JP3693981B2 - Pitch converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To convert a pitch while maintaining a formant of a former audio signal by synthesizing segmented phoneme pieces. SOLUTION: A pitch converting means 24 reads the phoneme pieces from an input waveform data stored in a RAM12 between intervals of two cycles, gives an envelope, synthesizes in a cycle corresponding to the reproducing pitch, and outputs.

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a pitch converter for converting the pitch of an audio signal or a musical sound signal.
[0002]
[Prior art]
Conventionally, various pitch converters have been proposed. For example, Japanese Patent Laid-Open No. 62-65098 discloses a music vocoder configured to cut out an input audio signal and reproduce the audio signal at a pitch selected by a keyboard. In this vocoder, even if the pitch of the reproduced audio signal is different from that of the input audio signal, the formant of the reproduced audio signal can be substantially maintained as the formant of the input audio signal.
[0003]
[Problems to be solved by the invention]
In the vocoder, one period of the original voice is cut out and reproduced and synthesized at a time interval corresponding to a desired pitch. In this method, a rectangular wave window is applied, and as a result, extra harmonics are generated. An object of the present invention is to provide a pitch conversion device that generates as little extra harmonics as possible.
[0004]
[Means for Solving the Problems]
The present invention provides a memory for storing an audio signal having a plurality of periods, a pitch detection means for detecting the pitch of the audio signal, a reproduction pitch information storage means for storing reproduction pitch information for setting a reproduction pitch, and the pitch detection. The audio signal stored in the memory according to a pitch period detected by the means; Is read at a reading speed set independently of the reproduction pitch information. A phoneme segmentation means for segmenting a phoneme segment, and an envelope that gradually shifts from the minimum value to the maximum value with a length equal to the time required to read the phoneme segment to the segmented segment. And stored in the reproduction pitch information storage means In units of 2 cycles based on playback pitch information Reproduction audio signal synthesis means for synthesizing and generating a reproduction audio signal is provided.
[0005]
According to the present invention, a plurality of periods of speech segments are cut out from an audio signal, given a predetermined envelope, and then synthesized at a period corresponding to the playback pitch, so that unnecessary harmonics are less generated and the original audio signal is reduced. The pitch conversion is performed while maintaining the formant.
[0006]
DETAILED DESCRIPTION OF THE INVENTION
As shown in FIG. 1A, the first embodiment of the present invention has an input terminal 2 to which an audible frequency signal such as an analog musical sound signal or a voice signal is input, and is supplied to this input terminal 2. The audible frequency signal is converted into a digital audible frequency signal (sampling data) by the A / D converter 4. In order to prevent the occurrence of aliasing between the A / D converter 4 and the input terminal 2, the audio frequency signal is limited to a frequency that is ½ or less of the sampling frequency in the A / D converter 4. A low-pass filter 6 is provided.
[0007]
Sampling data from the A / D converter 4 is supplied to a DSP (digital signal processor) 8 and then supplied to the RAM 12. The RAM 12 is used as a ring memory that sequentially stores input sampling data.
[0008]
The DSP 8 reads the sampling data from the ring memory, performs processing, and supplies it to the D / A converter 14. The D / A converter 14 converts the sampling data processed by the DSP 8 into an analog audible frequency signal, supplies the analog audio frequency signal to the low-pass filter 16, removes unnecessary signal components, and supplies the signal to the output terminal 18.
[0009]
The DSP 8 processes sampling data according to a preset program. For the processing, parameters such as formant change coefficient FORMAT-VR and pitch change coefficient PITCH-VR set by the operator 20 operated by the user are used, for example. The formant change coefficient FORMAT-VR and the pitch change coefficient PITCH-VR are detected by the CPU 22 and supplied to the DSP 8. In addition to this, the CPU 22 also controls the DSP 8.
[0010]
1 (b) and 1 (c) show functions executed by the DSP 8 in a functional block diagram. FIG. 1 (b) shows a basic embodiment of the present invention, and FIG. Other embodiment of this invention is shown. In these drawings, there are functional blocks for sequentially storing sampling data input to the input terminal 23 in the ring memory of the RAM 12, but the functional blocks are omitted for ease of explanation.
[0011]
In the basic embodiment of FIG. 2B, input means, for example, an input terminal 23, a pitch conversion means 24, a pitch detection means 26, a control means 28, and an output terminal 29 are provided. The control unit 28 uses the pitch detection signal from the pitch detection unit 26 and parameters such as the formant change coefficient FORMANT-VR and the pitch change coefficient PITCH-VR set by the operation of the operator 20 via the CPU 22. The sampling data stored in is processed by the pitch conversion means 24.
[0012]
The processing performed by the pitch conversion unit 24, the pitch detection unit 26, and the control unit 28 will be briefly described. The formant change coefficient FORMANT− is obtained with the passage of time as a phoneme from the sampling data stored in the ring memory. Cut out sequentially at a reading speed according to VR, change formant characteristics as appropriate, and reproduce the cut out phoneme at a period corresponding to the pitch detected by the pitch detection means 26 and the pitch change coefficient PITCH-VR, This enables playback at a pitch and formant different from those at the time of input.
[0013]
In the embodiment of the present application, the phoneme is reproduced using a processing path for reproducing the first waveform and a processing path for reproducing the second waveform. Phonemes are reproduced at a period twice as long as
[0014]
This effect adding operation will be described with reference to the flowcharts shown in FIGS. The DSP 8 includes a register PITCH that stores a detection pitch of sampling data, and a register SADRS that stores an address for reading (reading) sampling data from the ring memory. Further, a counter PHASE for counting whether a reproduction pitch period length described later has been reached, a counter PH1 for counting the phase of the first waveform, and a counter PH2 for counting the phase of the second waveform are also provided. ing.
[0015]
Further, a register WIDTH for storing reproduction pitch information (reproduction pitch period length) in which the pitch detected by the pitch detection means 26 is changed by the pitch change coefficient PITCH-VR, and the detected pitch and the formant coefficient FORMAT-VR are determined. A register LENGTH for storing the length of the envelope, a register ENV1 for storing the envelope of the first waveform, and a register ENV2 for storing the envelope of the second waveform are provided.
[0016]
Further, the step rate of WINDOW1 and WINDOW2 determined based on the values of the register WINDOW1 for determining the shape of the envelope of the first waveform, the registers WINDOW2 and LENGTH for determining the shape of the envelope of the second waveform, Register W-RATE to be stored, register SADRS1 to store the first waveform cut-out start address, register SADRS2 to store the second waveform cut-out start address, determination of the first and second waveform cut-out start positions, etc. A flag F used for the purpose is also provided.
[0017]
These are initialized when the power is turned on. That is, the flag F is set to 1 and the others are set to 0. In the following description, it is assumed that appropriate values are already stored in each register and each counter. Each step of the flowcharts shown in FIGS. 2 and 3 is executed each time sampling data is input from the A / D converter 4 to the DSP 8.
[0018]
In FIG. 2, when sampling data is supplied from the A / D converter 4, it is written in the ring memory (step S2). Next, pitch detection processing is performed based on the input sampling data (step S4). This step S4 corresponds to the pitch detection means 26. This pitch detection process outputs the address of the zero cross position and the detected pitch when the pitch is detected.
[0019]
This pitch detection process sequentially compares time intervals between adjacent zero crosses, for example. For example, in FIG. 4A, it is assumed that the time intervals between the zero crosses of the input signal are a0, b0, c0, d0, a1, b1, c1, and d1. The time interval a0 between the first zero crosses and the time interval b0 between the next zero crosses are compared. If they are different, the time interval (a0 + b0) obtained by adding them is compared with the adjacent time interval (c0 + d0). . If both are different in this comparison, the time interval (a0 + b0 + c0) is compared with the adjacent time interval (d0 + a1 + b1). If they are still different, the time interval (a0 + b0 + c0 + d0) is compared with the adjacent time interval (a1 + b1 + c1 + d1). If the two values are almost the same in this comparison, the time interval (a1 + b1 + c1 + d1) is output as the pitch, and Z1 is output as the address of the zero cross position.
[0020]
In addition, as disclosed in, for example, Japanese Patent Laid-Open No. 3-288200, the zero cross position of the sampling data and the peak position of the waveform signal are detected, and the time interval between the zero cross position and the peak position is previously detected. It is also possible to use one that detects the pitch by comparing the time interval with the zero cross position.
[0021]
In step S4, it is determined whether or not pitch detection has been performed (step S6). If the pitch is detected, the zero-cross position detected in step S4 is used as a cut-out address, thereby updating and detecting the value of the register SADRS. The pitch is set to one cycle length, and the value of PITCH is updated by this (step S8). Thus, the pitch detection means 26 constitutes a part of the cutout means.
