JP6765124B2 - Voice processing device, voice processing method, and voice processing program - Google Patents

Description

The present invention relates to a voice processing device or the like that performs predetermined voice processing on a target signal to generate a processed signal.

ADPCM (adaptive differential pulse code modulation) is known as a signal processing technique. ADPCM is an important technology in the amusement field where a plurality of audio signals are simultaneously reproduced because the circuit scale of hardware implementation is small and high-speed decoding is possible.

However, ADPCM may be inferior in reproduced sound quality to other technologies capable of achieving the same degree of compression ratio, for example, a codec technology of a transform coding method such as MP3.

On the other hand, a technique using a post filter having a gentle cutoff frequency to improve the quality of the reproduced signal in consideration of the tendency of the frequency spectrum between the quantization noise generated by ADPCM and the original signal to be encoded by ADPCM. It has been known.

As related technologies, in order to perform noise shaping on the quantization processing unit for generating an audio signal formed by a plurality of time-series signals and the quantization noise generated when the quantization processing unit performs quantization. An audio signal processing device including the noise shaping means of the above is known. The noise shaping means includes a post filter having a characteristic capable of realizing noise shaping of a quantized audio signal and a pre filter having an inverse characteristic of the post filter, and the post filter is placed after the quantization processing unit and is prefiltered. Is disclosed in each of the preceding stages of the quantization processing unit (see, for example, Patent Document 1).

In addition, as another related technique, in the coefficient setting method of the noise shaping filter that reduces the quantization error, after taking the reciprocal of the power spectrum shape based on the frequency analysis result of a predetermined number of samples, the inverse orthogonal conversion is performed. A technique for calculating a filter coefficient based on a obtained autocorrelation coefficient is known (for example, Patent Document 2). According to this, it is possible to provide a coefficient setting method of a noise shaping filter that enables real-time processing for reducing quantization error by a normal so-called DSP (digital signal processor).

Japanese Unexamined Patent Publication No. 2016-213683 JP-A-4-72907

For example, the voice processing technique for improving the quality of the reproduced signal by using the above-mentioned post filter improves the characteristics of the high frequency band, and does not reduce the quantization noise in the low and middle frequency bands.

On the other hand, in order to reduce not only the high frequency band but also the low / medium frequency band quantization noise, a method of minimizing the energy of the quantization noise can be considered. When the energy of the quantization noise is minimized, the spectral envelope of the quantization noise is flat in all frequency bands.

However, even if the spectral envelope of the quantization noise is flattened in the entire frequency band, the quality of the reproduced signal is not necessarily improved for humans.

On the other hand, noise shaping is applied in consideration of the frequency characteristics of the original signal, the spectrum of the quantization noise is deformed, and the quantization noise is concentrated in the place where the energy of the original signal is strong, so that the energy of the original signal is weak. A method of reducing the quantization noise can be considered. However, if the intensity of noise shaping is fixed, there is a problem that the total amount of quantization noise is increased and the quality of the reproduced signal is deteriorated.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique capable of improving the quality of a reproduced signal.

In order to achieve the above object, the voice processing device according to the first aspect is a voice processing device having a voice processing unit that performs predetermined voice processing on a target signal to generate a processed signal, and performs voice processing. In each of the voice processing control unit that changes the value of one or more parameters in the voice processing by the unit to a plurality of values and causes the voice processing unit to execute the voice processing, and the case where the value of the parameter is changed to a plurality of values. , A difference detector that detects the difference between the energy of the target signal and the energy of the reproduced signal generated based on the processed signal for each predetermined frequency component, and when the parameter values are changed to multiple values. In each case, the difference for each frequency component detected by the difference detection unit is weighted according to the sensitivity corresponding to the frequency component of human hearing, and the feature amount calculation unit calculates the feature amount based on them. , The effective value, which is the value of the parameter suitable for the use of voice processing, is determined based on the plurality of feature quantities calculated by the feature quantity calculation unit in each case where the parameter value is changed to a plurality of values. It includes a valid value determination unit.

