JP2011197564A - Electronic music device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a harmony voice signal that hardly causes a user to have unnatural noise feeling even in a section where it is difficult to detect a pitch of an input speech signal.SOLUTION: When a pitch of an input voice signal can be detected, the detected pitch is held and input pitch (pitch corresponding to sound name) is specified based on the pitch. When a pitch cannot be detected, the input pitch is specified based on the pitch held immediately before it. By specifying the input pitch based on the pitch held immediately before it, the original pitch to be pitch shifted is made clear and the pitch shift amount can be calculated. Thus, without preparing special data, a harmony voice signal where the discontinuity with the pitch in consecutive vowel sections is small, the continuity of timbre is kept, and the user hardly has unnatural noise feeling in a section where the pitch of the input voice signal cannot be detected.

Description

  The present invention relates to an electronic music apparatus and program for generating one or more harmony audio signals by pitch-shifting an input audio signal. In particular, the present invention relates to a technique for generating one or a plurality of harmony audio signals that hardly cause unnatural noise to the user even in a pitch non-detection section in which it is difficult to detect the pitch of an input audio signal.

  Conventionally, based on an input sound signal such as a musical instrument performance sound or a sound input by a user via a microphone or the like, the input sound signal has a pitch up and down by a predetermined pitch such as 3 degrees or 5 degrees. 2. Description of the Related Art There are known electronic music apparatuses and programs that automatically generate one or more separated harmony audio signals and output them together with the input audio signals. A conventionally known device calculates a pitch corresponding to either a fundamental frequency, that is, a musical pitch name for each predetermined section (or a predetermined period) based on frequency information (pitch) obtained by frequency analysis of an input audio signal. The input voice signal (more specifically, the window function corresponding to the specified pitch is cut out and stored in accordance with a predetermined pitch shift amount determined according to the pitch of the specified input voice signal. By pitch-shifting one period of waveform element data), one or more harmony audio signals having a predetermined pitch (here, the pitch corresponding to one of the musical pitch names) are separately added as additional sounds. It is designed to generate.

  By the way, vowels that cannot detect the input speech signal as a vowel with a clear harmonic structure, such as when the input speech signal transitions from a consonant interval consisting of a non-periodic waveform with a sense of pitch to a vowel interval consisting of a periodic waveform. In the non-detection section (the consonant section), it is difficult to specify the pitch (pitch corresponding to one of the pitch names) because the pitch of the input voice signal cannot be detected correctly (clearly). . If the pitch of the input audio signal cannot be specified, it is impossible to generate a harmony audio signal by pitch-shifting the input audio signal as described above, so the pitch of the input audio signal is detected correctly (clearly). For a vowel non-detection interval (also referred to as a pitch non-detection interval) that is difficult to do, the input audio signal must be output as it is without being pitch-shifted as a harmony audio signal.

  However, if the input voice signal is output without being pitch-shifted as a harmony voice signal in the vowel non-detection section, in the case of a consonant having a sense of pitch even if it is a consonant that is a non-periodic waveform, the vowel sections that follow each other As a result, the harmony voice signal in the vowel non-detection section is regarded as unnatural noise. It will be perceived by the user, which is inconvenient. Therefore, Patent Document 1 shown below solves the above-mentioned inconvenience by determining the pitch of the harmony voice signal in the vowel non-detection section using melody sequence data to be sung by a separately prepared singer. An apparatus for doing so is disclosed.

Japanese Patent No. 3173310

  As described above, in the conventional apparatus described in Patent Document 1, special data such as sequence data is used to generate a harmony voice signal in a vowel non-detection section in which it is difficult to detect the pitch of the input voice signal. Such special data must be prepared in advance. However, there are a wide variety of audio signals that can be input by a singer (user) via a microphone or the like, and it is impractical to prepare the special data in advance for all of them. However, even if the special data can be prepared in advance, it takes a lot of labor and storage capacity to create and store them, which is inconvenient and costly.

  The present invention has been made in view of the above-described points. A pitch non-detection section (specifically, as described above) is difficult to detect the pitch of an input audio signal without preparing special data such as sequence data. An object of the present invention is to provide an electronic music apparatus and program that can generate a harmony voice signal that does not cause an unnatural noise to the user even for a consonant section that is a vowel non-detection section. .

