JPS6127596A - Musical scale converter - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] Industrial Application Field The present invention is a device that outputs different pitches in response to the pitch of an input signal, and is suitable for use as a special effect generating device in the record, broadcasting, and movie industries, or as a sound source with different pitches. The present invention relates to a pitch converting device which is used as a pitch correction device, and also performs accompaniment in a mixing device such as a so-called karaoke.

Structure of a conventional example and its problems FIG. 1 shows a conventional example in which a digital memory 1 is used and the write clock Fo and the read clock f can be tracked. Analog signal S input to input terminal 2.

is converted into a digital signal 310 via a low-pass filter (hereinafter abbreviated as LPF) 3 and an analog-to-digital converter (hereinafter abbreviated as DC) 4. After being written into the digital memory (for example, RAM) 1 with the signal 5xoi and the write clock Fo, it is read out with the read clock Fv and a digital signal "20" is output.

The signal S2o is a digital-to-analog converter (hereinafter referred to as DAC).
An analog output signal S3o is outputted to the output terminal 7 via the LPF 6 and the LPF 6. The pitch change rate P in this case is determined by equation (1).

P=□・・・・・・・・・・・・(1)O This conventional example has the following problems.

■ When the pitch is changed, discontinuous noise occurs at the data junction.

FIG. 2 shows a conventional example in which the pitch is changed. Input signal So
When the pitch change rate P is set to 0.67, the signal length To is extended to Tl and the joining part N with the next data is
, which generates discontinuity noise.

■ Since the purpose is to convert pitches in all bands, it would be costly to process data using a microcomputer, etc.
Furthermore, data processing to completely eliminate noise is extremely expensive. That is, in the case of a music signal, the conceptual diagram is shown in FIG.
, F) There is a method of finely decomposing frequency elements from 1 to N, treating each as a single signal, performing phase matching, and adding the processed signals, but this method is difficult to realize in reality.

■ If the noise is generated randomly and impairs the musicality or the sense of distortion increases, it becomes impossible to identify the original instrument.

OBJECT OF THE INVENTION The object of the present invention is to provide a high-quality, high-sound quality, low-cost pitch conversion device that solves the above-mentioned problems. It is an object of the present invention to provide a pitch conversion device which can arbitrarily make the pitch of an output signal higher or lower without losing the natural feeling and musicality when compared with the pitch of an input signal.

Structure of the Invention The present invention includes an analog-to-digital converter, a first and second digital memory, and the first and second digital memories. The first digital memories respectively correspond to the second digital memories. a second digital-to-analog converter; first clock generation means for generating a fixed frequency clock for the analog-to-digital converter; a second clock generating means for generating a clock that can vary the first clock; memory control means for generating a signal for controlling the second digital memory; control means for controlling the output of the second digital-to-analog converter; and stepwise decreasing the output signal of the first digital-to-analog converter to zero within the same time period; means for increasing a device output signal stepwise from zero and mixing the two output signals; at least one analog narrowband signal amplifier for amplifying the narrowband signal of the analog signal; and at least one analog narrowband signal amplifier for amplifying the narrowband signal of the analog signal; The structure includes an arithmetic circuit that performs arithmetic operations based on the output of the amplifier, and switching means that switches the cross-fade time and cross-fade pitch based on the output of the arithmetic circuit, thereby changing the pitch of the output signal to that of the input signal. Compared to pitch, it can be set arbitrarily high or low without losing its natural feel and musicality.

DESCRIPTION OF EMBODIMENTS The present invention is based on the results of a psychological experiment showing that humans have difficulty detecting changing components of noise that occurs at the same timing as a fast-tempo song or rhythm.

That is, the present invention is a pitch conversion device that connects pitch-converted signals not uniquely at a cycle determined by memory capacity or a constant frequency division value from a clock, but according to rhythm.

An embodiment of the present invention will be described in detail below with reference to FIGS. 4 to 12. Figure 4 shows the unity of tone and the richness of sound in frequency modulation.
8-16o) are cited.

In Figure 4 (a) to (d), the vertical axis shows the number of modulations per second, the horizontal axis shows the width of the modulation, the small circle shows the limit of the unity of tone, and the broken line shows the richness of the sound. The large circle shows the amplitude and modulation (vibrato) of the vocalist, and the oval shows the vibrato of the violinist. In this way, when frequency modulation and amplitude modulation (vibrato) are applied to a pure tone, the limit point at which the tonality of the pure tone ends can be seen by changing the number of modulations and the frequency width of the modulation. According to Figure 4, as the modulation width increases, the number of modulations hardly changes, about 7 times (about 140m5).
ec). It was confirmed in a J1 physical experiment that even a continuous signal using a music signal has a value of about 140m5elO.