In the above-mentioned voice processing device, the value of the parameter in the voice processing is set to the valid value detected by the valid value detection unit, the voice processing is executed, and the processed signal obtained by the voice processing is used as a valid post-processing signal. It is also possible to further have an effective signal generation control unit to be stored in the storage unit as a signal after effective processing.

Further, in the above-mentioned voice processing apparatus, the voice processing unit includes a noise shaping unit that performs noise shaping that changes the frequency characteristic of the quantization noise in the target signal, and the parameters are parameters related to noise shaping in the noise shaping unit. May be good.

Further, in the voice processing device, the difference detection unit detects the difference in units of blocks of a predetermined size of the target signal, the feature amount calculation unit calculates the feature amount in units of blocks, and the effective value determination unit determines the effective value. The effective value may be determined in units of blocks.

Further, the voice processing device further includes an auditory information storage unit that stores information on sensitivity corresponding to the frequency component of human hearing, and the feature amount calculation unit is weighted based on information on sensitivity of the auditory information storage unit. May be decided.

Further, in order to achieve the above object, the voice processing method according to the second aspect is a voice processing method using a voice processing device having a voice processing unit that performs predetermined voice processing on the target signal to generate a processed signal. When the value of one or more parameters in the voice processing by the voice processing unit is changed to a plurality of values, the voice processing unit is made to execute the voice processing, and the parameter is changed to each of the plurality of values. The difference between the energy of the original signal and the energy of the reproduced signal generated based on the processed signal is detected for each predetermined frequency component, and the frequency component of human hearing is compared with the difference for each detected frequency component. Weighting is performed according to the sensitivity corresponding to, the feature amount based on them is calculated, and the effective value which is the value of the parameter suitable for the use of the voice processing is determined based on the calculated feature amount.

Further, in order to achieve the above object, the voice processing program according to the third aspect performs voice processing to be executed by a computer constituting a voice processing device that performs predetermined voice processing on the target signal and generates a processed signal. The voice processing program is a voice processing control unit, which causes the computer to change the value of one or more parameters in the voice processing by the voice processing unit to a plurality of values, and causes the voice processing unit to execute the voice processing. Difference detector that detects the difference between the energy of the target signal and the energy of the reproduced signal generated based on the processed signal for each predetermined frequency component when the value of the parameter is changed to a plurality of values. And, in each case where the value of the parameter is changed to a plurality of values, the difference for each frequency component detected by the difference detection unit is weighted according to the sensitivity corresponding to the frequency component of human hearing. , The feature amount calculation unit that calculates the feature amount based on them, and the voice processing based on the plurality of feature amounts calculated by the feature amount calculation unit in each case where the parameter value is changed to a plurality of values. It functions as an effective value determination unit that determines an effective value, which is a value of a parameter suitable for the use of.

According to the present invention, the quality of the reproduced signal can be improved.

FIG. 1 is a functional block diagram of the voice processing device according to the embodiment. FIG. 2 is a functional block diagram of a voice processing unit of the voice processing device according to the embodiment. FIG. 3 is a diagram showing an A curve showing sensitivity to a human auditory frequency according to an embodiment. FIG. 4 is a hardware configuration diagram of the voice processing device according to the embodiment. FIG. 5 is a flowchart of the voice generation process according to the embodiment. FIG. 6 is a flowchart of the feature amount calculation process according to the embodiment.

The embodiment will be described with reference to the drawings. It should be noted that the embodiments described below do not limit the invention according to the claims, and all of the elements and combinations thereof described in the embodiments are essential for the means for solving the invention. Is not always.

First, the voice processing device according to the embodiment will be described.

FIG. 1 is a functional block diagram of the voice processing device according to the embodiment. FIG. 2 is a functional block diagram of a voice processing unit of the voice processing device according to the embodiment.

The voice processing device 1 includes a voice processing unit 11, a difference detection unit 12, a feature amount calculation unit 13, an effective value determination unit 14, a voice processing control unit 15 as an example of an effective signal generation control unit, and auditory information. A storage unit 20 is provided as an example of the storage unit.