  An electronic music apparatus according to claim 1 of the present invention is an electronic music apparatus that generates one or a plurality of harmony audio signals whose pitches are controlled in accordance with pitch fluctuations of an input audio signal. Input pitch specification that specifies one of the pitches corresponding to the pitch name for each predetermined section of the input voice signal in accordance with the input means to input and analyzing the input voice signal to detect the pitch Means, a harmony pitch determining means for determining pitches of one or more harmony sounds according to the specified pitches, and the specified pitches and the pitches of the determined one or more harmony sounds, Shift amount calculating means for calculating one or a plurality of pitch shift amounts, and performing the pitch shift of the input audio signal for each of the predetermined sections based on the calculated one or more pitch shift amounts; And a musical tone generating means for generating one or more harmony voice signals controlled to one or more pitches, and the input pitch specifying means is capable of detecting the pitch of the input voice signals. While holding the detected pitch and specifying the pitch of the input voice signal based on the pitch, if the pitch of the input voice signal cannot be detected, the pitch is determined based on the pitch held immediately before. The pitch of the input audio signal is specified.

  According to the present invention, when the pitch of the input voice signal can be detected, the detected pitch is held and the pitch of the input voice signal (one of the pitches corresponding to the pitch name) is based on the pitch. On the other hand, if the pitch of the input voice signal cannot be detected, the pitch of the input voice signal is specified based on the pitch held immediately before. That is, since the harmony voice signal is generated by shifting the pitch of the input voice signal, the pitch of the input voice signal (the pitch corresponding to the pitch name) is used to calculate the pitch shift amount used at that time. Either) is required. If this is the case, if the pitch of the input audio signal cannot be detected, the pitch cannot be determined and the pitch shift amount cannot be calculated, and the original pitch cannot be determined when the pitch is shifted. Cannot be generated. Therefore, when the pitch of the input voice signal cannot be detected, the pitch of the original pitch to be shifted is clarified and the pitch shift amount is determined by specifying the pitch of the input voice signal based on the pitch held immediately before. It can be calculated. In this way, the pitch discontinuity between the adjacent vowel sections without outputting the input voice signal as it is for the section in which the pitch of the input voice signal cannot be detected without preparing special data separately. As a result, it is possible to generate a harmony voice signal that can minimize the timbre and maintain the continuity of the timbre, and thus less cause an unnatural noise to the user.

  An electronic music apparatus according to claim 2 of the present invention is an electronic music apparatus that generates one or a plurality of harmony audio signals whose pitches are controlled following the pitch fluctuation of an input audio signal, Input pitch specification that specifies one of the pitches corresponding to the pitch name for each predetermined section of the input voice signal in accordance with the input means to input and analyzing the input voice signal to detect the pitch Means, a harmony pitch determining means for determining pitches of one or more harmony sounds according to the specified pitches, and the specified pitches and the pitches of the determined one or more harmony sounds, Based on the shift amount calculating means for calculating one or more pitch shift amounts, the shift amount holding means for holding the calculated one or more pitch shift amounts, and on the basis of the held one or more pitch shift amounts. A musical tone generating means for generating one or a plurality of harmony voice signals controlled to the determined one or a plurality of pitches, each of which performs a pitch shift of the input voice signal for each predetermined section; When generating a harmony voice signal in a section in which the pitch of the input voice signal could not be detected, the means performs a pitch shift based on the held one or more pitch shift amounts in the section immediately before the pitch was detected. It is characterized by that. In this way, one or a plurality of harmonies that hardly cause an unnatural noise to the user can be obtained without preparing special data separately in order to generate a harmony voice signal in a section in which the pitch of the input voice signal cannot be detected. An audio signal can be generated. Furthermore, this case is advantageous in that it is not necessary to calculate the amount of pitch shift in order to generate a harmony speech signal in a vowel non-detection section (consonant section), and the processing load can be reduced.

  The present invention can be constructed and implemented not only as a device invention but also as a method invention. Further, the present invention can be implemented in the form of a program of a processor such as a computer or a DSP, or can be implemented in the form of a storage medium storing such a program.

  According to the present invention, the detected pitch or the calculated pitch shift amount is held, and based on this, one or a plurality of harmony audio signals in a section where the pitch of the input audio signal could not be detected is generated. As a result, even if special data is not prepared as in the past, the discontinuity with the pitch in successive vowel sections is small and the timbre continuity is maintained, resulting in an unnatural noise for the user. There is an effect that it is possible to generate one or a plurality of harmony audio signals that are rarely generated.