FIG. 6 is a block diagram of a pitch converter according to an embodiment of the present invention. In FIG. 6, the input signal So input from the input terminal 8 passes through the LPF 9 for digital conversion, and becomes the signal S1 with high frequency components cut off, and the ADC 10
i is converted into a digital signal S2. After the signal S2 is written to the first digital memory 11 at the write clock F0, the signal S2 is written to the first digital memory 11 at the write clock F0, and then at the read clock Fv determined by the pitch changeover switch 25 of the crossfade pitch conversion circuit 17 shown in FIG. (hereinafter referred to as a pinch conversion circuit) 17 (hereinafter referred to as a pinch conversion circuit) reads the data of the specified address string A1, and outputs a digital signal S5 from the first DAC 13. The digital signal S5 passes through a first amplifier (hereinafter referred to as first VCA) 15 whose amplification degree can be varied, and outputs a signal S7 based on the output signal C1 of the pitch conversion circuit 17.

Similarly, after the digital signal S21''j is written into the second digital memory 12, the pinch conversion circuit 17 outputs the data readout digital signal S6 for the specified address column rudder using the readout clock Fv.

The digital signal S6 is outputted as the signal S8 via the second VCA 16 by the signal C2 from the pitch conversion circuit 17. Signal S7 and signal S8 are added by adder 18 to become signal S9, which is output to output terminal 19.

In addition, the first DAC 13 and the second DAC 1
Cross-fade is the addition of a signal in which the output of one of the four outputs is gradually decreased from zero and a signal in which the output of the other is increased in steps from zero.

On the other hand, the input signal So has the high frequency band of the music signal cut by the LPF 29, becomes a low frequency signal S10, is amplified by the amplifier 31, is shaped into a rectangular wave S11, and is input to the comparator circuit 2o. On the other hand, the low frequency oscillator (O20) 28 is 1. e 7
Hz oscillation is performed, and the signal S12 is amplified by the amplifier 30, and the signal 813 is input to the comparison calculation circuit 2o as a reference signal. The comparison calculation circuit 20U calculates how many times the signal S11 is compared to the reference signal 813 by sampling at a fixed time, and outputs it as a signal C3 to the pinch conversion circuit.

This will be explained in detail below using FIG. 12. FIG. 12 shows a specific example of the comparison calculation circuit 2o, in which the number of pulses of the signal 813 (specifically, 100 The comparator 36 compares how many pulses of the signal S11 are counted while the pulse counter 34 counts the number of pulses), and according to the logic set in the ROM 35, if there is no distance between the input signal S11 and the input signal S11, a fixed pitch is reached. The information is sent to the pitch conversion circuit 17 as a signal C3.

Next, the pinch conversion circuit 17 will be explained in detail using FIG. 6. FIG. 6 is a configuration diagram of the pinch conversion circuit 17.

In FIG. 6, the oscillation frequency of the read oscillation circuit (08G) 24 is switched by the pitch changeover switch 26. Fifth
Figure 1. Address signals for writing and reading of the second digital memories 11 and 12 are the address strings 1. It becomes A2. Further, the output signal C3 of the comparison arithmetic circuit 2Q is arithmetic data that generates the signals CI, 02, and the arithmetic circuit 22 controls the cross-fade time and cross-fade pinch using the signals CI, C2.

Switch 27 is a switch that can only change the crossfade pinch, and can be used for fast and slow tempo songs.Specifically, in slow tempo Enka songs, discontinuities in sound are not noticeable, so the crossfade pitch can be widened and the tempo can be changed. For faster songs, you can set the crossfade pitch narrower.

The operation of this embodiment will be explained below using the signals shown in FIG. In FIG. 7, the horizontal axis is time and the vertical axis is voltage. Since the input signal S0 is written to the first digital memory 11 at the clock Fo and read out at the read clock F7, the relationship between the fixed part molecule 0 of the input signal and the input waveform equivalent part TI of the output signal, Fo, and FT is expressed by the following equation. expressed.

1F0 ToF.

The signal S3 read from the first digital memory 11 and the signal S4 read from the second digital memory 12 are the same signal, and the address signal M1. A phase shift occurs due to pitch conversion. The input signal So starts being input to the first digital memory 11 at time 1. And its output S
3 starts reading from the first digital memory 11 at time tb after a delay of time T2 for reading and writing from the first digital memory 11.

