JP2003345375A - Device and system for reproducing voice - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To compensate masking caused by peripheral noise while performing compensation adapted to individual degradation in hearing. <P>SOLUTION: A voice hearing condition setting part 5 reads a profile being a user's voice hearing condition, from a profile server 20 through a communication network. A masking extent estimation part 2 compensate masking caused by peripheral noise, of received voice data. A reproducing control part 4 obtains an articulate voice compensated for an influence of masking and degradation in hearing by performing voice emphasis processing considering user's degradation in hearing. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice reproducing device for reproducing voice information obtained through communication or a network in a mobile environment such as outdoors, and accurately relates to a user having various hearing characteristics. The present invention relates to a voice reproduction device having a function for making it easier for a terminal to hear a voice that is difficult to hear due to a noise environment, and a voice reproduction system having the same.

[0002]

2. Description of the Related Art At present, a portable voice recording / reproducing apparatus for recording and reproducing voice, a voice communication service for a mobile phone, a service for transmitting voice and audio information using a network, etc. are widely provided. There is an opportunity to listen to audio information in various forms even outdoors. However, since these devices and services generally do not perform voice reproduction in consideration of the hearing characteristics and listening environment of the user, a user with a remarkable deterioration of hearing may wear a hearing aid to improve the intelligibility. It corresponds.
The hearing compensation characteristic of a hearing aid is generally determined as the optimum parameter for the wearer through a characteristic determination process called fitting, and this parameter value is effective when it is equivalent to the listening environment in which the fitting is performed. However, under other conditions, optimum adjustment has not been achieved.

As a conventional hearing aid that has dealt with this problem, there is a history holding type hearing aid of Japanese Patent No. 2638563. FIG. 6 shows the history holding type hearing aid, that is, the sound reproducing device. Hereinafter, a conventional technique will be described with reference to the drawings. In FIG. 6, reference numeral 211 is a microphone, 212 is a hearing compensation unit, 213 is an amplification unit, 214 is an earphone, 221 is a selection unit, 310 is a current recording unit, 3
Reference numeral 20 is a history recording unit, 261 is a storage medium, 262 is a connector, and 263 is a storage medium detection unit.

First, the current recording unit 310 records a current parameter for determining a hearing aid characteristic related to the current hearing characteristic, and a current fitting recording unit 2 for storing fittings corresponding thereto.
And 41. On the other hand, the history recording unit 320 includes a plurality of parameter history recording units 23 indicating recording of parameters at the time of updating parameters for which hearing aid characteristics have been determined in the past.
2-1 to 232-n, and a plurality of fitting history recording units 242-1 to 242-n showing recording of fittings corresponding to each other in a one-to-one manner. The selection unit 221 includes a current recording unit 310 and a history recording unit 32.
It functions as a selection unit for selecting one parameter from the parameters recorded in 0. This history recording unit 3
20 is built on a removable storage medium 261.
In order to transmit the parameter history and the fitting history from the storage medium 261 to the selection unit 221, the storage medium 261 is used.
And a connector 262 for electrically connecting the selection unit 221
Intervenes. The storage medium detection unit 263 determines whether or not a recording medium is attached to the connector 262, and if not attached, instructs the selection unit 221 to prohibit data selection on the history recording unit 320. Selector 22
1 transmits the selected parameter value to the hearing compensation unit 212. The hearing compensation unit 212 performs the hearing compensation process on the audio signal input from the selection unit 211 with the parameter selected by the selection unit 221 as the hearing compensation characteristic. Amplifier 2
Reference numeral 13 amplifies the signal subjected to the hearing compensation processing of the hearing compensation unit 212 until a sufficient volume is obtained and outputs the amplified signal to the earphone 214. The earphone 214 is the hearing compensation unit 21.
2 and the final output signal of the audio signal processed by the amplification unit 213.

With such a configuration, a plurality of hearing compensation parameters of the hearing aid that have been fitted in the past are owned, and the user selects the parameters according to the fitted situation to change and give the hearing compensation characteristics. It is possible that the hearing compensation ability can be improved as compared with the case where only fixed auditory characteristics are given.