1 is a block diagram of a hardware configuration showing an example of the overall configuration of an electronic music apparatus according to the present invention. It is a conceptual diagram which shows one Example of the data structure of a harmony table. It is a flowchart which shows one Example of a harmony sound production | generation process. It is a conceptual diagram which shows the specific example for demonstrating a harmony sound production | generation process. It is a flowchart which shows another Example of a harmony sound production | generation process.

  Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

  FIG. 1 is a hardware configuration block diagram showing an embodiment of the overall configuration of an electronic music apparatus according to the present invention. The electronic music apparatus shown in this embodiment is controlled by a microcomputer comprising a microprocessor unit (CPU) 1, a read only memory (ROM) 2, and a random access memory (RAM) 3. The CPU 1 controls the operation of the entire electronic music apparatus. The CPU 1 is connected to the ROM 1, RAM 3, detection circuits 4 and 5, display circuit 6, sound source / effect circuit 7, A / D conversion circuit 8, storage device 9, communication interface (I / O) via the data and address bus 1 D. F) 10 are connected to each other.

  The ROM 2 stores various control programs executed or referred to by the CPU 1, various data such as a harmony table (pitch determination table) shown in FIG. The RAM 3 is used as a working memory for temporarily storing various data generated when the CPU 1 executes a predetermined program, or as a memory for temporarily storing a currently executing program and related data. The A predetermined address area of the RAM 3 is assigned to each function, and is used as a register, a flag, a table, a temporary memory, or the like.

  The performance operator 4A is, for example, a keyboard provided with a plurality of keys for selecting the pitch of a musical tone, and has a key switch corresponding to each key. Of course, the keyboard etc. can be used for manual performance by the user's own playing, and in this embodiment, the user himself / herself plays the chords (chords) to generate the harmony audio signal. Code information for specifying a “harmony table” (see FIG. 2 to be described later) to be referred to at the time of generation can be input. The detection circuit 4 generates a detection output by detecting the pressing and release of each key of the performance operator 4A.

  The setting operator (switch, etc.) 5A is, for example, a harmony automatic generation start / stop button for instructing automatic generation start / stop of a harmony audio signal based on an input audio signal, a setting switch for setting various parameters related to signal control, and the like. It's okay. Of course, the setting operation element 5A has a numeric data input numeric keypad and a character data input keyboard for selecting / setting / controlling pitches, tones, effects, etc., or a pointer displayed on the display 6A. Various operators such as a mouse for operating the above may be included. The detection circuit 5 detects the operation state of the setting operation element 5A, and outputs switch information and the like corresponding to the operation state to the CPU 1 via the data and address bus 1D.

  The display circuit 6 includes, for example, musical scores and lyrics that can be referred to when a user inputs an audio signal to the display 6A including a liquid crystal display panel (LCD), a CRT, or the like via the microphone 8A or the like. Various information such as setting contents of various parameters set by a setting operator (switch or the like) 5A, a list of various data stored in advance, a control state of the CPU 1, and the like are displayed.

  The tone generator / effect circuit 7 can simultaneously generate musical sound signals in a plurality of channels, and is applied to the input audio signal, for example, input via the microphone 8A, which is given via the data and address bus 1D. Generates a harmony voice signal generated based on the above. These signals generated from the sound source / effect circuit 7 are generated by a sound system 7A including an amplifier and a speaker. When the sound source / effect circuit 7 generates an input voice signal, a harmony voice signal, etc., for example, gender (type and depth of voice quality such as male voice and female voice), vibrato (depth and cycle change rate, vibrato start) Delay time), tremolo, volume, pan (localization), detune, reverb (reverberation), and the like. It should be noted that any conventional configuration may be used for the configuration of the sound source / effect circuit 7 and the sound system 7A. For example, the tone generator / effect circuit 7 may employ any of various tone synthesis methods such as FM, PCM, physical model, formant synthesis, etc., may be configured with dedicated hardware, or may be a CPU 1 or DSP (Digital Signal). (Processor) software processing.

  The microphone 8A is a signal input device for inputting a sound signal such as a sound uttered by a user or a performance sound of a musical instrument performed by the user himself / herself. The A / D conversion circuit 8 digitally converts the input voice signal input from the microphone 8A. The digitally converted input audio signal is input to the sound source / effect circuit 7. The signal input device for inputting the audio signal is not limited to the microphone 8A, and may be a reproduction device that reproduces and supplies a previously stored audio signal, for example.