The voice processing unit 11 performs predetermined voice processing (for example, ADPCM encoding processing) on the original signal s (n) as an example of the target signal. As shown in FIG. 2, the speech processing unit 11 includes an adaptive quantization unit 31 as an example of a quantization unit, an inverse adaptive quantization unit 32 as an example of an inverse quantization unit, and arithmetic units 33 and 34. A noise shaping unit 35, a calculation unit 36, an adaptive prediction unit 37, and a calculation unit 38 are provided.

The calculation unit 33 outputs a difference signal which is a difference between the original signal s (n) and the prediction signal output from the adaptation prediction unit 37. The calculation unit 34 adds the difference signal output from the calculation unit 33 and the signal output by the noise shaping unit 35.

The adaptive quantization unit 31 outputs the processed signal x (n) by executing the adaptive quantization process (decoding process) on the signal (difference signal after noise shaping) output from the calculation unit 34. To do. The inverse adaptive quantization unit 32 executes an inverse adaptive quantization process (decoding process) on the processed signal x (n) output by the adaptive quantization unit 31.

The calculation unit 36 outputs a signal of the difference between the signal output from the inverse adaptive quantization unit 32 and the signal output from the calculation unit 34. The signal output from the arithmetic unit 36 is a quantization noise signal with respect to the signal input to the adaptive quantization unit 31.

The noise shaping unit 35 receives the signal output from the calculation unit 36 as an input, and performs noise shaping that deforms the shape of the quantization noise signal.

Here, regarding the transfer function (noise transfer function) H (z) of the quantization noise, the transfer function H (γz) in which a parameter (intensity γ (0 or more and 1 or less)) for changing the noise shaping intensity is introduced is introduced. Then, this transfer function can be expressed as shown in the following equation (1). In the formula ^{(1), H - (γz} ) ( In this specification, "H ^-" it is on the H. The "-" and represent a) is equivalent to the feedback element of the noise shaping section 35 To do.

In the present embodiment, the noise transfer function H (γz) is the transfer function A in which the above-mentioned parameter intensity γ is introduced into the transfer function A (z) when the reproduction signal y (n) is used as an autoregressive model. It is set to (γz). By setting the noise transfer function H (γz) to the transfer function A (γz) in this way, noise shaping can be performed by spectral wrapping similar to the original signal, and noise hiding by so-called auditory masking can be expected. .. Further, as will be described later, since the value of the intensity γ can be set to a value effective for improving the quality of the reproduced signal, the total energy of noise can be suppressed and the sound quality of the reproduced signal can be improved.

Here, the transfer function A (z) of the autoregressive model will be described.

First, assuming that the predicted value of the reproduction signal y (n) at time n is predicted by a linear sum using the past k samples (k is a predetermined number) and k prediction coefficients αi, the time is assumed. Assuming that the predicted value of the reproduced signal of n is y ^ _k (n) (in the present specification, “y ^” represents “^” directly above y), the following equation ( It is represented as shown in 2).

Assuming that the difference signal between the predicted value y ^ _k (n) and the actual value y (n) is d (n), the difference signal d (n) is expressed as shown in the following equation (3). ..

By substituting the equation (2) into the equation (3), the relationship shown in the following equation (4) can be obtained. This equation (4) is called an autoregressive model.

By z-transforming this equation (4), the transfer function A (z) of the equation (4) can be obtained as shown in the following equation (5).

The prediction coefficient αi can be obtained by using the Levinson-Durbin algorithm.

And H (γz) = A (γz ), according to the equations (5) and (1), feedback element H ^- (γz) is expressed as shown in the following equation (6).

The calculation unit 38 outputs a reproduction signal y (n) obtained by adding the signal output by the inverse adaptive quantization unit 32 and the prediction signal output by the adaptive prediction unit 37.

The adaptive prediction unit 37 takes the reproduction signal y (n) as an input and outputs a prediction signal for predicting the original signal s (n).

Returning to the description of FIG. 1, the storage unit 20 stores the original signal data 21, the auditory information 22, and the effective processed signal data 23.

The original signal data 21 is a digitalized original signal s (n) to be processed by the voice processing unit 11. The original signal s (n) is, for example, an analog original signal obtained by converting an analog value at each time point into a digital value corresponding to the analog value at predetermined sampling intervals.