[0022]
When pitch detection is not performed or when pitch detection is performed and both the above registers are updated, the values of the registers PHASE, PH1, and PH2 are each incremented by one (step S10).
[0023]
Next, the values of the registers PHASE and WIDTH are compared (step S12). WIDTH stores one cycle length obtained by multiplying the value of the register PITCH and the pitch change coefficient PITCH-VR, as will be described later, and determines whether the PHASE value has reached this WIDTH value. are doing. That is, it is determined whether or not the PHASE value has reached a predetermined reproduction pitch. The value of WIDTH corresponds to reproduction pitch information.
[0024]
If the value of PHASE has not reached the value of WIDTH, the process proceeds to waveform processing in step S32 described later. When the value of PHASE reaches the value of WIDTH, the value of the register PHASE is set to 0 (step S14), and the pitch change coefficient PITCH-VR and the formant change coefficient FORMAT-VR set by the operator 20 are set via the CPU 22. Read (step S16).
[0025]
In order to determine a new WIDTH value, the PITCH value and the pitch change coefficient PITCH-VR are multiplied and stored in the register WIDTH, and the envelope of the first waveform and the envelope of the second waveform are also stored. In order to determine the period, the value of PITCH is divided by the formant coefficient FORMAT-VR and stored in the register LENGTH (step S18).
[0026]
Next, the value of the register LENGTH is compared with the value of the register WIDTH (step S20). If the value of LENGTH is larger than the value of WIDTH, the value of LENGTH is set as the value of WIDTH (step S22). This is to prevent the LENGTH value from exceeding the WIDTH value. In other words, since only two processing paths for adding waveforms are prepared in the waveform reading process of step S32 described later, two or more waveforms are prevented from overlapping.
[0027]
If the LENGTH value is less than or equal to the WIDTH value, the process of step S24 is performed without performing the process of step S22. Following step S22, the process of step S24 is also performed. In step S24, the reciprocal of the LENGTH value is obtained and stored in the register W-RATE. This W-RATE value is used to step up the values of WINDOW1 and WINDOW2 as will be described later. Further, in step S24, the value of the flag F is also inverted from the previous value.
[0028]
Next, it is determined whether the value of the flag F reversed in step S24 is 1 or -1 (step S26). If the flag F is 1, the registers PH1 and WINDOW1 are set to 0 to start reading the first waveform, and the SADRS value (cutout address) determined in step S8 is stored in the cutout start address register SADRS1 (step S28).
[0029]
If the value of F inverted in step S24 is -1, the reading of the second waveform is started, so that the registers PH2 and WINDOW2 are set to 0, and the value (cutout address) of the register SADRS determined in step S8 is set. The cutout start address register SADRS2 is stored (step S30).
[0030]
By such steps S12, S14, S16, S18, S20, S22, S24, S26, S28, and S30, the first and second time intervals corresponding to the period length of two playback pitches (twice the value of WIDTH) are passed. The cut start address of the second waveform data is updated, and the value of the flag F is inverted every time the PHASE value reaches the WIDTH value.
[0031]
Following step S28 or S30, or if it is determined in step S12 that the value of the register PHASE has not reached the value of the register WIDTH, a waveform reading process is performed (step S32). In this manner, the waveform reading process is performed without changing the WIDTH value and LENGTH value (and thus the W-RATE value) until the PHASE value reaches WIDTH.
[0032]
In the waveform readout process described later, the values of envelopes ENV1 and ENV2 of the first and second waveforms are controlled by the value of W-RATE.
[0033]
Since the value of W-RATE, and hence LENGTH, is obtained by changing the detected pitch by the formant change coefficient FORMANT-VR in step S18, the pitch of the waveform to be subjected to the waveform reading process depends on the formant change coefficient FORMANT-VR. It will be.
[0034]
In this waveform reading process, as shown in FIG. 3, first, the value of the counter WINDOW1 is incremented by the value of W-RATE (step S34). Then, it is determined whether the value of the incremented WINDOW1 is less than 1, 1 or more, less than 2, or 2 or more (step S36).
[0035]
If it is smaller than 1, the value of WINDOW1 is stored in the register ENV1 (step S38). If it is 1 or more and smaller than 2, the value obtained by subtracting the value of WINDOW1 from 2 is stored in the register ENV1 (step S40). At this time, the value of ENV1 is set to 0 (step S42).
[0036]
In steps S34 to S40, a sawtooth wave whose value increases by the value of W-RATE is created, and this value is folded back to 1, thereby creating ENV1. However, if the value of WINDOW1 exceeds 2, ENV1 is set to 0 in step S42. That is, a triangular wave that increases to 1 by W-RATE, which is the reciprocal of the value of LENGTH determined based on the formant coefficient FORMANT-VR and the detection pitch, and then decreases to 0 by W-RATE is converted into an envelope of the first waveform. It is created as.
[0037]
Further, following step S38, S40, or S42, a register that stores the cut-out start address of the first waveform by a value obtained by multiplying the value of the register PH1 (step value of the cut-out address) by the formant coefficient FORMAT-VR. The value is added to the value of SADRS1 and stored in the register ADRS1 that stores the cut-out address of the first waveform (step S44). Following this, the waveform data DATA1 of the first waveform is read from the ring memory at the cut-out address ADRS1 (step S46).
[0038]
As described above, since the read address is changed by the formant coefficient FORMAT-VR, as a result, the read speed of the waveform data DATA1 is changed by the formant coefficient FORMAT-VR.
[0039]
Following this, the value of WINDOW2 is incremented by W-RATE (step S48). It is determined whether the value of the incremented WINDOW2 is less than 1, 1 or more and less than 2, or 2 or more (step S50). If the value of WINDOW2 is smaller than 1, the value of WINDOW2 is stored in the register ENV2 (step S52). If the value of WINDOW2 is 1 or more and less than 2, the value obtained by subtracting the value of WINDOW2 from 2 is set in the register ENV2. (Step S54), and if the value of WINDOW2 is 2 or more, the value of ENV2 is set to 0 (step S56). In this way, the envelope of the second waveform is prepared.
[0040]
Subsequent to step S52, S54, or S56, the value obtained by multiplying the value of the register PH2 by the formant coefficient FORMAT-VR and the value of the cut-out start address SADRS2 for the second waveform data are added to the second waveform data. Is stored in the register ADRS2 for the cut-out address (step S58). Then, the waveform data (DATA2) of the second waveform is read from the ring memory at the address ADRS2 (step S60).
[0041]
The output OUT is obtained by adding the value of ENV1 multiplied by the value of DATA1 read out in this way and the value of DATA2 multiplied by the value of ENV2 (step S62). After such waveform reading processing is performed, an output OUT is sent out as shown in FIG. 2 (step S64).
[0042]
2 and 3, steps S2, S4, S6, S8, S26, S28, S30, S46 and S60 correspond to the cutting means. 1 and the means for generating WIDTH in steps S16 and S18 in FIG. 2 are the reproduction pitch information generating means. Similarly, the operator 20 shown in FIG. 1 and the FORMAT-VR in step S16 in FIG. Means for generating formant change information generating means. Steps other than those described above in FIGS. 2 and 3 correspond to reproduction audio signal synthesis means.
[0043]
FIG. 4 shows a waveform of each part when the pitch change coefficient PITCH-VR and the formant change coefficient FORMAT-VR are both 1. FIG. 5 shows that the pitch change coefficient PITCH-VR is larger than 1, and the formant change coefficient FORMAT- FIG. 6 shows the waveform of each part when the pitch change coefficient PITCH-VR is 1 and the formant change coefficient FORMAT-VR is greater than 1.
[0044]
As apparent from these waveforms, in this embodiment, the pitch of the output signal is determined by the value of WIDTH obtained by multiplying the pitch change coefficient PITCH-VR and the detected pitch, that is, the pitch reproduction information, and the output signal. Is read by the formant change coefficient FORMANT-VR. The pitch change coefficient and the formant change coefficient FORMAT-VR can be changed independently of each other by the operation of the operation element 20.
[0045]
Next, an embodiment in the case of changing the formant change information with time in the same configuration as shown in FIG. 1B will be described. The effect addition processing performed by the pitch conversion unit 24, the pitch detection unit 26, and the control unit 28 in this case is shown in FIG. 2 and FIGS. 8 and 9 which are detailed processing of the waveform readout processing S32 of FIG. Has been. In this embodiment, the operator setting such as the modulation waveform selection parameters (MSTART, MLENG) from the operator 20 is further input.