[0006]

However, in the above configuration, it is necessary to store the hearing compensation parameters in advance after the fitting is performed in the noise environment to be used, but the fitting is performed in advance under many conditions. Often requires a lot of effort and is virtually impossible. Also, even in listening environments under various conditions, there is no guarantee that the environmental conditions will always have the same characteristics.
There are many fluid factors such as time of use and external factors, and in the above configuration, it is not possible to immediately set optimal hearing compensation parameters on the spot for both the listening conditions of the terminal user and environmental noise. Have specific challenges.

In view of the above-mentioned problems, the present invention has a configuration in which the amount of masking for a voice signal is estimated based on the user's voice listening condition and the estimated value of environmental noise so that the voice enhanced voice can be heard. (EN) Provided are a sound reproducing device and a sound reproducing system including the same for immediately obtaining an easily audible sound by setting optimum hearing compensation parameters under various environmental noises.

[0008]

In order to solve this problem, the present invention uses a communication network to obtain profile information in which the user's voice listening conditions are recorded, and to determine the user's hearing characteristics and sound quality preference. At the same time as using the characteristics, by estimating the environmental noise and estimating the masking amount for the voice signal to be reproduced based on this estimated value, and applying the voice enhancement processing in consideration of both of them, various conditions can be applied. Thus, it is possible to provide a general-purpose audio reproduction device and audio reproduction system that immediately reproduces clear audio that is easy to hear.

In order to achieve this object, a voice reproducing apparatus according to the present invention comprises a voice listening condition setting means for obtaining a voice listening condition of a user from profile information of the user acquired via a communication network, a masking estimated value and a voice listening. The reproduction control means controls the voice enhancement processing means based on the conditions, and the voice enhancement processing means performs predetermined enhancement processing on the input voice by signal processing.

In order to achieve this object, a voice reproduction system according to the present invention includes a profile server for transmitting profile information of a registered user in response to an access request via a communication network, and AV information including voice information and voice. And a sending device that transmits the received voice signal, and performs a predetermined voice enhancement process using the profile information of the user acquired through the communication network while estimating the masking amount based on the estimated value of environmental noise for the received voice signal. And a sound reproducing device for reproducing the reproduced sound.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION (Embodiment 1) Hereinafter, Embodiment 1 of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an audio reproducing system according to the first embodiment of the present invention. In FIG. 1, 10 is a voice reproduction terminal, 20 is a profile server, 30 is a voice transmission device, 1 is a noise estimation unit, 2 is a masking amount estimation unit, 3 is a voice enhancement unit, 4 is a reproduction control unit, and 5 is voice listening. It is a condition setting unit. The operation will be described below.

First, the voice listening condition setting unit 5 reads the profile in which the voice listening condition of the user is recorded from the profile server 20 through the communication network to obtain the type and degree of hearing loss. The noise estimation unit 1 periodically obtains a quasi-steady noise estimation value from the ambient noise signal on the reception side picked up by the microphone. Next, the masking amount estimation unit 2
Calculates an estimated value of the masking amount that the received voice signal from the voice transmitting device 30 is affected by the noise estimated value. Then, the reproduction control unit 4 periodically updates the masking estimation value obtained by the masking amount estimation unit 2 every moment while compensating for the type and degree of hearing deterioration, and the voice enhancement processing unit 3
Adaptively control. The voice enhancement processing unit 3 outputs a voice in which the user can clearly hear the received voice signal by applying the voice enhancement process. As described above, if the voice listening condition is registered, the voice is taken out from the profile server via the communication network, so that even when the public terminal is used, it is possible to easily and accurately compensate for the hearing loss of the individual.

This embodiment is an embodiment assuming a voice communication device used outdoors, a portable information terminal capable of accessing voice information, and the like. Here, the user is assumed to be an elderly person or a handicapped person whose hearing is deteriorated, but by adjusting the degree of speech enhancement depending on the degree of hearing deterioration, it can be used by general hearing-impaired people. You may think that. A user who listens to the received voice has insufficient clarity due to the degree of his own hearing deterioration and the amount of environmental noise in the surroundings, which hinders communication. The cause is
There is a case of so-called hearing deterioration, which is called sensitivity deterioration or recruitment phenomenon in a user's specific frequency band,
There are two possible cases of masking the received voice by a noise source. Here, the user simply tries to solve the problem by manually increasing the volume of the audio reproduction device, but then, only a specific frequency band remains insufficient in sensitivity, or the volume of the sound is increased due to the recruitment phenomenon. If the sensitivity of the fish deforms in a non-linear manner, it is rather annoying. In addition, there are many disadvantages in any case, such as the need to automatically change the volume every moment according to the temporal change of the noise source.