The auditory information 22 is information (auditory information) relating to the sensitivity corresponding to the frequency component of human hearing. As auditory information, for example, information corresponding to a curve having a mirror characteristic of an equal loudness curve indicating a frequency change of a sound pressure level that humans perceive as having the same loudness, and JIS C 15091-1 : Information corresponding to the 2005 A curve and C curve.

Here, the A curve will be described.

FIG. 3 is a diagram showing an A curve showing sensitivity to a human auditory frequency according to an embodiment. In FIG. 3, the horizontal axis is the frequency [KHz] and the vertical axis is the sound pressure [dB]. On the vertical axis, the sound pressure at a frequency of 1.0 [kHz] is 0 [dB].

As shown in the A curve of FIG. 3, human hearing differs in the value perceived as the same sound pressure depending on the frequency, that is, the sensitivity is different.

Returning to the description of FIG. 1, the signal data 23 after the effective processing is set by the voice processing unit 11 when the effective value determined by the effective value determining unit 14 is set for the intensity γ of the noise shaping unit 35. The processed signal x (n) that is generated. By decoding the signal data 23 after the effective processing, it is possible to generate a reproduction signal capable of reproducing high-quality sound.

The difference detection unit 12 sets a predetermined unit (block) of the original signal s (n) as a processing target in each case where the value of the parameter of the noise shaping unit 35 is changed to a plurality of values, and the block (processing target). The logarithmic power spectrum of the block) and the block to be processed of the original signal s (n) are encoded by the adaptive quantization unit 31 and decoded by the inverse adaptive quantization unit 32, and then the reproduced signal y (n) (FIG. 2). Then, the logarithmic power spectrum of the signal output from the calculation unit 38) is calculated, and the absolute value of the difference between the two obtained logarithmic power spectra is obtained for each frequency component. Specifically, the difference detection unit 12 calculates the logarithmic power spectrum S (f) for the block of the original signal s (n) by the following equation (7).

Further, the difference detection unit 12 calculates the logarithmic power spectrum Y (f) for the block of the reproduction signal y (n) by the following equation (8).

Next, the difference detection unit 12 uses the following equation (9) to obtain the logarithmic power spectrum S (f) of the original signal s (n) and the logarithmic power spectrum Y (f) of the block of the reproduced signal y (n). The difference power spectrum D (f), which is the difference between the absolute values for each frequency component of, is calculated.

The feature amount calculation unit 13 weights w based on human hearing for each frequency component of the difference power spectrum D (f) in each case where the value of the parameter of the noise shaping unit 35 is changed to a plurality of values. (F) is multiplied, and the feature amount (FDD: Frequency Domain Difference) is calculated based on all the obtained values (for example, by adding all).

Here, when the auditory information 22 has a correspondence relationship between the frequency shown by the A curve shown in FIG. 3 and the sound pressure, the weight w (f) for each frequency component is the value of the sound pressure on the vertical axis with respect to the frequency. Let _CA (f) be, and it is calculated by the following equation (10).

The feature amount calculation unit 13 calculates the feature amount FDD by the following formula (11).

Here, the feature amount FDD is the sum of the energy differences between the original signal s (n) and the reproduced signal y (n) of each frequency, weighted according to the characteristics of human hearing. From this, when the feature amount FDD is small, the influence of the quantization noise on the reproduction signal y (n) is small on the human hearing, that is, the sound quality of the sound by the reproduction signal y (n) is good. Shown.

The voice processing control unit 15 sets each of the blocks of the original signal s (n) as a target (process target block), and changes the value of one or more parameters in the voice processing by the voice processing unit 11 to a plurality of values, respectively. The voice processing unit 11 is made to execute the voice processing in each state in which the value of is set, and the difference detection unit 12 and the feature amount calculation unit 13 are made to execute the processing for each voice processing. In the present embodiment, the voice processing control unit 15 changes, for example, the intensity γ in the noise shaping unit 35 to a plurality of values in the range of 0 or more and 1 or less. Further, when the valid value of the parameter is determined by the valid value determining unit 14, the voice processing control unit 15 is one or more in the voice processing for the processing target block of the original signal s (n) by the voice processing unit 11. The value of the parameter is changed to an effective value corresponding to the processing target block, and the processed signal x (n) generated by the voice processing unit 11 is used as a valid post-processing signal (effective post-processing data) of the processing target block. , Stored in the storage unit 20.

