JP3668459B2 - Sound correction apparatus and sound correction method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a sound correction device and a sound correction method, and more particularly to recording and correction for removing noise from a sound signal.
[0002]
[Prior art]
Conventionally, as such a sound correction device and sound correction method, taking an electronic musical instrument as an example, an actual sound, for example, an actual piano sound, is collected by a microphone instead of an artificially created sound. There is a sample that is sampled and digitized and recorded, and the recorded sound is reproduced and used as a sound of an electronic musical instrument. Such sampling recording is also possible for sounds of musical instruments other than piano, sounds other than musical instruments, and animal calls.
[0003]
[Problems to be solved by the invention]
However, in such a sampled and recorded sound, in addition to the original data of the sound, data such as noise may be mixed, and noise may be mixed particularly immediately before the start point of the sound.
[0004]
The present invention has been made to solve the above-described problems, and a first object of the present invention is to realize a beautiful sound by removing the noise at the beginning of such a sound.
[0005]
Further, when noise at the head portion of the sound is removed in this way, as shown in FIG. 6 (2), a discontinuous point where the level of the head point of the sound data suddenly rises from “0” to a certain value. If this sound data is reproduced as it is, noise is generated at this discontinuous point. Of course, in removing noise, it is sufficient to match the “0” cross point of the sound data, but such processing is difficult. In addition, when the noise removal area is set to the “0” cross point of the sound data, noise often remains and cannot be removed cleanly.
[0006]
The present invention has been made in order to solve the above-mentioned problems, and secondly, it is possible to realize a beautiful sound by removing new noise generated when noise at the head of the sound is removed as described above. The purpose is.
[0007]
[Means for Solving the Problems]
In order to achieve each of the above objects, the present invention clears the head portion of the sampled sound data, and further increases the value from level “0” to the sound data that clears the head portion. The correction data reaching “1” is synthesized.
[0008]
As a result, the noise at the beginning of the recorded sound data is removed, the beginning point of the sound data gradually increases from level “0”, and the discontinuous points after the noise removal are also removed.
BEST MODE FOR CARRYING OUT THE INVENTION
[0009]
1. Overall Circuit FIG. 1 shows an overall circuit of a sound correction recording apparatus. In the microphone 1, the sound to be recorded is collected and converted into an electric signal, that is, a sound signal. The sound signal SS is amplified by the amplifier 2, converted to digital data by the A / D converter 3, stored in the latch 4, and sequentially written in the first waveform RAM 5.
[0010]
The sampling signal oscillator 8 is a programmable type oscillator, generates a sampling signal SP having a frequency corresponding to set data from the CPU 10, and supplies the sampling signal SP to the latch 4 as a latch signal. Thereby, sampling of the sound signal SS taken in with the period of the sampling signal SP is performed. The sampling signal oscillator 8 may use a programmable type frequency divider, a timer, or the like.
[0011]
The sampling signal SP from the sampling signal oscillator 8 is given as a read / write command signal R / W to the first waveform RAM 5, and writing is performed in synchronization with the sampling signal SP. The sampling signal SP is input to the address counter 9, and the address data in the first waveform RAM 5 is counted. The sampling signal oscillator 8 and the address counter 9 are reset by the CPU 10 or switched between enable / disable.
[0012]
In the second waveform RAM 6, the noise data NS at the head portion is removed from the sound data SD recorded in the first waveform RAM 5, and the envelope data EN rising gradually from the level “0” is multiplied and synthesized, and corrected. The corrected sound data MS is stored. The control RAM 7 stores the envelope data EN and other processing data, and is provided with a working register group 30 described later. The envelope data EN may be stored in the ROM 11. The first waveform RAM 5, the second waveform RAM 6, and the control RAM 7 may be a single RAM. The ROM 11 stores a program for the CPU 10 to perform various processes such as a program according to a flowchart described later.
[0013]
The interface 12 transmits and receives various data to and from a device such as an electronic musical instrument connected to the sound correction recording device. The transmission / reception data includes the above-described sound data SD, modified sound data MS, envelope data EN, and the like. Various data such as sound data SD, modified sound data MS, envelope data EN, programs, and the like are displayed on the CRT 15 through the CRT controller 16 or printed by the printer 18 through the printer controller 17.
[0014]
Various data such as the sound data SD, the modified sound data MS, and the envelope data EN, programs, and the like are saved on the floppy disk 20 through the floppy disk driver 19 or loaded and taken in on the contrary.
[0015]
The keyboard 13 is a computer keyboard, and inputs data such as numbers, characters, symbols, and commands. Each key of the keyboard 13 is scanned by the key scanning circuit 14, data indicating on / off is detected, and written to the control RAM 7 by the CPU 10, and an on event and an off event of each key are discriminated.
