second pitch stable sections TSB in the music are stored in the storage device 12. The pitch distribution generation unit 225 generates a first frequency distribution indicating the frequency distribution of the representative value RPA over the plurality of first pitch stable sections TSA, while representing the representative value RPB over the plurality of second pitch stable sections TSB. A second frequency distribution indicating the frequency distribution is generated.

  FIG. 11 is an explanatory diagram for the processing of the pitch distribution generation unit 225. The pitch distribution generation unit 225 reads the representative value RPA included in the plurality of first pitch stable sections TSA from the storage device 12, and creates a first frequency distribution (histogram) HA of the representative value RPA. The frequency distribution HA is a frequency distribution of the representative value RPA in each of a plurality of classes in which the numerical value range of the representative value RPA is sufficiently finely divided as compared with the pitch difference (100 cent) of each tone adjacent to each other. . Similarly to the generation of the first frequency distribution HA, the pitch distribution generation unit 225 reads the representative value RPB included in the plurality of second pitch stable sections TSB from the storage device 12, and the frequency of the representative value RPB. A second frequency distribution HB indicating the distribution is created.

  The analysis processing unit 227 in FIG. 10 sums up each of a plurality of distributions that are distributions of a predetermined range including the pitches of the scales in the first frequency distribution HA generated by the pitch distribution generation unit 225. Specifically, the analysis processing unit 227 sums up the distribution of each unit range Tu over a plurality of unit ranges Tu (Tu1, Tu2, Tu3, Tu4, Tu5, Tu6) obtained by dividing the first frequency distribution HA for each scale sound. Then, one first evaluation distribution QA is created and a representative value of the first evaluation distribution QA is calculated. In the fourth embodiment, the average value A1 of the first evaluation distribution QA is calculated as a representative value of the first evaluation distribution QA. The unit range Tu corresponding to one arbitrary scale sound is a range of ± 50 cent centered on the pitch of the scale sound. Similar to the generation of the first evaluation distribution QA, the analysis processing unit 227 totals the distributions of the unit ranges Tu over a plurality of unit ranges Tu obtained by dividing the second frequency distribution HB for each scale sound, thereby generating one second An evaluation distribution QB is created and a representative value of the second evaluation distribution QB (average value A2 of the second evaluation distribution QB) is calculated.

  FIG. 12 is an explanatory diagram of the first evaluation distribution QA created by the analysis processing unit 227. The analysis processing unit 227 has each unit within a range of numerical value α (−50 ≦ α ≦ + 50) over ± 50 cents, where 0 is the pitch of the scale tone (1200cent, 1300cent, 1400cent, 1500cent, 1600cent, 1700cent in FIG. 11). The frequencies of the distribution over a plurality of unit ranges Tu are summed for each numerical value α after the distribution of the range Tu is made to correspond. The analysis processing unit 227 then adds each total frequency for each numerical value α within a range of ± 50 cents so that an integrated value (total area under the distribution) becomes a predetermined value (for example, 1). Is normalized to generate the first evaluation distribution QA. In other words, the analysis processing unit 227 divides the frequency distribution H (HA, HB) into a plurality of unit ranges Tu centered on the pitch of the scale sound, and overlaps each unit range Tu with each other to determine each unit range Tu. The first evaluation distribution QA is generated by summing the frequency of distribution for each numerical value α (pitch) over a plurality of unit ranges Tu (and normalizing the total frequency). The function G (α) in FIG. 12 is a function representing the first evaluation distribution QA after normalization using the numerical value α as a variable. The analysis processing unit 227 calculates an average value A1 of the first evaluation distribution QA. Further, the analysis processing unit 227 generates the second evaluation distribution QB by the same method as the first evaluation distribution QA and calculates the average value A2.

  As understood from the above description, the first evaluation distribution QA expresses the tendency of the distribution of the representative value RPA of the pitch PA for each of a plurality of scale sounds. Specifically, when the average value A1 of the first evaluation distribution QA is large, it can be estimated that the user U1 tends to sing a higher pitch with respect to the original pitch of the scale sound. Similarly, the second evaluation distribution QB expresses the tendency of the distribution of the representative value RPB of the pitch PB with respect to the pitch of each scale sound. Specifically, when the average value A2 of the second evaluation distribution QB is large, it can be estimated that the user U2 tends to sing a higher pitch with respect to the original pitch of the scale sound.