  The storage device 9 stores various information such as a prepared harmony table (see FIG. 2) and various control programs executed by the CPU 1. Or you may enable it to memorize | store the input audio | voice signal input, the produced | generated harmony audio | voice signal, etc. When the control program is not stored in the ROM 2 described above, the control program is stored in the storage device 9 (for example, a hard disk) and read into the RAM 3 to store the control program in the ROM 2. It is possible to cause the CPU 1 to execute the same operation as in FIG. In this way, control programs can be easily added and upgraded. The storage device 9 is not limited to a hard disk (HD), but may be a flexible disk (FD), a compact disk (CD-ROM / CD-RAM), a magneto-optical disk (MO), or a DVD (Digital Versatile Disk). Any storage device may be used as long as it uses a storage medium in the form. Alternatively, a semiconductor memory such as a flash memory may be used.

  A communication interface (I / F) 10 is an interface for transmitting and receiving various information such as a control program and various data between the apparatus and an external device (not shown). The communication interface 10 may be, for example, a MIDI interface, a LAN, the Internet, a telephone line, or the like, and may include both wired and wireless ones.

In the above-described embodiment, the performance operator 4A is not limited to a keyboard instrument, and may be any type such as a stringed instrument, a wind instrument, or a percussion instrument. Further, the electronic music apparatus is not limited to one in which the performance operator 4A, the display 6A, the sound source / effect circuit 7 and the like are built in one apparatus body, but each is configured separately, and a communication interface 10 such as a MIDI interface or various networks. Needless to say, the apparatus may be configured to connect the respective devices using the.
The electronic music apparatus according to the present invention may be any form of apparatus / equipment such as a karaoke apparatus, an electronic musical instrument, a personal computer, a mobile communication terminal such as a mobile phone, or a game apparatus. In the case of a mobile communication terminal, not only the case where a predetermined function is completed with only the terminal, but also a part of the function is provided on the server side so that the predetermined function is realized as a whole system including the terminal and the server. May be.

  The electronic music apparatus shown in FIG. 1 detects the pitch by analyzing the frequency of the input audio signal input via the microphone 8A (finally at a specific pitch corresponding to one of the musical pitch names). Specific), based on the specified pitch and the chord information input from the keyboard or the like, one or more new target pitches (specifically, specific pitches corresponding to any of the musical pitch names). And a harmony sound generation function (or a harmony addition function) for automatically generating a harmony sound signal of the determined target pitch.

  Here, the target pitch is prepared in advance based on a specific pitch corresponding to one of musical pitches obtained by frequency analysis of the input voice signal and chord information input from a keyboard or the like. According to the harmony table (pitch determination table) shown in FIG. 2, any one of twelve scale names (sound names) is determined. FIG. 2 is a conceptual diagram showing the data structure of the harmony table. However, here is an example of a table that is referred to when “C major” is designated as the code information, and is referred to when three series of harmony audio signals are generated simultaneously.

  In the harmony table, a plurality of tables are stored in advance for each code, and one corresponding table is specified according to the code information. As can be understood from FIG. 2, the harmony table includes one or more harmony sequences for each specific pitch (input pitch) corresponding to one of the musical pitch names obtained by frequency analysis of the input voice signal. Define the target pitch for. However, in FIG. 2, the input pitch is indicated by the pitch name “C, D, E, F, G, A, B”, and the input pitch with # (sharp) or ♭ (flat) is shown. Is omitted. Similarly, the target pitches of the three harmony series are also indicated by their pitch names.

  Regarding the pitch notation of the target pitch, for example, the target pitch “G” indicates a “G” tone in the same octave region as the input pitch, and the target pitch “C +” is one level higher than the input pitch. The target pitch “E−” indicates that the sound is the “E” sound in the octave region one level lower than the input pitch. Therefore, in the illustrated example, for example, when the input pitch is “E3”, “C4” is set as the target pitch of the first sequence harmony voice signal, and “G3” is set as the target pitch of the second sequence harmony voice signal. "G2" is determined as the target pitch of the third series harmony voice signal. In FIG. 2, the pitch shift amount (unit: cent), which is the difference between the input pitch and each target pitch, is shown in parentheses for the target pitch for convenience. In this embodiment, an example in which the octave region is divided between “C” and “B” is shown.