[0016]
In the tone generator 21, a sound signal corresponding to the sound information given from the CPU 10 or the like is generated and generated by the sound system 22. The sound information includes the sound data SD, the modified sound data MS, the envelope data EN, and the like. The tone generator 21 and the sound system 22 may be replaced with a clock oscillator, a piezoelectric speaker, or the like.
[0017]
2. Envelope data EN
FIG. 2 shows the envelope data EN stored in the control RAM 7. As shown in FIGS. 6 (2) to (3), the envelope data EN is multiplied and synthesized with the sound data SD, and the level of the head point of the sound data SD suddenly rises from “0” to a certain value. Discontinuous points are eliminated. The envelope data EN is waveform data that gradually increases from the level “0” to the exponential and maintains the level 1 when reaching the level “1”.
[0018]
The envelope data EN stored in the control RAM 7 is as shown in FIG. 2 (1), but the data shown in FIGS. 2 (2) and 2 (3) may be stored. If the envelope data EN is of a type that maintains the level “1” after reaching the level “1”, the control RAM 7 stores the hold point HL that reaches the level “1” from the beginning. May be. The envelope data EN may have a hold level other than “1”, or may have no hold level.
[0019]
3. Working register group 30
FIG. 3 shows the working register group 30. The register group 30 stores clear point data CP, delay length data DL, and the like. The clear point data CP indicates the address data at the end of the area from which the noise data NS at the beginning of the sound data SD is removed, as shown in FIGS. The sound data SD before the clear point CP is cleared.
[0020]
The delay length data DL indicates the number of addresses of the data level “0” added to the head of the modified sound data MS modified by multiplying and synthesizing the envelope data EN. By adding the data level “0”, stereo control can be performed.
[0021]
This stereo control is disclosed in, for example, Japanese Patent Application No. 2-408859 (Japanese Patent Application No. 3-204404), Japanese Patent Application No. 3-071256, and the like. The clear point data CP and the delay length data DL are input from the keyboard 13 and stored in the working register group 30 by the CPU 10.
[0022]
4). Waveform Recording Process FIG. 4 shows a flowchart of the waveform recording process. This process starts when the operator operates the keys on the keyboard 13 to instruct sound waveform recording. This instruction is a specific key operation or, for example, a command of “RECORDING START” is input to operate an enter key.
[0023]
In this case, the operator instructs the microphone 1 toward the sound source of the sound to be sampled and recorded. As the sampling source, a reproduced sound signal from an analog record, a reproduced digital sound signal from a compact disc, or the like can be used.
[0024]
In this process, the CPU 10 resets the address counter 9 and the sampling signal oscillator 8 (step 01), sets programmable data for determining the period of the sampling signal SP in the sampling signal oscillator 8 (step 02), and sets the address counter 9 and The sampling signal oscillator 8 is enabled (step 03).
[0025]
As a result, a sampling signal SP of a clock signal having a fixed period is output from the sampling signal oscillator 8, the latch 4 is latched for each period of the sampling signal SP, and the address counter 9 is incremented for each period of the sampling signal SP. A write command signal is given to the waveform RAM 5 for each cycle of the sampling signal SP. As a result, the sound data SD obtained by sampling the sound signal SS is written into the first waveform RAM 5.
[0026]
Then, the CPU 10 waits until the address data of the address counter 9 becomes “11... 1” (step 04). Thereby, the sampling recording process of the sound signal SS described above is repeated. When the address data of the address counter 9 becomes the end address “11... 1”, the CPU 10 disables the address counter 9 and the sampling signal oscillator 8 and stops the operation of the address counter 9 and the sampling signal oscillator 8 (step S1). 05) The sound information of the notification sound is set in the tone generator 21, and the notification sound is generated by the sound system 22 (step 06). Thereby, the operator knows that the sampling recording of the sound signal SS is completed. Thereafter, the CPU 10 performs other processes (step 07) and ends the waveform recording process.
[0027]
Thus, after sampling and recording the sound signal SS, the operator inputs a command for reading and displaying the sound data SD from the keyboard 13. As a result, the CPU 10 sequentially reads the sound data SD sampled and recorded in the first waveform RAM 5, displays it on the CRT 15, and further prints it on the printer 18 if necessary. As a result, as shown in FIG. 6A, the sound signal SS is simulated and displayed by the dot group representing the level of each point of the sound data SD.
[0028]
On the other hand, the operator inputs a vertical line cursor display command from the keyboard 13 to display the vertical line cursor, and moves the cursor to the clear point CP at the beginning of the sound data SD using the mouse or the like to enter. push. Then, the CPU 10 starts a waveform correction process described below. In this case, another means such as a light pen may be used for enlarging and displaying the head portion of the sound data SD on the CRT 15 or inputting a clear point.