  The evaluation unit 228 sings the singing sound of the first singing voice and the second singing voice according to the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB calculated by the analysis processing unit 227. Assess the degree of high similarity. When the singing pitch is approximated between the first singing voice and the second singing voice (for example, both the user U1 and the user U2 tend to pronounce at a higher pitch than the pitch of the scale sound. The average value A1 and the average value A2 are approximated to each other. On the other hand, when the singing pitch deviates between the first singing voice and the second singing voice (for example, the user U1 tends to pronounce at a high pitch with respect to the scale tone, while the user U2 It is estimated that the average value A1 and the average value A2 are different from each other. Considering the above tendency, the evaluation unit 228 of the fourth embodiment performs the first singing throughout the entire song as the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB approximate. The evaluation value S is set to a large numerical value because the singing pitches of the voice and the second singing voice are similar. On the other hand, as the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB deviate, the singing pitches of the first singing voice and the second singing voice diverge throughout the entire music. Set the value S to a small number.

  FIG. 13 is a flowchart of an operation process performed by the song evaluation unit 22 (pitch distribution generation unit 225, analysis processing unit 227, and evaluation unit 228) according to the fourth embodiment. The representative value calculation unit 226 selects the representative value RPA of the pitch PA of each first pitch stable section TSAn (n = 1, 2, 3, 4,...) And each second pitch stable section TSBn (n = 1). , 2, 3, 4,...) And the representative value RPB of the pitch PB are calculated over the entire music, the evaluation process of FIG. 13 is executed.

  The pitch distribution generation unit 225 creates the first frequency distribution HA of the representative value RPA over the plurality of first pitch stable sections TSA in the music and the plurality of second pitch stable sections TSB in the music. A second frequency distribution HB of the representative value RPB is created (SD1). The analysis processing unit 227 creates the first evaluation distribution QA from a plurality of unit ranges Tu obtained by dividing the first frequency distribution HA generated by the pitch distribution generation unit 225 for each scale sound, and the second frequency distribution HB. A second evaluation distribution QB is created from a plurality of unit ranges Tu divided for each scale sound (SD2). The analysis processing unit 227 calculates the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB (SD3).

  The evaluation unit 228 sings the singing sound of the first singing voice and the second singing voice according to the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB calculated by the analysis processing unit 227. Assess the degree of high similarity (SD4). For example, as the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB approximate, the singing pitch of the first singing voice and the singing pitch of the second singing voice are similar throughout the music. As a result, the evaluation value S is set to a large numerical value. On the other hand, as the average value A1 of the first evaluation distribution QA deviates from the average value A2 of the second evaluation distribution QB, the singing pitch of the first singing voice and the singing pitch of the second singing voice are more different throughout the music. Then, it evaluates and sets evaluation value S to a small numerical value. The display processing unit 28 displays the evaluation value S on the display device 16 (SD5).

  As understood from the above description, in the fourth embodiment, the average value A1 (first evaluation distribution) of the first evaluation distribution QA generated from the first frequency distribution HA of the representative value RPA in the first pitch stable section TSA. According to the average value A2 of the second evaluation distribution QB generated from the second frequency distribution HB of the representative value RPB in the second pitch stable section TSB (representative value of the second evaluation distribution QB). Thus, the degree of similarity of the singing pitch between the first singing voice and the second singing voice is evaluated. Therefore, similar to the effect of the first embodiment, the configuration of the fourth embodiment is similar to the singing pitch between the user U1 and the user U2 without requiring the score data for each singing part of the music. It is possible to appropriately evaluate the degree of. In the fourth embodiment, since the representative value RPA in the first pitch stable section TSA and the representative value RPB in the second pitch stable section TSB are used as objects for singing evaluation, the pitch transition of the music is changed. It is possible to appropriately evaluate the singing pitch as compared with the configuration in which the singing pitch over the entire section including the unstable section is the evaluation target.