  Next, “harmonic sound generation processing” for realizing the above-described harmony sound generation function will be described with reference to FIGS. 3 and 4. FIG. 3 is a flowchart showing an embodiment of “harmonic sound generation processing”. The processing is started in response to an instruction to start automatic generation of harmony sound in accordance with, for example, the operation of the automatic harmony sound generation start / stop button, and is repeatedly executed until an instruction to stop automatic generation of harmony sound is instructed. FIG. 4 is a conceptual diagram showing a specific example for explaining the harmony sound generation processing. The above processing will be described with reference to FIG. 4 as appropriate.

  In step S1, the digitized input voice signal is analyzed, and the pitch is detected at predetermined intervals (for example, every several ms to several tens of ms). That is, an input voice signal input via the microphone 8A or the like is digitized by the A / D conversion circuit 8, and the digitized input voice signal is converted into a frequency signal by “frequency detection” processing. For this “frequency detection” process, any known technique such as the zero-cross method, which is a technique well-known in the field of speech analysis, may be used.

  In step S2, it is determined whether or not the pitch of the input audio signal has been detected. If the pitch of the input audio signal can be detected (YES in step S2), the detected pitch is held (step S3). That is, a pitch holding storage area prepared in advance in the RAM 3 or the like to hold the detected pitch is updated as needed. At this time, a process of updating a storage area for storing one period of waveform element data cut out by the window function corresponding to the detected pitch with the newly cut out waveform element data is performed.

  On the other hand, if the pitch of the input audio signal cannot be detected (NO in step S2), the pitch detected and held immediately before is held (step S9). That is, the storage area for holding the pitch is not updated. At this time, the waveform element data is not updated, and the waveform element data for one cycle stored immediately before is held as it is. Note that the processing in step S9 is substantially omitted because the processing is achieved unless the pitch holding processing in step S3, that is, the update of the pitch holding storage area by the detected pitch is executed. Also good.

  In step S4, the pitch of the input voice signal is specified (detected) by quantizing the held pitch in pitch (note) units. That is, the frequency signal is flattened (or also referred to as smoothing) by “flattening (quantizing)” the frequency signal. The flattened frequency signal is discretized into, for example, any one of twelve-tone scale names (pitch names) every predetermined time by “floor name detection” processing. Specifically, the flattened frequency signal is a plurality of musical pitch names “C, D, E, F, G, A, B” in which semitones (100 cents) are defined, and # (sharp) Or it is rounded to the pitch corresponding to one of the pitch names with ♭ (flat). In this way, the input voice signal that has been input is identified as one of the pitches corresponding to the musical pitch name.

  Here, FIG. 4A shows from the first vowel section having a small pitch fluctuation (less than a semitone and about several to several tens of cents) on the basis of the pitch of the pitch name “E”. , The pitch indicated by the time t1 to t2 is difficult to detect clearly, but the pitch is small up and down based on the pitch of the pitch name “G” through a consonant interval (vowel non-detection interval) with a sense of pitch. It is a series of input speech signals that transition to a second vowel section having shaking (similarly about several to several tens of cents). For example, when the lyrics are “Ika”, the phoneme “i” of the syllable “i” corresponds to the first vowel segment, and the phoneme of the syllable “ka” corresponds to the consonant segment. It is “k”, and the phoneme “a” of the syllable “ka” corresponds to the second vowel section.

  In the case of such an input voice signal, as shown in FIG. 4B, the pitch of the pitch name “E” can be specified in accordance with the held pitch for the first vowel section where the pitch can be detected. Two vowel intervals can be specified for the pitch of the pitch name “G”. However, in the consonant section indicated by the times t1 to t2, the pitch is large and irregular in a short time and cannot be clearly detected. Therefore, the pitch (sound corresponding to one of the pitch names) is not detected. High) cannot be specified. Therefore, in the present invention, the pitch of the consonant section is specified according to the last detected pitch (holding pitch shown in FIG. 4A) in the first vowel section detected and held immediately before. For this reason (see step S9 above), in this example, the pitch of the pitch name “E” can be specified hypothetically as the pitch of the consonant section (described as the assumed pitch in FIG. 4B).