Because the pitch is converted, a large phase difference occurs from the original signal, so it is necessary to return to the write address at a constant period T1. The addresses of time - and tb are the same, and time t
o and t61d have the same address.

The signal 9S5 output from the first DAC 13 is a noise N2f due to a signal shift at time tb, and a noise N3 is generated at time t4. Read signal S4.

The signal S4 becomes an analog signal S6 via the second DAC 14. Noise N4 is generated as shown in FIG.

Noise N26. N3. The time at which N4 occurs has the following relationship.

T 1 = 2 ・ T4 jJI D A C13 (D output signal S5 [1st VC
A15 shapes the signal into signal S7. The first VCA 15 from the pinch conversion circuit 17 receives the control signal C1 (seventh
(shown by a dashed line in the figure), the signal S5 (broken line) is output as an output signal Sy (solid line).

The output signal S6 of the second DAC 14 is also the control signal C2.
is output as an output signal S8.

The output obtained by passing the output signals 87 and 88 through the adder 1B is the output signal S9. As can be seen in Figure 7, conventional N
A sudden signal change like that shown in does not occur.

At this time, the increasing and decreasing portions of the control signals C1 and C02, that is, the cross-fade time T3, is called a cross-fade time, and T4 is called a cross-fade pitch.

Although the circuit shown in Figure 5 configured as above uses two digital memories, it is also possible to use only one digital memory for time-sharing processing.
C and VC systems can also perform arithmetic processing on digital data to obtain the same effect.

・The crossfade pitch of the aforementioned “Psychology of Hearing (written by Soichi Kuroki)” is 140m! ! As mentioned above, it has to be more than eo, but in actual music signals containing various rhythms and pitches, it has been found that even more than 140mg6C causes discomfort. Through psychological experiments, it was discovered that there is a correlation between discomfort and each of the following factors. FIG. 8 is a diagram showing the relationship between tremolo ratio and discomfort. (Tremolo is a fixed-period amplitude modulation.) In addition, Fig. 8 shows 20 test subjects, and cross-fade time T4 =
It shows the results of a psychological experiment when the pitch was 40mS6c and the pitch change rate was 0.94. As for the music signal, a song played at a slow speed (1-60) was selected. As a result, the tremolo ratio is 0.4
When it exceeds 5, everyone recognizes the tremolo and feels uncomfortable.
Below 0.1, you won't notice the tremolo. In music signals, we recognize the tremolo and the lack of the V sound, leading to a sense of discomfort. Therefore, it can be seen that the tremolo ratio is preferably 0.3 or less.

FIG. 9 is a diagram showing the relationship between the crossfade pinch error rate, which is the error rate between the crossfade pitch and the rhythm, and discomfort, and this crossfade pitch error rate is defined as follows.

■n of the rhythm is selected to be a value close to the crossfade pinch (for example, one with a small error rate). The experimental conditions were the same as those shown in Figure 8, and as a result, the crossfade pitch error rate was 0.
If it's between 8 and 1.3, you won't feel any discomfort, but if it's outside that range, you'll suddenly feel discomfort. Therefore, it can be seen that a cross-fade pitch error rate of 0.77 to 1.3 is good.

FIG. 10 is a diagram showing the relationship between the crossfade time rhythm error rate, which is the error rate between the crossfade time and the rhythm, and discomfort.The crossfade time rhythm error rate is defined by 14n of the rhythm. However, the rhythm is 172 n ij 301
1! Set between IeO and csom5ec.

The reason why I set the rhythm V2n between 30 m5ec and 60 m81so is that the crossfade time is 30 m!
! It has been confirmed in another experiment that distortion can be heard as noise when the value is less than 60, and the upper limit was set to e o m5ec because this would uniquely determine the crossfade time.

Also, from another experiment, the difference between the product of crossfade time and pitch change rate and crossfade time is 40 mBej.

It has also been found that, depending on the music, the sound may become interrupted if the limit is exceeded. The music source is enka, and if the crossfade time rhythm error rate increases, the noise will increase at the beginning of the song and at the end of the song, making it difficult to sing. From this, it can be seen that the cross-fade time rhythm error rate is preferably 0.86 to 1.2.