<Fifth Embodiment>
In each above-mentioned form, composition which utilizes terminal unit D1 which user U1 uses as a song evaluation device was illustrated. In 5th Embodiment, the structure which utilizes the management apparatus 500 which performs communication between terminal device D (D1, D2) as a song evaluation apparatus is illustrated.

  FIG. 14 is a schematic diagram of the singing evaluation system 1 of the fifth embodiment. As illustrated in FIG. 14, the singing evaluation system 1 according to the fifth embodiment includes a management device 500, a terminal device D1 used by the user U1, and a terminal device D2 used by the user U2. The The configurations of the terminal device D1 and the terminal device D2 are the same as the terminal device D2 illustrated in FIG. That is, the pitch PA of the first singing signal V1 indicating the first singing voice of the user U1 is sequentially transmitted from the terminal device D1 to the management apparatus 500, and the second singing signal V2 indicating the second singing voice of the user U2. Are sequentially transmitted from the terminal device D2 to the management device 500.

  The management device 500 evaluates the degree of pitch harmony between the user U1 singing voice and the user U2 singing voice, and uses the evaluation value S indicating the evaluation result with the terminal device D1 of the user U1. And a singing evaluation unit 52 and a communication device 54. The communication device 54 communicates with each of the terminal device D1 and the terminal device D2. Specifically, the communication device 54 sequentially receives the pitch PA transmitted by the terminal device D1 and the pitch PB transmitted by the terminal device D2. The singing evaluation unit 52 evaluates the degree of harmony between the pitch of the singing voice by the user U1 and the pitch of the singing voice by the user U2. Specifically, the singing evaluation unit 52 of the fifth embodiment includes a section setting unit 224, a representative value calculation unit 226, and an evaluation unit 228, similarly to the singing evaluation unit 22 of the second embodiment. Since the processing of the section setting unit 224, the representative value calculation unit 226, and the evaluation unit 228 is the same as that of the second embodiment, detailed description thereof is omitted. The evaluation value S calculated by the evaluation unit 228 is transmitted from the communication device 54 to the terminal device D1 and the terminal device D2, and displayed on each display device.

  In the fifth embodiment, the same effects as those of the above-described embodiments are realized. In the configuration of the fifth embodiment, the management device 500 evaluates the degree of harmony between the pitch of the first singing voice of the user U1 and the pitch of the second singing voice of the user U2. Therefore, the terminal device D1 has an advantage that the singing evaluation unit 22 does not need to be mounted. 14 illustrates the configuration in which the terminal device D1 transmits the pitch PA of the first singing signal V1 and the terminal device D2 transmits the pitch PB of the second singing signal V2, the terminal device D1 is the first. A configuration in which the terminal device D2 transmits the second singing signal V2 to the management device 500 while transmitting the one singing signal V1 to the management device 500 may be employed. Specifically, a pitch analysis unit 222 that analyzes the pitch PA of the first song signal V1 and the pitch PB of the second song signal V2 is added to the song evaluation unit 52 of the management device 500.

<Modification>
Each of the above-described embodiments can be variously modified. Specific modifications are exemplified below. Two or more modes arbitrarily selected from the following examples can be appropriately combined.

  (1) In the fourth embodiment, the second singing signal V2 is transmitted from the terminal device D2 to the terminal device D1, and the pitch PB of the second singing signal V2 is analyzed by the terminal device D1, but the second embodiment and the second embodiment Similarly to the third embodiment, the pitch PB of the second singing signal V2 analyzed by the terminal device D2 may be sequentially transmitted from the terminal device D2 to the terminal device D1.