Therefore, a signal processing method for clearly listening to a difficult-to-hear voice is adaptively controlled according to the hearing deterioration of the listener and the surrounding environmental noise.

First, as a user's voice listening condition, in the case of a typical elderly hearing-impaired person, in the case of a high-pitched sound type, the air-conducted hearing characteristics shown in Table 1 are obtained by an audiometer. It is also possible to use such data.

[0017]

[Table 1] To compensate for such hearing loss, one
It is often limited to the amount of correction that applies the half-gain rule that corrects about / 2 (see "Hearing aid utilization guide" P.
(111, Naoki Onuma), and it is necessary to perform fitting based on a subjective evaluation that reflects the user's sound quality preference. Therefore, a configuration as shown in FIG. 2 can be considered as a mechanism for reflecting the sound quality preference of the user. The configuration of FIG. 2 is obtained by adding a sound quality adjustment setting terminal to the configuration of FIG. The sound quality preference of the user is already registered in the audio listening condition of the user profile, and the reproduction control unit 4 reads this information and instructs to correct the processing parameter of the reproduction control unit based on the read sound quality adjustment setting terminal. I do. As a typical example, a user who prefers soft sound with emphasis on naturalness rather than clear voice modifies the degree of emphasis processing slightly, and conversely a user who prefers clear and crisp speech changes the degree of emphasis processing. It is conceivable to set it stronger. In this way, in addition to physical hearing compensation, it is possible to realize voice enhancement processing that reflects subjective preferences.

In estimating the frequency component of the noise, the sound becomes unnatural if compensation is made following the sudden noise. Therefore, the estimated noise value is based on the frequency envelope integrated with a time constant that is long to some extent. By obtaining, it is possible to obtain a natural reproduced sound without a sudden change in the sound after the sound enhancement processing.

Next, a method of estimating the masking amount will be described. First, the voice signal and the noise estimation value are subjected to frequency analysis on a frame-by-frame basis to obtain the noise power for each specific bandwidth.
Then, the powers of the voice and the noise are compared for each predetermined frequency bandwidth, the masking amount at which the received voice is simultaneously masked by the ambient environmental noise is estimated and output to the reproduction control unit 4. The reproduction control unit 4 controls the enhancement parameter of the voice enhancement unit for each frequency bandwidth. The voice emphasis processing unit 3 adjusts the emphasis processing parameter to change the emphasis degree of the reproduced voice.

Next, the operation of the masking amount estimation unit 2 will be described. Frequency analysis is performed on the voice signal and the estimated noise value, and the average energy for each critical bandwidth is obtained. Then, the masking amount estimation unit estimates the masking amount in the corresponding critical bandwidth. This value is shown, for example, in the literature: Murase, Nakamura, Iida, “Sound Quality Control Method Considering Masking by Ambient Noise” (Proceedings of the Acoustical Society of Japan, March 2, 3/10, 1997). As
It is expressed in the form of a function with the values of both the signal source and the noise source as parameters. Here, regarding the simultaneous masking effect,
For example, B.I. C. J. Since it is detailed in Chapter 3 (Seishin Shobo) of Moore's book “Introduction to Auditory Psychology”, translated by Kengo Ogushi, the explanation is omitted. The masking amount estimated value for each critical band obtained in this way is used as a parameter that determines the degree to which the speech enhancement processing unit performs the enhancement processing.

Further, the reproduction control section 4 determines the processing parameter of the voice enhancement processing in consideration of the masking amount estimated value and the deterioration of the hearing of the user.