The valid value determination unit 14 is a parameter capable of performing voice processing with little deterioration based on the feature amount FDD calculated by the feature amount calculation unit 13 corresponding to each value of the parameter changed by the voice processing control unit 15. Determine the value (effective value) of (intensity γ). In the present embodiment, the effective value determining unit 14 determines the value of the parameter having the smallest feature amount FDD as the effective value for the block to be processed.

Next, the hardware configuration of the voice processing device 1 will be described in detail.

FIG. 4 is a hardware configuration diagram of the voice processing device according to the embodiment.

The voice processing device 1 includes a control circuit 101, a storage device 102, a reader / writer 103, a communication interface (communication I / F) 104, an input / output interface (input / output I / F) 105, an input device 106, and the like. It is composed of a computer including a display device 107. The control circuit 101, the storage device 102, the reader / writer 103, the communication I / F 104, the input / output I / F 105, and the display device 107 are connected via the bus 108.

The control circuit 101 is, for example, a processor, and controls the entire voice processing device 1. The control circuit 101 executes various processes by executing a program stored in the storage device 102. In the present embodiment, the control circuit 101 executes the voice processing program stored in the storage device 102 to execute the voice processing unit 11, the difference detection unit 12, the feature amount calculation unit 13, the effective value determination unit 14, and the voice. The processing control unit 15 is configured.

The storage device 102 is, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), a RAM, a ROM, or the like, and stores a program (voice processing program or the like) executed in the control circuit 101 and various information. The recording device 102 constitutes the storage unit 20 shown in FIG.

The reader / writer 103 has a removable recording medium 110, and reads data from the recording medium 110 and writes data to the recording medium 110. Examples of the recording medium 110 include non-temporary recording media (nonvolatile recording media) such as SD memory cards, FDs (floppy disks: registered trademarks), CDs, DVDs, BDs (registered trademarks), and flash memories. In the present embodiment, the recording medium 110 stores the voice processing program, the original signal s (n) used for the processing, the auditory information, and the like, and the read-dryer 103 reads them out and uses them. May be good.

The communication I / F 104 is connected to the network 111, and transmits / receives data to / from another device connected to the network 111.

Next, the operation of the voice generation processing in the voice processing device 1 according to the present embodiment will be described.

FIG. 5 is a flowchart of the voice generation process according to the embodiment.

The voice processing control unit 15 extracts an unprocessed block (process target block) within the processing range of the original signal s (n) from the storage unit 20 (step S101). The processing range of the original signal s (n) may be, for example, the entire original signal s (n) or a part of the original signal s (n) designated by the user.

Next, the voice processing control unit 15 sets the intensity γ of the noise shaping unit 35 of the voice processing unit 11 to 0 (step S102), inputs the extracted processing target block to the voice processing unit 11, and executes voice processing. (Step S103).

Next, the voice processing control unit 15 tells the difference detection unit 12 and the feature amount calculation unit 13 the original signal s (n) used for voice processing by the voice processing unit 11 and the reproduction signal y output by the voice processing unit 11. Using (n), a process of calculating the feature amount FDD (feature amount calculation process) is executed (step S104).

Next, the voice processing control unit 15 changes the intensity γ of the voice processing unit 11 (step S105). For example, the voice processing control unit 15 adds a predetermined value (for example, 0.01) to the value of the intensity γ.

Next, the voice processing control unit 15 determines whether or not the intensity γ is greater than 1 (step S106). As a result, when the intensity γ is not greater than 1 (step S106: No), it is necessary to calculate the feature amount FDD based on the intensity γ, so the voice processing control unit 15 advances the processing to step S103. On the other hand, when the intensity γ is larger than 1 (step S106: Yes), it means that the required feature amount FDD is calculated by sufficiently changing the intensity γ, so that the voice processing control unit 15 steps the process. Proceed to S107.