  (2) In the fourth embodiment, the average value A1 of the first evaluation distribution QA and the average value A2 of the second evaluation distribution QB are exemplified as representative values of each distribution. However, the first evaluation distribution QA and the second evaluation distribution The representative value representing each characteristic of QB is not limited to the average value. For example, for each of the first evaluation distribution QA and the second evaluation distribution QB, various statistics such as an index value other than the average value (median value, mode value, etc.), variance value, and second moment are used as representative values. Can be calculated. For example, it can be evaluated that the singing pitches of the first singing voice and the second singing voice are more similar as the variance value and the second moment are approximated between the first evaluation distribution QA and the second evaluation distribution QB. is there.

  (3) In the fifth embodiment, whether or not the representative value RPA of the pitch PA of the first singing signal V1 and the representative value RPB of the pitch PB of the second singing signal V2 have a predetermined pitch relationship. Accordingly, the degree of harmony between the first singing voice and the second singing voice was evaluated. In addition to the above examples, a pitch distribution generation unit 225 and an analysis processing unit 227 are added to the management device 500 of the fifth embodiment, and the degree of similarity in singing pitch between the first singing voice and the second singing voice. May be evaluated.

  (4) In each of the above-described embodiments, the configuration in which the music data L including the accompaniment data B and the lyrics data Q is stored in the storage device 12 of the terminal device D1, but the illustration is omitted for the terminal device D1. It is good also as a structure which receives the music data L by performing communication via a communication network (for example, mobile communication network or the internet) 400 between music provision servers.

  (5) In each above-mentioned form, the composition which displays evaluation value S as an evaluation result of singing was illustrated. Instead of the evaluation value S, text information such as “OK” or “NG” may be displayed as the evaluation information.

  (6) In each of the above-described embodiments, the user U1 sings toward the terminal device D1, while the user U2 sings toward the terminal device D2, but the user U1 and the user U2 It is good also as a structure which sings toward one terminal device. Specifically, for example, a plurality of sound collecting devices (14A and 14B) are added to the terminal device D1, the user U1 sings toward the sound collecting device 14A, and the user U2 faces the sound collecting device 14B. You may sing. With the above configuration, the same effect as in the first embodiment is realized.

DESCRIPTION OF SYMBOLS 1 ... Singing evaluation system, 10 ... Arithmetic processing device, 12 ... Memory | storage device, 14 ... Sound collecting device, 15 ... Communication device, 16 ... Display device, 18 ... Sound emitting device, 22 ... Singing Evaluation unit, 26... Reproduction processing unit, 28... Display processing unit, 34... Sound collection device, 35 .. communication device, 36 .. pitch analysis unit, 52. , 222 …… Pitch analysis unit, 224 …… Section setting unit, 225 …… Pitch distribution generation unit, 226 …… Representative value calculation unit, 227 …… Analysis processing unit, 228 …… Evaluation unit, 500 …… Management Device, B ... Accompaniment data, Q ... Lyric data, L ... Music data, TSA ... First pitch stable section, TSB ... Second pitch stable section, RTS ... Overlapping section, Tu ... Unit range.

Claims (6)