  In step S5, for example, chord information input by the user operating a keyboard or the like is acquired. Step S6 refers to the corresponding harmony table stored in the ROM 2 or the storage device 8 on the basis of the acquired code information and the pitch of the obtained input voice signal, and stores one or more harmony voice signals. Determine the pitch (target pitch). For example, if the input chord information is “C major” and the harmony table shown in FIG. 2 is used, the target pitches in the first vowel section shown in FIG. 4 are “C +”, “G”, “G−”. The target pitches in the second vowel section are “E +”, “C +”, and “C”. Further, in this embodiment, the target pitch in the consonant section is presumed to be the pitch of the pitch name “E” as the pitch of the consonant section as described above, so that the target sound in the first vowel section is determined. High “C +”, “G”, “G-”.

  In step S7, one or more target pitches determined in step S6 and the pitch of the input audio signal detected in step S1 are compared to obtain a difference, and one or more pitch shift amounts are determined according to the obtained difference. Ask for. In step S8, one or more sequences of the target pitch are shifted by pitch-shifting an input audio signal (specifically, the stored waveform element data of one cycle) based on the obtained one or more pitch shift amounts. Generates a harmony voice signal.

  FIG. 4C shows the pitch shift amount in each section obtained in accordance with the process of step S7 by a solid line. However, in the consonant section, for comparison, when the input speech signal is output as it is without being pitch-shifted as a conventional harmony speech signal, that is, when the pitch shift amount of all harmony sequences is once set to “0”. Shown in bold lines. As can be understood from FIG. 4C, the pitch shift amount is “+800” for the first vowel section in which the target pitches are determined as “C +”, “G”, and “G−” as described above. ”,“ +300 ”,“ −900 ”, and the pitch shift amount is“ +900 ”,“ +500 ”in the second vowel section where the target pitch is determined to be“ E + ”,“ C + ”,“ C ”. , “-700”. Conventionally, for a consonant section (pitch non-detection section) in which the pitch shift amount of all the harmony sequences is temporarily set to “0”, the transition from the first vowel section to the second vowel section is performed. The pitch shift amounts “+800”, “+300”, and “−900” with respect to the input speech signal immediately before are substantially maintained, and the first vowel interval and the second vowel interval are compared with the conventional case. It can be understood that the discontinuity of the pitch shift amount is reduced.

  Thus, in the present invention, even in a consonant section where it is difficult to detect the pitch clearly, the pitch shift amount for realizing the target pitch is not once set to “0” for all the harmony sequences. The pitch shift amount can be obtained (set) for each harmony sequence in the same manner as in the vowel section. Accordingly, by performing pitch shift on the input voice signal (the stored waveform element data) in each section, it is possible to generate the harmony voice signals composed of three sequences as shown in FIG. it can.

  As can be understood from FIG. 4 (d), in the first harmony sequence, signals have conventionally transitioned in the order of "C +", "E", and "E +", but in this embodiment, "C +", The signal transitions in the order of “C +” and “E +”. In the second harmony sequence, the signal has conventionally transitioned in the order of “G”, “E”, “C +”, but in this embodiment, the signal has transitioned in the order of “G”, “G”, “C +”. It is supposed to be. That is, in the first and second harmony sequences, in the present embodiment, the time t1 at which transition from the first vowel section to the consonant section and the time t2 at which transition from the consonant section to the second vowel section occur in the present embodiment. In both cases, the pitch discontinuity of the harmony voice signal is small.

  On the other hand, in the third harmony sequence, the signal has conventionally transitioned in the order of “G-”, “E”, “C”, but in this embodiment, “G-”, “G-”, “C”. The signal transitions in this order. That is, in the third harmony sequence, it can be understood that the pitch discontinuity at the time t2 when the present embodiment transitions from the consonant section to the second vowel section is larger than in the conventional example. However, the discontinuity of the pitch that occurred at the time t1 in the past has been eliminated, and the discontinuity of the pitch occurs only at the time t2, and in the past, what was a simple input voice in the consonant section clearly has a sense of pitch. From the point of being captured as a part of the harmony sound, the pitch discontinuity is increased in the present embodiment at time t2, but the above advantages are greater than the influence thereof.