In step S107, the effective value determining unit 14 has an intensity γ capable of generating a processed signal capable of obtaining a reproduced signal with less quantization noise based on the plurality of feature quantities FDD calculated in step S104. The value (valid value) is determined, and the voice processing control unit 15 sets the determined effective value to the intensity γ of the voice processing unit 11. In the present embodiment, the effective value determining unit 14 determines the value of the intensity γ having the smallest feature amount FDD as the effective value for the block to be processed.

Next, the voice processing control unit 15 inputs the processing target block to the voice processing unit 11 to execute voice processing, and stores the processed signal x (n) generated by the voice processing unit 11 as effective processed signal data. It is stored in the unit 20 (step S108).

Next, the voice processing control unit 15 determines whether or not processing has been performed on all the blocks in the processing range of the original signal s (n) (step S109). As a result, when the processing is not performed for all the blocks in the processing range (step S109: No), the voice processing control unit 15 advances the processing to step S101 and performs the processing for the next block. On the other hand, when processing is performed on all the blocks in the processing range (step S109: Yes), the voice processing control unit 15 ends the voice generation processing.

Next, the feature amount calculation process shown in step S104 of FIG. 5 will be described.

FIG. 6 is a flowchart of the feature amount calculation process according to the embodiment.

The difference detection unit 12 calculates the logarithmic power spectrum S (f) of the original signal s (n) according to the equation (7) (step S201). Further, the difference detection unit 12 calculates the logarithmic power spectrum Y (f) of the reproduction signal y (n) by the equation (8) (step S202). Then, the difference detection unit 12 calculates the difference power spectrum D (f), which is the difference between the absolute values of the logarithmic power spectrum S (f) and the logarithmic power spectrum Y (f), for each frequency by the equation (9). (Step S203). Next, the feature amount calculation unit 13 assigns a weight w (f) according to the human auditory characteristic corresponding to the corresponding frequency to each difference power spectrum D (f). That is, the feature amount calculation unit 13 multiplies each difference power spectrum D (f) by the corresponding weight w (f) according to the equation (10) (step S204). Next, the feature amount calculation unit 13 calculates the feature amount FDD by summing the difference power spectra D (f) for each weighted frequency according to the equation (11) (step S205).

As described above, according to the voice processing device 1 according to the present embodiment, the values of one or more parameters in the voice processing are changed to a plurality of values, the voice processing unit 11 is made to execute the voice processing, and the original signal s. The difference between the energy of (n) and the energy of the reproduction signal y (n) for each predetermined frequency component is detected, and the sensitivity corresponding to the frequency component of human hearing with respect to the difference for each detected frequency component. The quantization error is based on the plurality of feature quantities calculated by the feature quantity calculation unit 13 when the weighting is performed according to the above, the total feature quantity is calculated, and the parameter is changed to each of the plurality of values. Since the effective value of the parameter suitable for the generation of the processed signal that can obtain the reproduced signal with less is determined, the quality of the audio by the reproduced signal generated by the processed signal generated by the audio processing can be improved. Can be done.

The present invention is not limited to the above-described embodiment, and can be appropriately modified and implemented without departing from the spirit of the present invention.

For example, in the above embodiment, the feature amount FDD is calculated by constantly changing the value of the intensity γ of the noise shaping unit 35 within a predetermined range (range of 0 or more and 1 or less) for each block of the original signal. However, the present invention is not limited to this, and the value of the intensity γ is within a limited range around the effective value based on the effective value of the intensity γ for a block before a predetermined number (for example, immediately before). May be changed. The effective value of the intensity γ for the block tends to be close to the effective value of the intensity γ of the block in a time range (for example, immediately before), and by doing so, the effective value can be appropriately determined. At the same time, the processing load required to determine the effective value can be reduced.