[0046]
In the pitch converting means 24, a modulation waveform memory composed of, for example, a ROM is provided. In this modulation waveform memory, for example, a plurality of, for example, three modulation waveforms having different waveforms starting with 1 and ending with 1 as shown in FIG. 7 are stored. SA1, SA2, and SA3 shown in FIG. 7 represent the start addresses of these modulation waveforms, and LENG1, LENG2, and LENG3 represent the lengths of the respective modulation waveforms.
[0047]
One of these modulated waveforms is selected by the operation of the operation element 20. The start address of the selected waveform is stored in the register MSTART provided in the pitch converting means 24, and the length of the selected waveform is stored in the same register. Each stored in MLENG.
[0048]
Other registers and counters are used in the same manner as in the above-described embodiment, and further, the values of registers MAD1, MAD2, MAD1, MAD2, and the value of MLENG that store values obtained by halving the values of registers WINDOW1, WINDOW2 Registers MODAD1 and MODAD2 are provided for storing addresses for reading the waveform data from the modulation memory using the value of MSTART and the value of MSTART.
[0049]
In this embodiment, since FIG. 2 is the same as the above-described embodiment, the flowcharts of FIGS. 8 and 9 different from the above-described embodiment will be described. Therefore, in the waveform reading process, as shown in FIG. 8, in step S34, the current WINDOW1 value is updated by the value of W-RATE, and ½ of the updated WINDOW1 value is stored in the register MAD1 (see FIG. 8). Step S64). Next, the value of the register MAD1 is compared with 1 (step S66). If the value of MAD1 is greater than 1, the value of MAD1 is set to 1 (step S68).
[0050]
In step S66, if the value of register MAD1 is smaller than 1, or following step S68, the value obtained by multiplying the value of register MAD1 and the value of MLENG and the value of MSTART is the first waveform data. Is stored in the register MODAD1 as the address of the modulation data of the reading speed (step S70). Subsequently, data MODDATA1 is read from an address corresponding to the value of MODAD1 in the modulation waveform memory (step S72).
[0051]
In step S64, the value of WINDOW1 is halved so that the value of MODAD1 in step S70 does not exceed the final address of the selected modulation waveform memory.
[0052]
Following this, steps S36, S38, S40, and S42 are executed to store the envelope for the first waveform in the register ENV1.
[0053]
Next, the value of the counter PH1 advanced in step S10 is multiplied by the formant coefficient FORMAT-VR and MODDATA1 obtained in step S72, and this multiplied value is multiplied by the cut-out start address register SADRS1 determined in step S28. A value obtained by adding the values is stored in the register ADRS1 (step S44a). This is used as the read address of the waveform data of the first waveform, and the waveform data DATA1 of the first waveform is read from the read address of the waveform data in the ring memory (step S46a). In this “other embodiment”, steps S34, S64, S66, S68, S70, S72 and the means for generating the formant change coefficient FORMAT-VR correspond to the formant change information generating means.
[0054]
Next, as shown in FIG. 9 following the flowchart of FIG. 8, the value of the counter WINDOW2 is incremented by the value of W-RATE (step S48). Then, the value of WINDOW2 is halved and stored in the register MAD2 (step S74). Then, the value of MAD2 is compared with 1 (step S76).
[0055]
If the value of MAD2 is greater than 1, the value of MAD2 is corrected to 1 (step S78). In step S76, when the value of MAD2 is 1 or less, or following step S76, a value obtained by adding the value of MSTART to the product of the value of MAD2 and the value of MLENG is stored in register MODAD2 (step S80). . The modulation data MODDATA2 for the second waveform data is read from the address corresponding to the value of MODAD2 in the modulation waveform memory (step S82).
[0056]
In step S74, the value of WINDOW2 is halved for the same reason that the value of WINDOW1 is halved in step S64.
[0057]
Following this, steps S50, S52, S54, and S56 are executed to store the envelope for the second waveform in the register ENV2.
[0058]
Next, the value of the counter PH2 advanced in step S10 is multiplied by the formant coefficient FORMAT-VR and MODDATA2 obtained in step S82, and this multiplication value is stored in the cut-out start address register SADRS2 determined in step S30. A value obtained by adding the values is stored in the register ADRS2 and used as a read address of the waveform data of the second waveform (step S58a). Similarly to the above, the steps S48, S74, S76, S78, S80, S82 and the means for generating the formant change coefficient FORMAT-VR correspond to the formant change information generating means. The second waveform data DATA2 is read from the read address of the waveform data in the ring memory (step S60a).
[0059]
A value obtained by multiplying the first waveform data DATA1 and the first envelope of ENV1 and a value obtained by multiplying the second waveform data DATA2 and the second envelope of ENV2 are added to the output OUT. (Step S62a). Thereafter, step S64 shown in FIG. 2 is executed.
[0060]
Thus, in this embodiment, the read addresses of the first and second waveform data are modulated using the modulation waveform stored in the modulation waveform memory. Therefore, it is possible to add an unprecedented unique effect to the input audio frequency signal. In addition, a plurality of modulation waveform data used for applying this modulation are prepared, and any of them can be used, so various effects can be added. In addition, since the degree of modulation can be changed by changing the value of the formant change coefficient FORMAT-VR by operating the operation element 20, the same modulation waveform data can be obtained by changing the value of the formant change coefficient FORMAT-VR. Even if is used, different effects can be added.
[0061]
Next, the case where the present invention is implemented in the configuration shown in FIG. 1C instead of the configuration in FIG. 1B in which one of the two embodiments described above is implemented will be described. In the embodiment shown in FIG. 1 (c), a structure in which the waveform converting means 30 and the filter means 32 are cascade-connected is added between the pitch converting means 24 and the input terminal 23 in the configuration of FIG. 1 (b). ing. In addition to the pitch change coefficient PITCH-VR and the formant change coefficient FORMAT-VR, the control means 28 is supplied from the operator 20 such as setting parameters of the waveform converting means 30 and setting parameters of the filter means 32 via the CPU 22. The operator settings are also entered.
[0062]
The waveform converting means 30 is a so-called distortion circuit that clips a predetermined level or more of the input waveform, and the details thereof are well known, and thus the description thereof is omitted.
[0063]
As the filter means 32, a known comb filter, low-pass filter, high-pass filter, or the like, or a combination thereof can be used. In this embodiment, for example, it is disclosed in Japanese Patent Laid-Open No. 6-4076. In such a case, the one whose characteristics change according to the pitch of the input musical sound signal is used.
[0064]
That is, in the filter means disclosed in Japanese Patent Laid-Open No. 6-4076, the filter characteristics are controlled in accordance with the pitch of the input musical sound signal to obtain an appropriate frequency characteristic for the harmonics of the input musical sound signal. ing.
[0065]
Therefore, in this embodiment, after the audible frequency signal supplied to the input terminal 23 is distorted by the waveform converting means 30, the filter means 32 has an appropriate frequency characteristic with respect to the overtone, and as described above. The formant read speed is modulated, the modulation state is adjusted by the formant change coefficient FORMAT-VR, and the output pitch is also adjusted by the pitch change coefficient PITCH-VR. In addition to the effects, it is possible to add unique effects that are not found in the prior art.
[0066]
In both of the above embodiments, two processing paths for generating the first and second waveform data that are phonemes are provided, and in each processing path, the phonemes are reproduced at a cycle twice the reproduction pitch. However, three or more processing paths may be provided and combined. Furthermore, in the above embodiment, the pitch detecting means is used. However, instead of this, for example, a means for detecting the frequency of harmonics of the input audio frequency signal may be used.
[0067]
A second embodiment of the present invention will be described. In this embodiment, as shown in FIG. 10, for example, a digital voice signal obtained by digitizing a singing voice sung according to a karaoke performance from the karaoke performance unit 40 as shown in FIG. It is input to the unit 44, where it is possible to add a chorus effect and output it. In this case, the chorus adding unit 44 adds the chorus effect by synthesizing the original voice input with the voice input pitch and formant changed.
[0068]
The present invention is used to change the pitch and formant. Information for changing the pitch and changing the formant is given to the chorus adding unit 44 by operating the operator 46. Further, pitch information and level information are supplied from the karaoke performance unit 40 to the chorus adding unit 44 as MIDI signals. The pitch information and level information may be supplied not to the karaoke performance unit 40 but to the chorus adding unit 44 as an external input MIDI signal from another device provided outside, for example, a keyboard that can be played. .
[0069]
There are two choruses added by the chorus adding unit 44: a karaoke mode in which chorus is added according to pre-stored performance information, and a choir mode in which a chorus is performed by a large number of people. In this embodiment, pitch information and level information are supplied from the karaoke performance unit 40 to the chorus adding unit 44 in the karaoke mode, and are supplied from the external device in the quire mode.