As described above, in the present invention, when the pitch of the input voice signal can be detected, the detected pitch is held, and the pitch of the input voice signal (pitch corresponding to the pitch name) is based on the pitch. On the other hand, if the pitch of the input voice signal cannot be detected, the pitch of the input voice signal is specified based on the pitch held immediately before. That is, since the harmony voice signal is generated by shifting the pitch of the input voice signal, the pitch of the input voice signal (the pitch corresponding to the pitch name) is used to calculate the pitch shift amount used at that time. Any) is required. Therefore, when the pitch of the input voice signal cannot be detected, the pitch of the original pitch to be shifted is clarified and the pitch shift amount is determined by specifying the pitch of the input voice signal based on the pitch held immediately before. I am trying to calculate. In this way, in a section in which the pitch of the input voice signal cannot be detected, the pitch discontinuity between the adjacent vowel sections is minimized and the timbre continuity is reduced without outputting the input voice signal as it is. Can be held. That is, it is possible to easily generate one or a plurality of harmony audio signals that hardly cause an unnatural noise to the user without preparing special data separately.
In the above-described processing, even if the chord is arbitrarily changed in the pitch non-detection section, a harmony voice signal having a pitch corresponding to the change can be advantageously generated.

  As mentioned above, although an example of embodiment was demonstrated based on drawing, this invention is not limited to this, It cannot be overemphasized that various embodiment is possible. For example, the “harmonic sound generation process” for realizing the harmony sound generation function is not limited to the above-described embodiment. FIG. 5 is a flowchart showing another embodiment of the harmony sound generation process.

  In step S11, the digitized input voice signal is analyzed to detect the pitch. In step S12, it is determined whether or not the pitch has been detected. If it is determined that the pitch cannot be detected (NO in step S12), the process jumps to step S18. On the other hand, if it is determined that the pitch has been detected (YES in step S12), the pitch of the input voice signal is obtained by quantizing the detected pitch in pitch (note) units (step S13). A step S14 acquires code information. Step S15 refers to the corresponding harmony table stored in the ROM 2 or the storage device 8 on the basis of the acquired code information and the pitch of the obtained input voice signal, and stores one or more harmony voice signals. Find the pitch (target pitch).

  In step S16, one or more target pitches obtained in step S15 are compared with the pitch of the input audio signal detected in step S11 to obtain a difference, and one or more pitch shift amounts are determined according to the obtained difference. Ask for. Step S17 holds the obtained one or more pitch shift amounts. In step S18, the input voice signal is pitch-shifted based on the held one or more latest pitch shift amounts, thereby generating one or more harmony series harmony voice signals of the target pitch. However, in a section where the pitch of the input voice signal cannot be detected, the input voice signal of the section immediately before the pitch can be detected is pitch-shifted to generate one or a plurality of harmony-sequence harmony voice signals in the section. To do.

  In this way, as in the above-described embodiment (see FIG. 3), the discontinuity in the pitch shift amount between the first vowel section and the second vowel section is alleviated as compared with the conventional example, and the consonant. One or more series of harmony audio signals can be obtained even in the section. In this embodiment, since the pitch shift amount with respect to the input speech signal immediately before the consonant section is maintained, there is no need to calculate the difference between the held pitch and the target pitch and calculate the pitch shift amount. For this reason, it is advantageous to reduce the processing burden.

The chord information input to generate the harmony audio signal is detected from the input information input in response to a user operation from a performance operator such as a keyboard connected to the apparatus or the apparatus as described above. May be obtained, or may be obtained by sequentially inputting chord names. Alternatively, the chord information may be acquired by reproducing the automatic performance data including chord name data according to the progress of the song.
In the above-described embodiment, the pitch of the harmony voice signal is determined based on the chord information (more specifically, the harmony table). However, the present invention is not limited to this, and the harmony voice is not based on the chord information. Other known methods for determining the pitch of the signal may be used. For example, a method may be employed in which the pitch of the harmony voice signal is determined based on a predetermined pitch separated from the pitch of the input voice signal (for example, three degrees above).

In the above-described embodiment, although three series of harmony voice signals are simultaneously generated with respect to the input voice signal, the present invention is not limited to this, and three or more series harmony voice signals are simultaneously generated. It may be a thing.
The pitch detection result of the harmony voice signal is not limited to the pitch detection result of the voice signal input as described above, and the pitch detection result of the input voice signal is, for example, one octave or three semitones. You may make it determine the pitch of a harmony audio | voice signal using what converted pitch, such as going up and down only a predetermined pitch.
In the above-described embodiment, the input audio signal that is the basis for generating the harmony audio signal is described as an example of the audio input through the microphone 8A. However, for example, the musical instrument performance input through the microphone 8A is used. Sound may be used. In the case of a musical instrument performance sound, the additional sound may be an accompaniment sound.