Next, the audio signal processing performed by the audio enhancement processing section 3 will be described. The received voice signal is masked by the environmental noise around the listener to generate a component that cannot be heard auditorily, and therefore a process for compensating the masked frequency band is performed. First, the masking amount obtained for each critical bandwidth can be compensated for the effect of masking by adjusting the gain at a constant value in that band. However, when the number of points in the analysis frame is increased to improve the frequency resolution, it is effective as an average gain adjustment value in that section, but it is large in the transient part where the amplitude is not steady in the frame. The sound in the amplitude part may be over-amplified and may be offensive to the ear. Therefore, dynamic range compression processing that is often used in hearing aids and the like is applied.
FIG. 3 shows the configuration of the voice enhancement processing unit 3. First, it is assumed that masking compensation is performed by dividing a band into frequency bands having a critical bandwidth and performing dynamic range compression for each band. The band division unit 131 performs band division for each critical bandwidth, and the dynamic range compression processing unit 132 at the next stage determines the minimum audible level (HTL) based on the masking amount given for each band and sets the uncomfortable threshold (UCL). ) And the dynamic range compression processing for accommodating the audio signal between the two. The input / output characteristics as shown in FIG. 4 are shown as the dynamic range compression processing at this time. In this figure, for the masking compensation, a polygonal line type input / output characteristic which gives a gain increase of 20 dB when the input signal is 40 dB (HL) is given. With this characteristic, the input signal is 90 dB (H
L) is assumed to be UCL, and the output signal is not amplified beyond this value. Further, by performing such a non-linear gain adjustment, it becomes possible to perform a dynamic range compression process to a predetermined range, and as a result, masking compensation can be performed for each band. By considering the hearing ability of the user in this input / output characteristic, compensation can be performed at the same time.

Further, if the recording medium is a removable medium, it is possible to perform a process suitable for an accurate audio listening condition for each user, which is convenient for use in public telephones and public facilities. Even in the case of such a general-purpose audio reproducing device, since the user profile is read, it is possible to obtain a voice-enhanced voice in which a desired sound quality is added according to the type of voice information (conversation voice, music, news, etc.) to be heard. You can

It should be noted that other than the dynamic range compression, it is conceivable that the received voice becomes clear. For example, a graphic equalizer that performs an operation of limiting the upper limit value by a limiter operation can perform the same operation.

(Second Embodiment) A second embodiment of the present invention will be described below with reference to the drawings. FIG. 5 is a block diagram of the audio reproduction system according to the second embodiment of the present invention. In FIG. 5, 10 is a voice reproducing terminal, 20 is a profile server, 30 is a voice transmitting device,
1 is a noise estimation unit, 2 is a masking amount estimation unit, 3 is a voice enhancement processing unit, 4 is a reproduction control unit, 5 is a voice listening condition setting unit,
Reference numeral 6 is a terminal listening condition sending unit. The operation will be described below.

First, the voice listening condition setting unit 5 reads the profile in which the voice listening condition of the user is recorded from the profile server 20 through the communication network to obtain the type and degree of hearing deterioration. The noise estimation unit 1 periodically obtains a quasi-steady noise estimation value from the ambient noise signal on the reception side picked up by the microphone. Next, the terminal listening condition sending unit 6
Sends the estimated noise value and the acoustic reproduction condition on the terminal side to the voice transmitting side. Next, the masking amount estimation unit 2 obtains an estimated value of the masking amount in which the received voice signal from the voice transmitting device 30 is affected by the noise estimated value. Then, the reproduction control unit 4 adaptively controls the voice enhancement processing unit 3 based on the masking estimated value sent from the voice transmitting side while compensating for the type and degree of hearing deterioration. The voice enhancement processing unit 3 outputs a voice in which the user can clearly hear the received voice signal by applying the voice enhancement process. In this way, if the voice listening condition is registered, it can be easily taken out from the profile server through the communication network and compensated for the deterioration of the hearing ability of the individual even when using a public terminal.