[0070]
The karaoke performance unit 40 includes an automatic performance means 48 and a musical sound generation means 50, and also includes a karaoke control means 52 for controlling them. The automatic performance means 48 corresponds to a conventionally known automatic performance device and stores automatic performance information of a plurality of songs. The musical sound generating means 50 corresponds to a conventionally known sound source device.
[0071]
In response to a command generated by the karaoke control means 52 in accordance with information supplied to the karaoke control means 52 by an operator control means 54 (to be described later) in response to an operation of the operation element 46, the automatic performance means 48 performs performance music from the automatic performance information. Information is selected, and the musical sound generating means 50 is automatically played according to the information, and accompaniment (karaoke) is played. The performance information of the chorus part is also stored in the automatic performance information, which includes the above-described pitch information and level information, and is supplied to the chorus adding unit 44 in the form of a MIDI signal.
[0072]
The chorus adding unit 44 includes chorus effect adding means 56 and chorus control means 58. As shown in FIG. 11, the chorus effect adding means 56 has a main channel 60M to which a digital audio signal supplied from the audio input terminal 42 is supplied, and a plurality of, for example, four chorus channels 60C1 to 60C4.
[0073]
The main channel 60M includes a multiplier 62M that adjusts the level of the digital audio signal, and an effect adding unit 64M that adds an effect such as a reverberation effect to the digital audio signal whose level has been adjusted.
[0074]
The chorus channels 60C1 to 60C4 change the pitch and formant of the input digital audio signal, and further add pitch conversion means 66C1 to 66C4 for adding effects such as a reverberation effect, and outputs of these pitch conversion means 66C1 to 66C4. Delay means 68C1 to 68C4 for delaying and multipliers 62C1 to 62C4 for adjusting the output levels of the delay means 68C1 to 68C4, respectively.
[0075]
The pitch converting means 66C1 to 66C4 correspond to the structure shown in FIG. 1C, for example, excluding the pitch detecting means, and the waveform converting means 30 and the filter means 32 correspond to the effect adding means here. As this effect adding means, for example, a known reverberation adding device, or a delay effect of the delay means 76 is changed by a low frequency oscillator 78 as shown in FIG. A pseudo chorus effect adding device may be used. For use in each of the pitch converting means 66C1 to 66C4, a pitch detecting means 74 for detecting the pitch of the digital audio signal is provided. In some cases, it is not necessary to provide effect adding means for each of the pitch converting means 66C1 to 66C4.
[0076]
The outputs of the chorus channels 60C1 to 60C4 are combined by the adder 70, and then an effect such as a reverberation effect is added by the effect adding means 64C. The effect adding means 64C also functions as a pseudo chorus effect adding device in FIG. 12 by setting effect parameters as will be described later. The output of the effect adding means 64C is combined with the output of the effect adding means 64M by the adder 72 and output.
[0077]
Also, each of the pitch conversion means 66C1 to 66C4 includes formant change information such as formant change coefficients F1 to F4, pitch change information such as pitch change coefficients P1 to P4, pitch information SP1 to SP4 of chorus sound, and pitch conversion means 66C1 to 66C1. The effect addition information e1 to e4 of the effect addition means in 66C4 is supplied from the chorus control means 58, respectively.
[0078]
Further, delay information D1 to D4 is supplied from the chorus control means 58 to each of the delay means 68C1 to 68C4. Further, level information L0 to L4 is supplied from the chorus control means 58 to the multipliers 62M and 62C1 to 62C4. Similarly, effect addition parameters EM and EC are supplied from the chorus control unit 58 to the effect addition units 64M and 64C, respectively. The effect addition parameters are each composed of a plurality of effect parameters.
[0079]
As will be described later, the chorus controller 58 generates each piece of information based on information from the operator controller 54 and a MIDI signal from the karaoke performance unit 40 or an external device.
[0080]
The chorus adding unit 44 can be constituted by, for example, the DSP and RAM shown in FIG. 1A, and the karaoke control means 52, the automatic performance means 48, and the operator control means 54 are controlled by the CPU shown in FIG. Can be configured.
[0081]
The pitch conversion means 66C1 to 66C4 may use the pitch conversion means 24 shown in FIG. 1C as it is, but in this embodiment, it is possible to obtain a natural audio signal that is more connected than the pitch conversion means 24. In addition, the pitch converting means 66C1 to 66C4 are improved.
[0082]
FIG. 13A shows an original waveform of a digital audio signal supplied to the audio input terminal 42, for example, which includes harmonics and changes the harmonics. The pitch is measured with reference to zero cross points of this waveform, for example, s1, s2, s3..., And the pitch detection means 74 detects pitch data. Usually, since a sudden and temporary change in pitch does not occur in an audio signal, if such a sudden change occurs during pitch detection, it is determined that pitch detection has failed, and the pitch detection means 74 The detection data is not output. Specifically, the pitch c in FIG. 13A is not output as pitch data because the difference is too large compared to b. Similarly, since the difference d is too large compared to the same c, it is not output as pitch data. Eventually, the pitch detection means 74 outputs a, b, e,... As pitch data.
[0083]
FIG. 13 (b) shows the fundamental wave of the waveform of FIG. 13 (a), which will be described below using this fundamental wave. In order to change the pitch or formant, the waveform is read (cut out) from the ring memory configured in the RAM. For example, as shown in FIG. 5C, the waveform corresponds to the pitch data a from the zero cross point s1. The waveform is cut out, and then the waveform in the period corresponding to the pitch data b is cut out from the zero cross point s2. Since the detection of the pitch data c has failed, the pitch data b is read from the zero cross point s2 again. When the pitch detection is successful, the waveform corresponding to the pitch data e is cut out from the zero cross point s5 when the pitch detection is successful, and discontinuity occurs in the basic waveform.
[0084]
As shown in FIG. 13D, first, a waveform of a period corresponding to the pitch data a is cut out with the zero cross point s1 as a reference. After the cutout, s1 + a is calculated, and as a result, a waveform corresponding to the pitch data b is cut out based on s2. Next, the calculation of s2 + b is performed, and as a result, the detection of the pitch c has failed with reference to s3. Next, the calculation of s3 + b is performed, and as a result, the detection of the pitch d has failed with reference to ss4. Next, the calculation of ss4 + b is performed, and as a result, the detection of the pitch data e is successful with ss5 as a reference, and therefore a waveform corresponding to the pitch data e is cut out. Next, calculation of ss5 + e is performed, and as a result, a waveform of a period corresponding to the pitch data f is cut out with ss6 as a reference. As described above, the cut-out reference is updated according to the detection pitch, so that the discontinuity of the fundamental wave component does not occur.
[0085]
In this embodiment, after the waveform is cut out in this way, the pitch is changed or the formant is changed. Hereinafter, the cutout, pitch and formant changes will be described with reference to FIGS. 14 to 17. In order to simplify the description, only the pitch conversion means in one chorus channel will be described. In the above description, the pitch change coefficients supplied to the pitch conversion means have been described as P1 to P4. However, in the following description, the pitch change coefficients are expressed as PITCH-VR, and similarly, the formant change coefficients are expressed as F1 to F4. In the description of, it is expressed as FORMAT-VR.
[0086]
As in the first embodiment, the registers PITICH, SADRS, counter PHASE, PH1, PH2, registers WIDTH, LENGTH, ENV1, ENV2, W-RATE, SADRS1, SADRS2, ADRS1, ADRS2, and flag F are used. .
[0087]
When the sampling data (digital audio signal) is supplied to the input terminal 42, the sampling data is written to the write address IBUF-PTR of the ring memory configured in the RAM (not shown) as in step S2 of FIG. Step S2a). However, the write address IBUF-PTR is incremented by one every time sampling data is supplied to the input terminal 42. Next, pitch detection processing is performed in the same manner as in step S4 (step S4a), and it is determined whether pitch detection has been performed in the same manner as in step S6 (step S6a), and the pitch information PITCH is updated with the detection result (step S8a). ). In step S8, in addition to the update of the pitch information PITCH, the cut-out address SADRS is also updated to the zero cross position obtained by the pitch detection. However, in step S8a, the SADRS is not updated, and step S100 is separately performed. Is done by. The update of the cut-out address in step S100 updates the cut-out reference according to the detection pitch as described with reference to FIG. 13, and details will be described later. Further, step S100 is executed also when the pitch is not detected in step S4a.