FIG. 11 is a diagram showing the relationship between the ratio of the cross-fade time and the truncated data length shown by T6 in FIG. 7 and the actual recognition of discontinuity, which is determined as the truncated data length. In other words, discontinuity, or the state where a sound is omitted or the same sound is heard twice, is
It can be seen that the shorter the cross-fade time and the lower the recognition rate of discontinuity, the less. However, T3= 30 m5ec
However, it can be seen that it is sufficient to keep the discontinuity recognition rate (Tz) below 1.1 and the truncated data length not to exceed approximately 4QmB60. Conceptually, it is better to make the crossfade time longer than the length of the truncated data, but another factor is that the absolute length of the truncated data (T♀xTz) is within 33m19150 when the crossfade time is 30m5ec. ，
It can be seen that if the X fade time is 60 m5ec, discontinuity will be recognized unless it is within 48 m5ec. Since there are large differences depending on the instrument and preference, the data length is rounded down to 40m5e.
Better results can be obtained if it is within c.

In this way, in this example, it was discovered that there is a correlation between crossfade time, truncated data length, tremolo ratio, crossfade pitch and rhythm, and crossfade time and rhythm, and this was realized using a simple method. It is.

In addition, we have installed an arithmetic circuit that detects the low frequency components of the music signal, digitizes the deviation from the reference frequency, and then sets the crossfade time, truncated data length, tremolo ratio, and crossfade pitch to optimal values. It is a characteristic.

Effects of the Invention As described above, the present invention can provide a simple pitch conversion device with high quality and high sound quality, and can arbitrarily change the pitch of an output signal by making it higher or lower than the pitch of an input signal. It detects the rhythm of the music and automatically sets itself to provide the best natural sound quality.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of a conventional pitch conversion device, Fig. 2 is a schematic diagram of signals during pitch conversion, Fig. 3 is a conceptual diagram of another conventional pitch conversion device, and Fig. 4 is a diagram of the tone in frequency modulation. A diagram showing the relationship between unity and richness of sound. Figure 5 is a block diagram of a pitch converter according to an embodiment of the present invention. Figure 6 is a block diagram of the same pitch variable circuit. Figure 7 is a diagram showing the same pitch. Fig. 8 is a signal schematic diagram at the time of conversion, which shows the relationship between tremolo ratio and discomfort.
Figure 9 shows the relationship between crossfade pitch error rate and discomfort, Figure 10 shows the relationship between crossfade time rhythm error rate and discomfort, and Figure 11 shows the discontinuity recognition rate (
Figure 12 is a diagram showing the relationship between Tz) and discontinuity recognition, and Figure 12 is a configuration diagram of the comparison arithmetic circuit.
Digital memory, 12... Second digital memory, 13... First digital-to-analog converter, 14
...Second digital-to-analog converter, 16...
...First amplifier, 16...Second amplifier, 1st...Cross fade pitch conversion circuit, 18...
... Adder, 2o... Comparison calculation circuit, 22
... Arithmetic circuit, 25 ... Pitch selection switch, 2 ... Switch, 29 ... Low pass filter, 31 ... Amplifier. Name of agent: Patent attorney Toshio Nakao and 1 other person @1
Figure 3 @ Figure 4 ((L) V, 1 Modulation Radiation U) α] @ Figure 6 Figure 8 Ll-to-D ratio (/V=20) Figure 9

Claims (2)

[Claims] (1) an analog-to-digital converter for digitally converting an analog signal; first and second digital memories for storing the output of the analog-to-digital converter;
The first and second digital memories respectively correspond to the second digital memory.
a digital-to-analog converter; a first clock generating means for generating a fixed frequency clock for the analog-to-digital converter;
a second clock generating means for generating a clock that can change the frequency for analog conversion; a memory control means for generating a signal for controlling the first and second digital memories; and the first and second digital-to-analog memories. means for controlling the output of the converter; and means for controlling the output of the converter within the same time.
Means for stepwise decreasing the output of the analog converter to zero, increasing the output of the second digital-to-analog converter stepwise from zero, and mixing the two output signals (hereinafter referred to as cross-fade). at least one analog narrowband signal amplifier that amplifies the narrowband signal of the analog signal; an arithmetic circuit that performs an arithmetic operation based on the output of the analog narrowband signal amplifier; and a cross-fade time determined by the output of the arithmetic circuit. A pitch converter comprising a means for switching between (time required to increase a signal stepwise from zero or decrease a signal stepwise from zero) and a crossfade pitch (interval at which a crossfade is performed) . (2) The pitch conversion device according to claim 1, further comprising a second means for switching the cross-fade pitch.