[0029]
Further, the address data determined in step 04 may be other than the end address of the first waveform RAM 5, and in short, is the address data at the end of the area for storing the sound data SD.
[0030]
5. Waveform Correction Process FIG. 5 shows a flowchart of the waveform correction process. This process starts when the operator operates the keys on the keyboard 13 to instruct the correction of the sampled sound signal SS. This instruction is a specific key operation or, for example, operating the enter key by inputting a command of “MODIFY WAVEFORM”.
[0031]
In this process, the CPU 10 copies the sound data SD of the first waveform RAM 5 to the second waveform RAM 6 as it is, and performs the waveform correction process described below (step 11). This waveform correction process may be performed directly on the sound data SD of the first waveform RAM 5.
[0032]
Next, the CPU 10 sets the designated address of the second waveform RAM 6 to “00... 0” (step 12), writes the data “00... 0” to the address “00. Is cleared (step 13), and the designated address data is incremented by 1 (step 14). Then, the clear processing of the head portion of the sound data SD in steps 13 and 14 is performed until the designated address data reaches the clear point data CP (step 15).
[0033]
Thus, all data from the beginning of the sound data SD to the clear point CP is cleared, and the noise data NS at the beginning is removed. The sound data SD after this clearing is as shown in FIG.
[0034]
Next, the CPU 10 sets the designated address of the control RAM 7 to “00... 0” (step 21), reads the envelope data EN at the address “00... 0” of the control RAM 7 (step 22), and further designates the second waveform RAM 6. The address, that is, the sound data SD of the clear point CP is read (step 23). Then, the read envelope data EN and sound data SD are multiplied and synthesized (step 24) and written to the designated address of the second waveform RAM 6 (step 25).
[0035]
Thereafter, the designated address of the control device RAM 7 is incremented by 1 (step 26), the designated address of the second waveform RAM 6 is also incremented by 1 (step 27), and the multiplication / synthesis processing of the envelope data EN with respect to the sound data SD in the above steps 22-27. Is repeated until the end of the envelope data EN (step 28).
[0036]
In this case, at the clear point CP of the sound data SD, the envelope data EN at the level “0” is multiplied and synthesized at the discontinuous portion where the data level suddenly rises from “0” to a certain value. The clear point CP of SD is corrected to “0” level.
[0037]
After the clear point CP of the sound data SD, the envelope data EN that gradually increases from “0” is multiplied and synthesized. Therefore, as shown in FIG. 6 (3), the clear point CP of the sound data SD. The discontinuous portion is eliminated, and corrected sound data MS that starts smoothly is created, and noise at the head point is removed.
[0038]
Thereafter, if the transfer mode of the modified sound data MS is set (step 31), the CPU 10 sends data “00... 0” to the interface 12 by the number corresponding to the delay length data DL (steps 32 to 32). 34) Next, the modified sound data MS is read from the address corresponding to the clear point data CP of the second waveform RAM 6 (step 35), and sent to the interface 12 (step 36). Repeat until the address (step 37).
[0039]
In addition, you may perform the process of these steps 31-37 separately from the process of steps 11-15, 21-28. In this case, a sound data SD transfer instruction is given from the keyboard 13, and the processing of steps 31 to 37 is performed accordingly. Finally, the CPU 10 sets the sound information of the notification sound in the tone generator 21 and causes the sound system 22 to generate the notification sound (step 38). Thereby, the operator knows that the correction of the sound data SD has been completed. Thereafter, the CPU 10 performs other processing (step 39) and ends the waveform correction processing.
[0040]
The present invention is not limited to the above embodiments, and various modifications can be made without departing from the spirit of the present invention. For example, the address data determined in steps 28 and 37 may be other than the end address of the control RAM 7 and the end address of the second waveform RAM 6. In short, the address data at the end of the area storing the envelope data EN and the modified sound data MS is important. If it is.
[0041]
The process of clearing the head portion of the sound data SD and multiplying and synthesizing the envelope data EN to create the modified sound data MS removes specific data from the sound data SD in addition to removing the noise data NS. You may adapt to do.
[0042]
Further, the envelope data EN may be calculated and generated based on a specific arithmetic expression in addition to storing and reading the envelope data EN. This arithmetic expression is, for example, an exponential function expression, a logarithmic function expression, a linear function expression, a quadratic function expression, a hyperbolic function expression, a trigonometric function expression, or the like. The sampling recording means of the microphone 1, the amplifier 2, the A-D converter 3, the latch 4, the first waveform RAM 5, the sampling signal oscillator 8 and the address counter 9 are not limited to these, and any sampling recording can be performed. Good.