[0088]
Subsequent to step S100, the values of the registers PHASE, PH1, and PH2 are each incremented by one as in step S10 (step S10a). Next, as in step S12, the values of the registers PHASE and WIDTH are compared (step S12a). If the PHASE value does not reach the WIDTH value, the process proceeds to a waveform reading process in step S32a described later. If the value of PHASE has reached the value of WIDTH, the value of register PHASE is set to 0 (step S14a) as in step S14.
[0089]
Next, the pitch change coefficient PITCH-VR, formant change coefficient FORMAT-VR, and reproduction pitch information SPITCH are input from the chorus control means 58 (step S16a). In step S16, only the pitch change coefficient PITCH-VR and the formant change coefficient FORMAT-VR set by the operator are input, whereas in step S16a, the pitch change set in accordance with the operation of the operator 46 is input. In addition to the coefficient PITCH-VR and the formant change coefficient FORMAT-VR, reproduction pitch information SPITCH based on the MIDI signal from the karaoke performance unit 40 or an external device is input.
[0090]
In order to determine a new WIDTH value (corresponding to the reproduction pitch information described in the claims), the reproduction pitch information SPITCH is multiplied by the pitch change coefficient PITCH-VR, and this is stored in the register WIDTH. In order to determine the period of the envelope of the first waveform and the envelope of the second waveform, the value of PITCH is divided by the formant coefficient FORMAT-VR and stored in the register LENGTH (step S18a). In step S18, the detected pitch information PITCH is multiplied by PITCH-VR, whereas in step S18a, SPITCH and PITCH-VR are multiplied.
[0091]
Next, as shown in FIG. 15, the value of the register LENGTH is compared with a period P of 100 Hz, for example (step S102). If the LENGTH value is larger than P, the LENGTH value is changed to P (step S104). These steps S102 and S104 are for preventing a large delay from being added even if a digital audio signal having a low frequency is input. Specifically, even when a digital audio signal having a frequency lower than 100 Hz is input, the delay is limited to a delay time of 10 milliseconds of 100 Hz.
[0092]
Subsequent to step S104 or when the LENGTH value is smaller than P in step S102, the LENGTH value is compared with the WIDTH value as in step S20 (step S20a), and the LENGTH value is smaller than the WIDTH value. If it is larger, the LENGTH value is set to the WIDTH value (step S22a) as in step S22.
[0093]
If the LENGTH value is less than or equal to the WIDTH value, the process of step S24a is performed without performing the process of step S22a. Further, even after step S22a, the process of step S24a is performed. In the process of step S24a, the reciprocal of the LENGTH value is obtained and stored in the register W-RATE, as in step S24. The value of W-RATE is used to increment the value of registers WINDOW1 and WINDOW2. In step S24a, the value of the flag F is inverted. Since step S24a is performed when the PHASE value becomes equal to or greater than the WIDTH value in step S12a, the inversion of the flag F is also performed when the PHASE value becomes equal to or greater than the WIDTH value.
[0094]
Next, similarly to step S26, the value of the flag F is compared with 0 (step S26a), and it is determined whether the flag F is 1 or -1. If the value of the flag F is 1, the registers PH1 and WINDOW1 are set to 0 as in step S28, and the SADRS value determined in step S100 is stored in the cut-out start address register SADRS1 (step S28a).
[0095]
If the value of the flag F is −1, as in step S30, the registers PH2 and WINDOW2 are set to 0, and the value of the register SADRS determined in step S100 is extracted and stored in the start address register SADRS2 (step S30a). .
[0096]
Subsequent to step S28a or 30a, or when it is determined in step S12a that the value of the register PHASE has not reached the value of WIDTH, the same waveform reading process as in step S36 is performed (step S32a). Since this waveform readout process is the same as that shown in FIG. 3, a detailed description thereof will be omitted. Then, the waveform subjected to the waveform reading process is sent out in the same manner as in step S64 (step S64a).
[0097]
For example, when SPITCH = PITCH and PITCH-VR and FORMAT-VR are each 1, an output waveform similar to that in FIG. 4 is obtained. In addition, when SPITCH = PITCH, PITCH> 1, and FORMAT-VR = 1, an output waveform similar to that in FIG. 5 is obtained. Further, when SPITCH = PITCH, PITCH-VR = 1, and FORMAT-VR> 1, a waveform similar to that in FIG. 6 is obtained.
[0098]
When PITCH = SPITCH, PITCH-VR <1, and FORMAT-VR = 1, in FIG. 5, the periods of 2 * WIDTH, ENV1, and ENV2 are equal, and levels 0 in FIGS. The part of disappears. Similarly, when PITCH = SPITCH, PITCH-VR = 1, and FORMAT-VR <1, in FIG. 5, the periods of 2 * WIDTH, ENV1, and ENV2 are equal to each other, and the levels shown in FIGS. The 0 part disappears.
[0099]
Next, with reference to FIG. 16 and FIG. 17, the update process of the cut-out address in step S100 will be described. In step S2a, the ring memory in which the sampling data is written has addresses 0 to FFF (hex), and the write address IBUF-PTR changes like a sawtooth wave as shown by a solid line in FIG. That is, when the address FFF (hex) is reached, the next address becomes 0.
[0100]
The update processing of the cut-out address is a premise for setting the read addresses ADRS1 and ADRS2 that are the reference when cutting out the sampling data in the waveform read processing, and these addresses ADRS1 and ADRS2 overtake the write address IBUF-PTR. There is no way. Therefore, first, the distance (address difference) DELAY at which the read address does not overtake the current write address IBUF-PTR is subtracted from the current write address IBUF-PTR, and the AND of (1000 (hex) -1). The result obtained is stored as a reference read address ADRS (step S106). The reason for taking this AND is to make the above subtraction value within the address range 0 of the ring memory within FFF (hex). This DELAY is set equal to the value of P used in step S102 of FIG.
[0101]
Next, this ADRS is compared with the current cut-out address SADRS (step S108). If SADRS is larger than ADRS, 1000 (hex) is added to ADRS (step S110). This is because, for example, as shown in FIG. 17, ADRS may be temporarily smaller than SADRS, and if it is left as it is, it adversely affects the subsequent processing, and this is to be corrected.
[0102]
Subsequent to step S110 or when it is determined in step S108 that ADRS is larger than SADRS, ADRS is compared with (SADRS + PITCH) (step S112). This PITCH is updated in step S8a. When ADRS ≧ (SADRS + PITCH), that is, when a difference greater than or equal to PITCH occurs between ADRS and SADRS, the calculation of (SADRS + PITCH) & (1000 (hex) −1) is performed, and the result of SADRS is calculated. The value is updated (step S114). If it is determined in step S112 that ADRS <(SADRS + PITCH), step S114 is not executed. As a result, the cut-out reference SADRS is updated according to the detected pitch PITCH as shown in FIG. 13D, and the SADRS changes by PITCH as indicated by the solid line shown in the staircase pattern in FIG.
[0103]
Next, the operation element 46 will be described with reference to FIG. The controller 46 includes a MODE switch 80 for setting a karaoke mode for adding a chorus in karaoke performance and a choir mode for adding a chorus by a MIDI signal from an external device. The switch 80 is a switch for switching between two on / off states. When the switch 80 is on, the chore mode is set. When the switch 80 is off, the chorus mode is set. When the switch is on, the display means provided in the switch 80, for example, the LED 82 is turned on, and when it is turned off, the LED 82 is turned off.
[0104]
The operation unit 46 is provided with a PITCH switch 84, a FORMAT switch 88, an EFFECT switch 92, and a level operation unit 96. These are for setting various parameters in the pitch converting means 66C1 to 66C4, delay means 68C1 to 68C4, and multipliers 62C1 to 62C4 of the chorus channels 60C1 to 60C4 of the chorus effect adding means 56. Note that it is determined by operating the CH-SELECT switch 100 which chorus channel the various parameters are set to. The CH-SELECT switch 100 includes four switches 100a to 100d corresponding to the respective chorus channels 60C1 to 60C4, and the switches 100a to 100d are provided with LEDs 101a to 101d, respectively.
[0105]
The PITCH switch 84 includes a -1OCT switch 84a, a NORMAL switch 84b, and a + 1OCT switch 84c. The PITCH switch 84 is for changing the tone quality (tone color) of the chorus audio signal to be added by pitch conversion. More specifically, the pitch of the chorus audio signal is changed from the karaoke performance unit 40 or an external device. This is for instructing whether to output with the pitch specified by the MIDI signal, or to raise or lower the pitch in octave units.