1 ... CPU, 2 ... ROM, 3 ... RAM, 4,5 ... detection circuit, 4A ... performance operator, 5A ... setting operator, 6 ... display circuit, 6A ... display, 7 ... sound source / effect circuit, 7A ... sound System: 8 ... A / D conversion circuit, 8A ... Microphone, 9 ... Storage device, 10 ... Communication interface, 1D ... Data and address bus

Claims (4)

An electronic music device that generates one or a plurality of harmony voice signals whose pitches are controlled in accordance with pitch fluctuations of an input voice signal,
An input means for inputting an audio signal;
Analyzing the input voice signal and detecting the pitch, an input pitch specifying means for specifying any of the pitches corresponding to the pitch name for each predetermined section of the input voice signal;
Harmony pitch determination means for determining the pitch of one or more harmony sounds according to the specified pitch;
Shift amount calculating means for calculating one or more pitch shift amounts based on the identified pitches and the determined pitches of the one or more harmony sounds;
Based on the calculated one or more pitch shift amounts, the input audio signal is pitch-shifted for each predetermined interval, and the determined one or more harmony audio signals controlled to the plurality of pitches are obtained. A musical sound generating means for generating,
When the pitch of the input voice signal can be detected, the input pitch specifying means holds the detected pitch and specifies the pitch of the input voice signal based on the pitch. An electronic music apparatus characterized in that, when a pitch of a signal cannot be detected, a pitch of the input audio signal is specified based on the pitch held immediately before. An electronic music device that generates one or a plurality of harmony voice signals whose pitches are controlled in accordance with pitch fluctuations of an input voice signal,
An input means for inputting an audio signal;
Analyzing the input voice signal and detecting the pitch, an input pitch specifying means for specifying any of the pitches corresponding to the pitch name for each predetermined section of the input voice signal;
Harmony pitch determination means for determining the pitch of one or more harmony sounds according to the specified pitch;
Shift amount calculating means for calculating one or more pitch shift amounts based on the identified pitches and the determined pitches of the one or more harmony sounds;
Shift amount holding means for holding the calculated one or more pitch shift amounts;
Based on the held one or more pitch shift amounts, the input audio signal is pitch-shifted for each predetermined section, and the determined one or more harmony audio signals controlled to the plurality of pitches are obtained. A musical sound generating means for generating,
The musical sound generation means, when generating a harmony voice signal in a section in which the pitch of the input voice signal could not be detected, based on the held one or more pitch shift amounts in the section immediately before the pitch was detected. An electronic music apparatus characterized by pitch shifting. A computer-executable program for generating one or more harmony voice signals whose pitches are controlled following the pitch fluctuation of an input voice signal, the program being stored in the computer,
Input audio signal,
A procedure for identifying one of pitches corresponding to a pitch name for each predetermined section of the input voice signal, by detecting the pitch by analyzing the input voice signal, the procedure comprising: When the pitch of the input audio signal can be detected, the detected pitch is held and the pitch of the input audio signal is specified based on the pitch, while the pitch of the input audio signal cannot be detected Identifying the pitch of the input audio signal based on the pitch held immediately before;
Determining a pitch of one or more harmony voices according to the identified pitch;
Calculating one or more pitch shift amounts based on the identified pitch and the determined pitches of the one or more harmony voices;
Based on the calculated one or more pitch shift amounts, the input audio signal is pitch-shifted for each predetermined interval, and the determined one or more harmony audio signals controlled to the plurality of pitches are obtained. A program that executes the generating procedure. A computer-executable program for generating one or more harmony voice signals whose pitches are controlled following the pitch fluctuation of an input voice signal, the program being stored in the computer,
Input audio signal,
A procedure for identifying one of pitches corresponding to a pitch name for each predetermined section of the input voice signal, by analyzing the input voice signal and detecting a pitch;
Determining a pitch of one or more harmony voices according to the identified pitch;
Calculating one or more pitch shift amounts based on the identified pitch and the determined pitches of the one or more harmony voices;
A procedure for holding the calculated one or more pitch shift amounts;
Based on the held one or more pitch shift amounts, the input audio signal is pitch-shifted for each predetermined section, and the determined one or more harmony audio signals controlled to the plurality of pitches are obtained. A step of generating, in the step of generating a harmony voice signal in a section in which the pitch of the input voice signal could not be detected, the held one or more pitches in the section immediately before the pitch was detected. A program that executes what shifts the pitch based on the shift amount.

Images