Further, in the above embodiment, the feature amount is calculated for each block to determine the effective value of the intensity γ, but the present invention is not limited to this, and for example, among a plurality of blocks. The intensity γ may be changed for one block to determine an effective value, and the voice processing may be performed using one effective value determined as the intensity γ for the plurality of blocks. In this way, the processing load can be reduced as compared with the case where the feature amount is calculated for each block to determine the effective value of the intensity γ, and a relatively high quality reproduced signal is generated in a short time. It is possible to generate a post-processing signal that can be produced.

Further, in the above embodiment, the feature amount when the value of the intensity γ of the noise shaping unit 35 is changed is calculated, and the effective value of the intensity γ is determined based on the feature amount. Is not limited to this, and the values of other parameters in speech processing (for example, if the speech processing section has an LPF (low-pass filter), the cutoff frequency of the filter and the self-regression model used as the noise transfer function of the noise shaping section 35) The feature quantity at each parameter value is calculated by changing at least one parameter such as the number of prediction coefficients and the block size at the time of quantization, and the effective value of the parameter is calculated based on the feature quantity. It may be used when generating a signal after effective processing so as to determine.

Further, in the above embodiment, the weighting is determined and the feature amount is calculated based on the common auditory information 22, but the present invention is not limited to this, and for example, the reproduction generated from the processed signal is reproduced. For each age group (for example, teens, 20s, 30s, etc.) of the user who listens to the voice based on the signal, the auditory information corresponding to that age is prepared, and based on the auditory information. The weighting may be determined and the feature amount may be calculated. In this case, the feature amount may be calculated using the auditory information 22 of the main age of listening to the voice, the validity of the parameter may be determined, and the processed signal may be generated for each age. The feature amount may be calculated, the validity of the parameter for each age may be determined, and the processed signal suitable for each age may be generated.

Further, in the above embodiment, the auditory information 22 mainly shows an example of information corresponding to the equal loudness curve, but the present invention is not limited to this, and for example, instead of the information corresponding to the equal loudness curve, Or, in addition to the information corresponding to the equal loudness curve, the information corresponding to time masking, which is a phenomenon in which the subsequent sound is drowned out by the previous sound in time, and other information when a certain sound is generated. Information corresponding to spectrum masking, which is a phenomenon in which the sound of the sound is drowned out, may be used.

Further, in the above embodiment, the voice processing by the voice processing unit is set to ADPCM, but the present invention is not limited to this, and for example, another DPCM or a codec of another method (for example, MP3) may be used. When the voice processing by the voice processing unit is MP3, for example, when the frequency (parameter) of the sound to be cut in the voice processing is changed, which of the auditory characteristics (for example, the equal loudness curve and the information corresponding to masking) The feature amount FDD is calculated by using the weight according to (auditory characteristic based on 1 or more), the feature amount FDD can be set to be equal to or less than the threshold value, and the compression rate can be increased (for example, cut). It is only necessary to select a parameter (so that the frequency of the sound can be increased). In this way, in MP3, the compression rate can be increased by cutting a small amount of sound that does not affect the sound quality while maintaining the sound quality at a certain level or higher according to the auditory characteristics.

Further, in the above embodiment, a part or all of the processing performed by the control circuit 101 may be performed by the hardware circuit. For example, at least one of the functional units 11 to 15 shown in FIG. 1 may be configured by a hardware circuit. Further, at least a part of each configuration of the voice processing unit 11 may be configured by a hardware circuit.

1 ... voice processing device, 11 ... voice processing unit, 12 ... difference detection unit, 13 ... feature amount calculation unit, 14 ... effective value determination unit, 15 ... voice processing control unit, 20 ... storage unit, 21 ... original signal data, 22 ... Auditory information, 23 ... Signal data after effective processing

Claims (8)