[0106]
The -1OCT switch 84a is a mode for reproducing a chorus voice signal one octave lower than the pitch specified by the MIDI signal. For example, when a female voice is inputted, a male chorus is added. To do. The NORMAL switch 84c is operated when adding a chorus while keeping the pitch represented by the MIDI signal. The + 1OCT switch 84c is operated when the reproduced chorus audio signal is set to an octave higher than the pitch specified by the MIDI signal. For example, when a male voice is input, a female chorus is added. Operated.
[0107]
The three switches 84a to 84c are provided with LEDs 86a to 86c, respectively, and among these LEDs 86a to 86c, the LED provided for the operated switch is lit. For example, when the switch 84b is operated in the state where the switch 84a is already operated, the operation by the switch 84b is effective for these three switches 84a to 84c. Thus, the operation performed later is configured to be prioritized. Note that the CH-SELECT switch 100 is similarly configured to give priority to subsequent operations.
[0108]
The FORMAT switch 88 is for changing the sound quality of the chorus audio signal by changing the formant of the input audio signal, and includes a LOW switch 88a, a NORMAL switch 88b, and a HIGH switch 88c. The LOW switch 88a is operated to change the formant of the chorus audio signal to a thicker voice quality than the input audio signal. The NORMAL switch 88b is operated when the formant of the chorus audio signal is the same as the formant of the input audio signal. The HIGH switch 88c is operated when the formant is raised to make a fine voice quality. These switches 88a to 88c are also provided with LEDs 90a to 90c, respectively, and the LED provided on the operated one of the switches 88a to 88c is lit. These three switches 86a to 86c are also configured so that the operation performed later becomes effective.
[0109]
Examples of operations of the PITCH switch 84 and the FORMAT switch 88 include the following. If the input audio signal is male and the chorus pitch information specified by the MIDI signal is in the male range, the + 1OCT switch 84c in the PITCH switch 84 is operated to raise the chorus audio signal by one octave, and the FORMAT The HIGH switch 88c in the switch 88 is operated to increase the formant of the chorus audio signal. In this case, even if a male audio signal is input, the chorus audio signal is synthesized in the female sound range with the female formant.
[0110]
The effect parameters e1 to e4 and the delay information D1 to D4 supplied to the pitch converting means 66c1 to 66c4 and the delay means 68c1 to 68c4, respectively, are illustrated in advance for karaoke and choir, for example, with four types having different contents. The EFFECT switch 92 is used for designating which one of them is to be used, and has four switches 92a to 92d provided for each type. Yes. These switches 92a to 92d are also provided with LEDs 94a to 94d, respectively, and the LED corresponding to the operated switch is lit. These switches 92a to 92d are also configured with priority on the post-operation.
[0111]
The level operation unit 96 includes an up operation unit 102, a down operation unit 104, and an 8-segment display unit 106. The level operation unit 96 includes a display unit 106. The displayed numerical value is increased or decreased to set the multiplier level of the channel designated by the CH-SELECT switch 100 among the multipliers 62c1 to 62c4 to an appropriate value.
[0112]
Further, the operation element 46 is provided with a DELAY switch 98 for designating whether the delay means 68c1 to 68c4 are turned on or off in common to the respective chorus channels 60c1 to 60c4. The DELAY switch 98 is set to 1 when it is on, 0 when it is off, and the LED 108 is lit when it is on.
[0113]
Further, for example, four kinds of effect parameters having different contents for the effect adding means 64M are stored in karaoke and chore preset memories (not shown), respectively, for selecting which one to use. In addition, the operation element 46 is provided with an EFFECT M switch 110. This EFFECT M switch 110 has four switches 110a to 110d corresponding to the above four types, and is set in the effect adding means 64M as will be described later by operating the switch corresponding to the type to be used. Is done. Whether to use karaoke or choir depends on the operation of the MODE switch 80. These switches 110a to 110d are also provided with LEDs 112a to 112d, and those provided for the operated switches are lit. These switches are also configured to give priority to post-operation.
[0114]
Similarly, for example, four types of effect parameters having different contents for the effect adding means 64C are stored in a preset memory (not shown) for karaoke and choir, in order to select which one to use. The operation element 46 is provided with an EFFECT C switch 114. The EFFECT C switch 114 has four switches 114a to 114d corresponding to the above four types, and is designated as the effect adding means 64C as described later by operating the switch corresponding to the type to be used. You can set parameters of the specified type. Whether to use the preset memory for karaoke or choir depends on the operation of the MODE switch 80. These switches 114a to 114d are also provided with LEDs 116a to 116d, and those provided for the operated switches are lit. These switches are also configured to give priority to post-operation.
[0115]
A direct level operator 118 is also provided on the operator 46. This is for setting the level information of the multiplier 62M of the main channel 60M, and has an up operator 120, a down operator 122, and an 8-segment display unit 124. The numerical value displayed on the display unit 124 is increased or decreased according to the operation, and the level of the multiplier 62M is set to an appropriate value.
[0116]
Next, processing performed by the chorus control means 58 will be described with reference to FIG. First, the setting state of the MODE switch 80 input from the operator control unit 54 is determined (step S116). As will be described later, each time the operating element control unit 54 operates the operating element 46 at a predetermined cycle to detect which operating element is being operated, the chorus control means 58 performs an operating element process described later. As a result, the setting state of the MODE switch 80 is input.
[0117]
If it is determined that the karaoke mode is selected, a MIDI input signal is set to be input from the karaoke performance unit 40, a preset memory is set for karaoke, and a channel memory to be described later is set for karaoke (step S118). Subsequently, karaoke mode parameter processing, which will be described later, is performed (step S120). Further, it is determined whether or not the mode change has been changed (step S122). This is also performed according to the state of the MODE switch 80 obtained as a result of the operator processing. If the mode has not been changed, step S122 is repeated, and if the mode has been changed, the process returns to step S116.
[0118]
If it is determined in step S116 that the mode is the choir mode, the MIDI signal is set to be input from an external device, the preset memory is set for the choir, and the channel memory described later is set for the choir (step). S124). Next, a parameter process for a chore mode described later is performed (step S126). Then, it is determined whether the mode change has been changed (step S128). This is also performed according to the state of the MODE switch 80 obtained as a result of the operator processing. If the mode has not been changed, step S128 is repeated. If the mode has been changed, the process returns to step S116.
[0119]
Next, the operator processing performed by the chorus control means 58 will be described with reference to FIGS. As is well known, the operator control means 54 scans the operator 46 every predetermined period, and supplies information on which operator is being operated to the chorus control means 58. The chorus control means 58 performs a manipulator process each time this information is input. As shown in FIG. 21, the operator control means 54 has a register group 126, a channel memory 128 for karaoke mode, and a channel memory 130 for choir mode.
[0120]
The register group 126 corresponds to each of the chorus channels 60C1 to 60C4, and the pitch change coefficients P1 to P4 (corresponding to PITCH-VR) set by the operation of the PITCH switch 84 and the formant change set by the operation of the FORMAT switch 88 are set. Coefficients F1 to F4 (corresponding to FORMAT-VR), level information L1 to L4 to the multipliers 62c1 to 62c4 set by operating the level operator 96, and effect parameters e1 to e4 set by operating the EFFECT switch 92, respectively. And registers for storing delay information D1 to D4, respectively. Further, it has a register for storing ON / OFF information set by operating the DELAY switch 98. Furthermore, a register for storing the effect parameter EM (= EM1 to EMm) set by the operation of the EFFECT M switch 110, and a register for storing the effect parameter EC (= EC1 to ECm) set by the operation of the EFFECT switch C And a register for storing level information L0 for the multiplier 62M of the main channel.
[0121]
The karaoke channel memory 128 stores pitch change coefficients P1 to P4, formant change coefficients F1 to F4, level information L1 to L4, effect information e1 to e4, and delay information D1 to D4 set in the register group 126 for the karaoke mode. , For memorizing. Further, the channel memory 130 for the choir includes pitch change coefficients P1 to P4, formant change coefficients F1 to F4, level information L1 to L4, effect information e1 to e4, delay information D1 to D4 set in the register group 126 for the choir mode. Is for memorizing.
[0122]
Whether to use the karaoke channel memory 128 or the choir channel memory 130 is determined in step S118 or S124. Further, although the preset memory described above is used, it is determined in step S118 or S124 whether to use a chorus memory or a choir memory. In this process, the flag Flg ON for storing the previous operation state of the CH-SELECT switch 100, Flg M for determining whether the writing to the channel memory is completed, and the CH-SELECT switch 100 Use a T-counter to count the time being operated.