In the present embodiment, the first embodiment has all the configurations on the receiving side to perform processing, whereas a part of the configuration is distributed to the voice transmitting side for transmitting voice. Will be realized in. Therefore, the final audio output is
Although the same effect can be obtained, by distributing a part of the processing to the transmitting side, the amount of calculation on the reproducing apparatus side on the receiving side can be reduced. On the other hand, although it is necessary to transmit the terminal listening condition from the receiving side to the transmitting side, the communication load is light because the amount of data is small. Therefore, for example, in a battery-driven device such as a portable information terminal, the present configuration that can reduce the amount of calculation is considered to be useful for the purpose of achieving size reduction and weight reduction.

It should be noted that other than the dynamic range compression, it is conceivable that the received voice becomes clear. For example, a graphic equalizer that performs an operation of limiting the upper limit value by a limiter operation can perform the same operation.

[0029]

As described above, according to the present invention, the masking compensation of ambient noise is performed on the received voice data, and at the same time, the voice emphasizing process considering the deterioration of the hearing ability of the user is performed. Thus, it is possible to realize a sound reproducing device capable of obtaining clear sound supplementing the above.

[Brief description of drawings]

FIG. 1 is a block diagram of an audio reproduction system according to a first embodiment of the present invention.

FIG. 2 is a block diagram of the audio reproduction system according to the first embodiment of the present invention.

FIG. 3 is an internal block diagram of a voice enhancement processing unit.

FIG. 4 is an input / output characteristic diagram of dynamic range compression processing.

FIG. 5 is a block diagram of an audio reproduction system according to a second embodiment of the present invention.

FIG. 6 is a block diagram of a conventional audio reproduction device.

[Explanation of symbols]

10 Audio playback terminal 20 profile server 30 Audio transmitter 1 Noise estimation section 2 Masking amount estimation unit 3 Speech enhancement processor 4 Playback control section 5 Audio listening condition setting section 6 Terminal listening condition sending unit

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Junichi Tagawa             1006 Kadoma, Kadoma-shi, Osaka Matsushita Electric             Sangyo Co., Ltd. (72) Inventor Tomomi Matsuoka             1006 Kadoma, Kadoma-shi, Osaka Matsushita Electric             Sangyo Co., Ltd.

Claims (5)

[Claims] 1. A profile server for transmitting profile information of a registered user in response to an access request via a communication network, a transmission device for transmitting audio information and AV information including audio, and a received audio signal. A voice reproduction device that reproduces a voice that has been subjected to a predetermined voice enhancement process by using the profile information of the user acquired via the communication network while estimating the masking amount based on the estimated value of environmental noise Playback system. 2. A profile server for transmitting profile information of a registered user in response to an access request via a communication network, an estimated value of environmental noise obtained from a receiving terminal side, and an AV including voice information and voice to be transmitted. A transmitting device that estimates the masking amount with information and transmits the masking amount at the same time, the received voice information and AV information including the voice, the masking amount estimated value, and the profile information of the user acquired via the communication network An audio reproduction system comprising: an audio reproduction device for reproducing the sound subjected to the emphasis processing. 3. A noise estimation means for estimating environmental noise, a voice enhancement processing means for performing predetermined enhancement processing on input speech by signal processing, and masking for a voice signal received based on an estimated value of environmental noise. Masking amount estimating means for estimating the amount, voice listening condition setting means for obtaining the user's voice listening condition from the profile information of the user acquired via the communication network,
A voice reproduction apparatus comprising: a reproduction control unit that controls a voice enhancement processing unit based on the masking estimated value and the voice listening condition. 4. A noise estimation means for estimating environmental noise, a voice enhancement processing means for performing predetermined enhancement processing by signal processing on an input voice, and a reception condition on a terminal side based on an estimated value of environmental noise. A terminal listening condition transmitting means for transmitting to the transmitting side, a voice listening condition setting means for obtaining the user's voice listening condition from the profile information of the user acquired via the communication network, and a transmitter side for obtaining the terminal listening condition and the transmitted voice. A voice reproduction apparatus comprising: a reproduction control unit that controls a voice enhancement processing unit based on a masking amount estimated value and the voice listening condition. 5. The reproduction control means controls the voice enhancement processing means on the basis of both the masking estimation value and the voice listening condition of the user, as well as the sound quality adjustment value reflecting the sound quality preference of the user. The audio reproducing device according to claim 3 or 4.