[0043]
In addition, the multiplication and synthesis of the sound data SD and the envelope data EN do not need to make the head of the envelope data EN coincide with the clear point CP of the sound data SD, and may be slightly shifted. The leading value of the envelope data EN does not have to be “0”, but may be small enough to remove noise at discontinuous points.
[0044]
The claims at the beginning of the filing of the parent application of the present application are as follows.
[A] Sampling means for sampling the input sound signal, sound recording means for recording the sound data sampled by the sampling means, and clear means for clearing the head portion of the sound data recorded in the sound recording means And a correction synthesis means for synthesizing correction data that gradually increases from level 0 to the sound data cleared by the clearing means.
[0045]
[B] The sound recording correction method according to claim A, wherein the sound processing method further comprises storage means for storing the corrected sound data corrected by the correction synthesis means.
[0046]
[C] The sound recording correction method according to claim A, wherein the modified sound data modified by the modification synthesizing means is used as a sound source of a stereo system.
[0047]
[D] Sampling means for sampling the input sound signal, sound recording means for recording the sound data sampled by the sampling means, and clear means for clearing the head portion of the sound data recorded in the sound recording means Sound recording correction method characterized by comprising
[0048]
[E] The sound recording apparatus according to claim A or D, wherein the sound recording apparatus is modified and recorded by the sampling means, the sound recording means, the clearing means, and the correcting and synthesizing means, or the sampling means, the sound recording means and the clearing means.
[0049]
【The invention's effect】
As described in detail above, the present invention clears the head portion of the sampled sound data, and further increases the value from level “0” to “1” with respect to the sound data that has cleared the head portion. Modified data was synthesized. Therefore, the noise at the head portion of the recorded sound data is removed, the head point of the sound data gradually increases from the level “0”, and the discontinuous points after the noise removal are eliminated.
[Brief description of the drawings]
FIG. 1 is an overall circuit diagram of a sound correction recording apparatus.
FIG. 2 is a diagram showing envelope data EN stored in a control RAM 7;
FIG. 3 is a diagram showing a working register group 30 in the control RAM 7;
FIG. 4 is a flowchart of waveform recording processing of a sound signal SS.
FIG. 5 is a flowchart of a waveform correction process for sound data SD.
FIG. 6 is a diagram illustrating a process of storing and correcting a sound signal SS.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Microphone, 4 ... Latch, 5 ... 1st waveform RAM, 6 ... 2nd waveform RAM, 7 ... Control RAM, 8 ... Sampling signal oscillator, 9 ... Address counter, 10 ... CPU, 11 ... ROM, 12 ... Interface, 13 ... Keyboard, 15 ... CRT, 30 ... Working register group.

Claims (5)

Means for sampling the incoming sound signal;
Means for recording the sampled sound data;
Means for specifying the clear point of the recorded sound data;
The first part before the clear point of the sound data is cleared, the delay length data of the value “0” is formed in a predetermined length before the clear point, and the level for the sound data after the clear point is set. Means for multiplying and storing corrected data that gradually increases from “0” and reaches “1”;
A sound correction apparatus comprising: means for using, as sound, the corrected sound data in which the delay length data is formed and the correction data is multiplied and synthesized. CPU to sound correction device
Sampling the incoming sound signal,
Record this sampled sound data,
Specify the clear point of the recorded sound data,
The head part before the clear point of the sound data is cleared, the delay length data of the value “0” is formed to a predetermined length before the clear point, and the level for the sound data after the clear point is set. The corrected data that gradually increases from “0” and reaches “1” is multiplied and synthesized , stored ,
A sound correction method, characterized in that the modified sound data, in which the delay length data is formed and the modified data is multiplied and synthesized, is used as a sound. Means for sampling the incoming sound signal;
Means for recording the sampled sound data;
Means for specifying the clear point of the recorded sound data;
The head portion before the clear point of the sound data is removed, and the sound data after the clear point is multiplied and synthesized by the correction data that gradually increases from level “0” and reaches “1”, and the sound Means for storing the delay length data of the value “0” in a predetermined length at the head portion of the data;
And a means for using the modified sound data obtained by multiplying and synthesizing the modified data and forming the delay length data as a sound. CPU to sound correction device
Sampling the incoming sound signal,
Record this sampled sound data,
Specify the clear point of the recorded sound data,
The head portion before the clear point of the sound data is removed, and the sound data after the clear point is multiplied and synthesized by correction data that gradually increases from level “0” and reaches “1”, and the sound The delay length data of the value “0” is formed in a predetermined length at the beginning of the data and stored,
A sound correction method characterized in that the corrected sound data obtained by multiplying and synthesizing the correction data to form delay length data is used as a sound. 4. The sound correction apparatus according to claim 1, wherein the corrected sound data is used as a sound source of a stereo system, and the delay length data is for performing stereo control.

Images