[0123]
First, the operation state of each operation element 46 supplied from the operation element control means 54 is detected, set to the corresponding one in the register group 126, and the display means corresponding to the operation element being operated is displayed (step S130). ). However, the operation states of the PITCH switch 84, the FORMAT switch 88, the EFFECT switch 92, and the LEVEL operator 96 are stored in a register corresponding to the channel selected by the CH-SELECT switch 100. In the PITCH switch 84, 0.5 is set when the -1OCT switch 84a is operated, 1 is set when the NORMAL switch 84b is operated, and 2.0 is set when the + 1OCT switch 84c is operated. The pitch change coefficient register P of the selected channel is stored. Similarly, the FORMAT switch 88 is 0.5 when the LOW switch 88a is operated, 1 when the NORMAL switch 88b is operated, and 2.0 when the HIGH switch 88c is operated. It is stored in the register F of the selected channel.
[0124]
Similarly, in the case of the level operator 96, the numerical values set by the operation of the up operator 102 and the down operator 104 are stored in the register L of the selected channel. In the case of the EFFECT switch 92, the effect parameter and the delay information corresponding to the operated switch are stored in the parameter register e and the delay register D of the selected channel, respectively. In FIG. 21, the contents e1 to e4 of the parameter register e of each channel are described in plural as e11 and e12, respectively. Actually, these parameters e1 to e4 are each composed of a plurality of parameters. It shows that.
[0125]
The operation result of the DELAY switch 98 is stored in the DELAY register. In the EFFECT M register and the EFFECT C register, effect parameters corresponding to the operated one of the EFFECT M switch 110 and the EFFECT C switch 114 are stored.
[0126]
Further, the DIRECT LEVEL register stores a numerical value set by the DIRECT LEVEL operator 118.
[0127]
Next, it is determined whether the CH-SELECT switch 100 detected in step S130 is on (step S132). If it is on, it is determined whether the Flg ON flag is 0 (step S134). If the Flg ON flag is 0, the Flg ON flag is set to 1 and the Flg M flag is set to 1 (step S136). Then, the T counter is reset (step S138), and this process ends.
[0128]
If it is determined in step S134 that the Flg ON flag is not 0, it is determined whether the Flg M flag is 0 (step S140). If the Flg M flag is 0, this process is terminated. If the Flg ON flag is not 0, the count value of the T counter is incremented by 1 (step S142). Then, it is determined whether or not the counted value corresponds to 4 seconds or more (step S144). If it is not 4 seconds or more, this process is terminated. If it is 4 seconds or longer, the contents of each register of the channel selected by the CH-SELECT switch are stored in the corresponding channel area of the channel memory (step S146), and the Flg M flag is set to 0 (step S148). This process ends.
[0129]
For example, in the karaoke mode, a pitch change coefficient, a formant change coefficient, an effect parameter, level information, and delay information are set corresponding to channel 1, and the switch 94a corresponding to channel 1 of the CH-SELECT switch 100 is operated for 4 seconds or more. Then, the above values are stored in the storage area for the channel 1 pitch change coefficient, formant change coefficient, effect parameter, level information, and delay information in the channel memory 128 for karaoke.
[0130]
If it is determined in step S132 that the CH-SELECT switch 100 is not on, it is determined whether the Flg ON flag is 1 (step S150). If not 1, the process ends. If it is 1, the Flg ON flag is set to 0 (step S152), and it is determined whether the Flg M flag is 1 (step S154). If the Flg M flag is not 1, this process is terminated.
[0131]
If Flg M is 1, the contents of the channel corresponding to the operated CH-SELECT switch 100 in the channel memory are stored in the register group 126 and displayed on the operator 46 (step S156). For example, in channel mode 1, when channel 1 is selected and channel 101 switch 101a of CH-SELECT switch 100 has been operated for less than 4 seconds, it is stored in channel 1 of channel channel memory 130. The stored contents are stored in the corresponding registers of the register group 126, and the LEDs provided on the controls corresponding to these registers are lit. Following this, the Flg M flag is set to 0 (step S158), and this process ends.
[0132]
Next, with reference to FIG. 22 and FIG. 23, a process performed by the chorus control unit 58 every time it receives a MIDI signal from the karaoke performance unit 40 or an external device will be described. The chorus control means 58 has a MIDI register as shown in FIG. 23, which has a tempo register 132 for storing the tempo, and a register for storing the chorus sound pitch information SPITCH for each channel. 134a to 134d, and registers 136a to 136d for storing chorus sound level information Lmidi.
[0133]
It is determined whether the input MIDI signal is a timing clock representing the performance tempo (step S160). If it is the timing clock, the tempo is measured from the timing clock and stored in the TEMPO register 132 of the MIDI register ( In step S162), this process ends. Note that what is stored is, for example, a product obtained by multiplying a measurement result by a predetermined coefficient and converting it into tempo information having a quarter note length. Since the period of the timing clock is measured, the measurement cannot be performed at the first timing clock input, and this process is terminated. In addition, tempo information may be transmitted as numerical data instead of the timing clock. In this case, step S160 is changed to determination of tempo information. In the case of tempo information, the tempo information is directly used as the TEMPO register 132. Remember me.
[0134]
If it is not a timing clock in step S160, it is determined whether or not the input MIDI signal is a note-on signal for instructing the production of a chorus sound (step S164). If it is a note-on signal, it has a key number that represents the pitch of the reproduced sound, a velocity that represents the level of the reproduced sound, and channel information that represents the reproduced channel. The key number is converted into chorus sound reproduction pitch information SPITCH by a conversion table that is not present, and stored in the reproduction pitch information register of the channel designated by the channel information (step S166).
[0135]
Next, the velocity is stored as level information Lmidi in the channel specified by the channel information in the level information register of the MIDI register (step S168), and the process is terminated.
[0136]
If it is determined in step S164 that it is not a note-on signal, it is determined whether it is a note-off signal (step S170). If it is a note-off signal, it has channel information indicating the channel to be note-off subsequently, so the value of the level information register of the channel specified by the channel information is set to 0 (step S172), and this processing is performed. finish.
[0137]
Next, the details of the karaoke mode parameter supply process in step S120 shown in FIG. 19 will be described with reference to FIG. This karaoke mode parameter supply process is executed each time a predetermined time elapses such that a parameter change due to an operation of the operation element 47 or the like is not unnatural as an effect change. Therefore, although not shown in the figure, when this process is executed, the value of the counter prepared in advance is advanced by 1. If the value is not a predetermined value, the process proceeds to step S122, and the value is set to a predetermined value. When the value is reached, the counter value is set to 0 and the following processing is executed. If the predetermined time is long, the response to the operation of the operation element 47 is delayed and unnatural, and if the predetermined time is short, a change in the effect cannot be perceived by humans. become.
[0138]
In this process, the parameter group of EFFECT M and EFFECT C and the DIRECT level are read from the register group shown in FIG. 21 and supplied to the effect adding means 64M and 64C and the multiplier 62M shown in FIG. 11 (step S174). ).
[0139]
Next, in order to designate a parameter for each chorus channel device, the value of the channel designation counter n is set to 1 (step S176). Next, the level information Lmidin of the channel designated by the counter n in the MIDI register shown in FIG. 23 is multiplied by the level information Ln of the channel designated by the counter n in the karaoke channel memory shown in FIG. The multiplication result is supplied as level information to the multipliers 62c1 to 62c4 designated by the counter n, and further, the tempo information of the MIDI register, the DELAY of the register group 126, and the counter n of the karaoke channel memory 128 are designated by the counter n. The channel delay information Dn is read out, multiplied by these, and supplied as the delay time to the delay means 68c1 to 68c4 designated by the counter (step S178). Since the delay information Dn is a value representing the number of delay times corresponding to quarter notes, the actual delay time can be determined by multiplying this by TEMPO. Since DELAY is 1 when delaying and 0 when not delaying, the actual delay time is set in the delay means when delaying, and 0 is set as the delay means when not delaying. Is done.
[0140]
Next, the pitch change coefficient Pn, formant change coefficient Fn, and effect parameter en are read from the channel specified by the counter n in the channel memory 128, and set to the one specified by the counter n among the pitch conversion means 66c1 to 66c4. (Step S180). The chorus pitch information specified by the counter n in the MIDI register is set to the pitch conversion means 66c1 to 66c4 specified by the counter n (step S182). Then, it is determined whether the value of the counter n is 4 or more (step S184). If it is 4 or more, this process is terminated. If it is not 4 or more, the value of the counter n is advanced by 1 (step S186). The process returns to S178.
[0141]
Note that the chore mode parameter supply process in step S126 will not be described in detail because the channel memory used is only the channel memory 130 for the choir.