A plurality of first pitch stable sections in which the pitch of the first singing signal indicating one singing voice of the music is stable, and a plurality of first pitches in which the pitch of the second singing signal indicating the other singing voice of the music is stable. A section setting unit for setting a two-pitch stable section;
For each overlapping section in which each of the plurality of first pitch stable sections and each of the plurality of second pitch stable sections set by the section setting unit overlap on the time axis, the pitch of the first singing signal is set. A representative value calculating unit for calculating a representative value and a representative value of the pitch of the second singing signal;
Depending on whether or not the representative value of the pitch of the first singing signal and the representative value of the pitch of the second singing signal in the respective overlapping sections are in a predetermined pitch relationship, the one singing voice and A singing evaluation apparatus comprising: an evaluation unit that evaluates a degree of harmony with the other singing voice. A pitch analyzer for sequentially analyzing the pitch of the first singing signal;
A receiver for receiving the pitches sequentially analyzed from the second singing signal;
The section setting section includes a plurality of first pitch stable sections in which the pitch analyzed by the pitch analysis section of the first singing signal is stabilized, and a sound received by the receiving section of the second singing signal. The singing evaluation apparatus according to claim 1, wherein a plurality of second pitch stable sections in which the height is stabilized are set. The pitch analysis unit sequentially analyzes the pitch of the first singing signal for each first analysis point on the time axis,
The receiving unit receives the pitch of the second singing signal sequentially analyzed for each second analysis point on the time axis,
The section setting unit matches the mutually corresponding ones on the time axis between the plurality of first analysis points and the plurality of second analysis points, and then sets the first pitch stable section and The singing evaluation apparatus according to claim 2, wherein an overlapping section in which the second pitch stable section overlaps is set. A plurality of first pitch stable sections in which the pitch of the first singing signal indicating one singing voice of the music is stable, and a plurality of first pitches in which the pitch of the second singing signal indicating the other singing voice of the music is stable. A section setting unit for setting a two-pitch stable section;
A representative value calculation unit for calculating a representative value of the pitch in each of the plurality of second pitch stable sections, while calculating a representative value of the pitch in each of the plurality of first pitch stable sections;
A first frequency distribution indicating a frequency distribution of representative values of pitches over the plurality of first pitch stable sections is generated, while a frequency distribution of representative values of pitches over the plurality of second pitch stable sections is generated. A pitch distribution generation unit for generating the second frequency distribution shown;
The first frequency distribution is divided into a plurality of unit ranges centered on the pitch of the scale sound, and each unit range is overlapped with each other, and the frequency of the distribution of each unit range is divided for each pitch over a plurality of unit ranges. The first evaluation distribution is created by summing up, while the second frequency distribution is divided into a plurality of unit ranges centered on the pitch of the scale sound, and the unit ranges are overlapped with each other. An analysis processing unit that creates a second evaluation distribution by summing the frequency of distribution for each pitch over a plurality of unit ranges;
A singing evaluation apparatus comprising: an evaluation unit that evaluates the degree of similarity in singing pitch between the one singing voice and the other singing voice based on the first evaluation distribution and the second evaluation distribution. A plurality of first pitch stable sections in which the pitch of the first singing signal indicating one singing voice of the music is stable, and a plurality of first pitches in which the pitch of the second singing signal indicating the other singing voice of the music is stable. A section setting unit for setting a two-pitch stable section,
For each overlapping section in which each of the plurality of first pitch stable sections and each of the plurality of second pitch stable sections set by the section setting unit overlap on the time axis, the pitch of the first singing signal is set. A representative value calculating unit for calculating a representative value and a representative value of the pitch of the second singing signal;
Depending on whether or not the representative value of the pitch of the first singing signal and the representative value of the pitch of the second singing signal in the respective overlapping sections are in a predetermined pitch relationship, the one singing voice and A program that causes a computer to function as an evaluation unit that evaluates the degree of harmony with the other singing voice. A plurality of first pitch stable sections in which the pitch of the first singing signal indicating one singing voice of the music is stable, and a plurality of first pitches in which the pitch of the second singing signal indicating the other singing voice of the music is stable. A section setting unit for setting a two-pitch stable section,
A representative value calculator for calculating a representative value of the pitch in each of the plurality of first stable pitch sections, and calculating a representative value of the pitch in each of the plurality of second stable pitch sections;
A first frequency distribution indicating a frequency distribution of representative values of pitches over the plurality of first pitch stable sections is generated, while a frequency distribution of representative values of pitches over the plurality of second pitch stable sections is generated. A pitch distribution generation unit for generating the second frequency distribution shown in FIG.
The first frequency distribution is divided into a plurality of unit ranges centered on the pitch of the scale sound, and each unit range is overlapped with each other, and the frequency of the distribution of each unit range is divided for each pitch over a plurality of unit ranges. The first evaluation distribution is created by summing up, while the second frequency distribution is divided into a plurality of unit ranges centered on the pitch of the scale sound, and the unit ranges are overlapped with each other. An analysis processing unit that creates a second evaluation distribution by summing the frequency of the distribution for each pitch over a plurality of unit ranges;
A program that causes a computer to function as an evaluation unit that evaluates the degree of similarity in singing pitch between the one singing voice and the other singing voice based on the first evaluation distribution and the second evaluation distribution.

Images