A voice processing device having a voice processing unit that performs predetermined voice processing on a target signal and generates a signal after processing.
The value of one or more parameters in the voice processing by the voice processing unit is changed to each of a plurality of values , and the voice processing unit is made to change each of the changed values of the parameter for the same target signal. A voice processing control unit that executes each of the used voice processes,
Detects the difference between the energy of the target signal and the energy of the reproduced signal generated based on the processed signal for each predetermined frequency component when the value of the parameter is changed to a plurality of values. Difference detector and
When the value of the parameter is changed to a plurality of values, the difference for each frequency component detected by the difference detection unit is weighted according to the sensitivity corresponding to the frequency component of human hearing. , The feature amount calculation unit that calculates the feature amount based on them,
When the value of the parameter is changed to a plurality of values, the voice processing can be used from among the values of the plurality of parameters based on the plurality of the feature amounts calculated by the feature amount calculation unit. A valid value determination unit that determines a valid value, which is a value of a suitable parameter,
A voice processing device including. The value of the parameter in the voice processing is set to the valid value detected by the valid value determination unit, the voice processing is executed, and the processed signal obtained by the voice processing is used as a valid post-processing signal. The voice processing device according to claim 1, further comprising an effective signal generation control unit for storing the effective processing post-signal in the storage unit. The voice processing unit includes a noise shaping unit that performs noise shaping that changes the frequency characteristics of the quantization noise in the target signal.
The voice processing device according to claim 1 or 2, wherein the parameter is a parameter related to noise shaping in the noise shaping unit. The difference detection unit detects the difference in units of blocks of a predetermined size of the target signal.
The feature amount calculation unit calculates the feature amount in units of the block, and then calculates the feature amount.
The voice processing device according to claim 3, wherein the effective value determining unit determines the effective value in units of the block. Further provided with an auditory information storage unit for storing information on sensitivity corresponding to the frequency component of human hearing.
The voice processing device according to any one of claims 1 to 4, wherein the feature amount calculation unit determines the weight based on information on the sensitivity of the auditory information storage unit. A voice processing method using a voice processing device having a voice processing unit that performs predetermined voice processing on a target signal to generate a processed signal.
The value of one or more parameters in the voice processing by the voice processing unit is changed to each of a plurality of values , and the voice processing unit is made to change each of the changed values of the parameter for the same target signal. Each of the used voice processes is executed,
When the parameter is changed to each of a plurality of values, the difference between the energy of the target signal and the energy of the reproduced signal generated based on the processed signal is detected and detected for each predetermined frequency component. The difference for each frequency component is weighted according to the sensitivity corresponding to the frequency component of human hearing, a total feature amount is calculated, and based on the calculated plurality of feature amounts, the feature amount is calculated. A voice processing method for determining an effective value which is a value of a parameter suitable for using the voice processing from a plurality of values of the parameter. A voice processing program to be executed by a computer constituting a voice processing device having a voice processing unit that performs predetermined voice processing on a target signal and generates a signal after processing .
The voice processing program
The computer
The value of one or more parameters in the voice processing by the voice processing unit is changed to each of a plurality of values , and the voice processing unit is made to change each of the changed values of the parameter for the same target signal. A voice processing control unit that executes each of the used voice processes,
Detects the difference between the energy of the target signal and the energy of the reproduced signal generated based on the processed signal for each predetermined frequency component when the value of the parameter is changed to a plurality of values. Difference detector and
When the value of the parameter is changed to a plurality of values, the difference for each frequency component detected by the difference detection unit is weighted according to the sensitivity corresponding to the frequency component of human hearing. , The feature amount calculation unit that calculates the feature amount based on them,
When the value of the parameter is changed to a plurality of values, the voice processing can be used from among the values of the plurality of parameters based on the plurality of the feature amounts calculated by the feature amount calculation unit. A valid value determination unit that determines a valid value, which is a value of a suitable parameter,
A voice processing program that works. A processing unit that performs processing including noise shaping on the original signal that is the target of audio processing to generate a playback signal,
A detector that detects the difference in energy between the original signal and the reproduced signal for each frequency component,
A calculation unit that weights the difference in energy of each frequency component according to the characteristics of hearing and calculates a feature amount by adding the difference in energy of each weighted frequency component.
A control unit that changes the noise shaping parameters and causes the calculation unit to calculate the feature amount in each state in which the changed parameters are set.
A determination unit that determines the value of the parameter corresponding to the minimum feature amount or the feature amount below the threshold value as the value of the noise shaping parameter used for the voice processing from the plurality of feature amounts calculated in each state. When,
A voice processing device including.