[0142]
Next, a state where the chorus effect is added in the karaoke mode will be described with reference to FIG. In order to simplify the description, it is assumed that only the chorus channel 60c1 is used, and level information 0 and delay information 0 are supplied to the other channels, respectively. Further, it is assumed that a delay time for one quarter note is set in the delay means 68c1, and level information other than 0 is set in the multiplier 62c1. In addition, it is assumed that pitch information of the pitch E is input to the pitch conversion means 66c1 as reproduction pitch information SP1 corresponding to the input MIDI signal. Then, when a pitch signal “CAME ON” having a pitch C is input to the pitch converting means 66c1, as shown in FIG. 10C, a voice signal “CAMEON” having a pitch E corresponding to the reproduction pitch information SP1. Is output. This output is output from the pitch converting means 66c1 as shown in FIG. Then, the delay means 68c1 delays the above delay time, and the effect adding means 64c adds an effect. On the other hand, the effect is also added to the input audio signal by the effect adding means 64M, added by the adder 72, and output as a chorus output signal as shown in FIG. A sound like the score shown is generated.
[0143]
As the delay means 68c1 to 68c4, the one which inputs 0 when not delaying by the DELAY switch 98 is used, but when not delaying, the delay means 68n is bypassed by the delay switch 69n as shown in FIG. In the case of using a delay means, the level information Ln is not controlled, but the delay on / off may be controlled by a delay switch provided for each channel instead of the DELAY switch 98. .
[0144]
Next, a state where the chorus effect in the choir mode is added will be described. This choir mode adds the effect of chorusing with a large number of people. In the choir mode, the effect adding means 64c shown in FIG. 11 functions as the pseudo chorus effect adding device shown in FIG. The parameter is set. Further, the parameters may be set so that the effect adding means provided in each of the pitch converting means 66c1 to 66c4 functions as the chorus effect adding device shown in FIG.
[0145]
By operating the FORMAT switch 88, for example, the voice quality of the chorus channels 1 and 2 is set to male, and the voice quality of the chorus channels 3 and 4 is set to female. Note that if the delay information set in each delay means 68c1 to 68c4 is slightly shifted and set, the singing timing of each chorus channel is slightly shifted, and it becomes natural as if a human is singing. . In addition, when the effect adding means 64c or the effect adding means of each of the pitch converting means 66c1 to 66c4 is used as the pseudo chorus effect device shown in FIG. It may not be necessary to cause a delay at 68c4.
[0146]
In the above-described embodiment, the parameters PITCH-VR and FORMAT-VR are set by operating the PITCH switch 84 and the FORMAT switch 88 of the operator 46, but both can be set by operating one switch.
[0147]
For example, as shown in FIG. 27, a CHARACTER switch 124 including a BIGMAN switch 124a, a NORMAL switch 124b, and a PRETTYGIRL switch 124c is provided in place of the PITCH switch 84 and the FORMAT switch 88 in the operation unit 46, and any of these switches is provided. If operated, PITCH-VR and FORMAT-VR may be set.
[0148]
In this case, when the BIGMAN switch 124a is operated in step S130 shown in FIG. 20, 0.5 is specified as PITCH-VR, 0.7 is specified as FORMAT-VR, and the CH-SELECT switch 100 of the register group 126 is designated. Stored in the P and F areas of the selected channel. Similarly, when the NORMAL switch 124b is operated, 1 is stored as PITCH-VR and 1 is stored as FORMAT-VR in the P and F areas of the channel specified by the CH-SELECT switch 100 of the register group 126, respectively. Is done. When the PRETTYGIRL switch 124c is operated, 2.0 is set as PITCH-VR, 1.5 is set as FORMAT-VR, and the P and F areas of the channel designated by the CH-SELECT switch 100 of the register group 126 are respectively set. Remembered.
[0149]
These switches 124a to 124c are also provided with LEDs 126a to 126c, respectively, and those corresponding to the operated switches are lit. Further, these switches 124a to 124c are also configured with priority on the post-operation.
[0150]
In the second embodiment, PITCH-VR and FORMAT-VR are set by selecting one of predetermined values by operating the switches 84 and 88. However, instead of these switches, potentiometers are provided, An arbitrary value may be set by an operation. The delay information for determining the delay time set in each of the delay means 68C1 to 68C4 is determined by the setting of the EFFECT switch 92, but may be set by a MIDI signal from the karaoke performance unit 40 or an external device, for example. In this case, since the delay information can be changed with the passage of time, the delay time can also be changed with time, and an effective performance can be performed. Similarly, the effect adding means included in the pitch converting means 66C1 to 66C4 and the effect parameters respectively set in the effect adding means 64M and 64C may be set by the MIDI signal. Also in this case, the effect added by these effect adding means changes with the passage of time.
[0151]
【The invention's effect】
As described above, according to the present invention, a plurality of periods of speech segments are cut out from an audio signal, a predetermined envelope is added, and then synthesized at a period corresponding to the reproduction pitch. The pitch conversion is performed while maintaining the formant of the audio signal.
[Brief description of the drawings]
FIG. 1A is a block diagram showing a first embodiment of an effect adding device according to the present invention, FIG. 1B is a block diagram showing an embodiment of a DSP in the block diagram, and FIG. It is a block diagram which shows embodiment.
FIG. 2 is a flowchart showing a part of functions executed by a DSP in the embodiment.
FIG. 3 is a flowchart showing detailed contents of a waveform reading process in the flowchart of FIG. 2;
FIG. 4 is an operation explanatory diagram when the pitch change coefficient and the formant change coefficient are both 1 in the embodiment.
FIG. 5 is an operation explanatory diagram when the pitch change coefficient is made larger than 1 and the formant change coefficient is set to 1 in the same embodiment;
6 is an operation explanatory diagram when the pitch change coefficient is 1 and the formant change coefficient is larger than 1 in the embodiment. FIG.
FIG. 7 is a diagram showing a modulation waveform stored in a modulation waveform memory used in a modification of the embodiment.
FIG. 8 is a flowchart showing the detailed contents of a part of a waveform read process in a modification of the embodiment.
FIG. 9 is a flowchart showing the detailed contents of the remaining part of the waveform reading process in the modification of the embodiment.
FIG. 10 is a block diagram of a second embodiment of the present invention.
FIG. 11 is a block diagram of chorus effect adding means of the second embodiment.
12 is a block diagram of a pseudo chorus adding device used in the effect adding means of FIG. 11. FIG.
FIG. 13 is an explanatory diagram of waveform cutout in the second embodiment.
FIG. 14 is a flowchart for explaining a part of the operation of the chorus effect adding means of FIG. 11;
FIG. 15 is a flowchart for explaining another part of the operation of the chorus effect adding means of FIG. 11;
16 is a flowchart for explaining operations related to waveform cutting in the chorus effect adding means of FIG. 11. FIG.
17 is a diagram showing a change state of an address of a ring memory used for waveform extraction in the chorus effect adding means of FIG.
FIG. 18 is a diagram showing an operator used in the second embodiment.
FIG. 19 is a flowchart for explaining the operation of the chorus control means shown in FIG. 10;
20 is a flowchart illustrating an operation performed by the chorus control unit shown in FIG. 10 in accordance with the operation state of the operator.
FIG. 21 is a diagram showing a register and channel memory provided in the chorus control means.
FIG. 22 is a flowchart illustrating processing performed by a chorus control unit in response to an input of a MIDI signal.
FIG. 23 is a diagram showing a MIDI register included in the chorus control means.
24 is a flowchart illustrating karaoke mode parameter supply processing shown in FIG.
FIG. 25 is a diagram illustrating a state in which a chorus effect is added in the second embodiment.
FIG. 26 is a diagram showing another example of delay means used in the second embodiment.
FIG. 27 is a diagram illustrating another example of the operation element used in the second embodiment.
[Explanation of symbols]
8 DSP
12 RAM
20 46 operator
22 CPU
26 74 Pitch detection means
66C1 to 66C4 pitch conversion means

Claims (1)

A memory for storing an audio signal having a plurality of periods;
Pitch detecting means for detecting the pitch of the audio signal;
Reproduction pitch information storage means for storing reproduction pitch information for setting a reproduction pitch;
According to the pitch period detected by the pitch detection means, a predetermined two-cycle section of the audio signal stored in the memory is read at a reading speed set independently of the reproduction pitch information. Phoneme segmentation means for segmenting phonemes;
An envelope that gradually shifts from the minimum value to the maximum value with a length equal to the time required to read the phoneme segment is given to the extracted phoneme segment, and is stored in the reproduction pitch information storage means. Reproduction audio signal synthesis means for synthesizing in units of two periods based on the reproduction pitch information and generating a reproduction audio signal;
A pitch converter provided.

Images