JPH02230898A - Voice reproduction system - Google Patents

Abstract

PURPOSE:To allow the content of conversation to be listened by concerned persons only by synthesizing a voice at a frequency band other than the voice effecting the understanding of the content of conversation with a masking voice and reproducing the synthesized voice through a nondirectional speaker. CONSTITUTION:The frequency band is split (split into A and B) at a frequency split circuit 11 to an audio signal of a monaural sound X(=A+B) transferred from an input terminal 10. For example, the voice A of a noticed characteristic frequency region is reproduced by a directivity speaker 12 and the voice B of the other frequency is transferred to a making sound addition section 25. The masking sound addition section 25 adds the 'masking sound C with respect to A' to the voice B and the synthesized sound (B+C) is reproduced by the nondirectional speaker 13. Thus, the directivity speaker 12 acts it that the content of conversation is not listened to a person not concerned resident therearound (not understood) but listened to a person concerned 15.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] This invention relates to an audio reproduction method used, for example, in a teleconferencing system. ``Prior art'' When realizing a communication conference system such as an audio conference or a video conference, playback devices are often used for long periods of time due to the nature of the conference, and using a telephone receiver or earphones places a heavy burden on listeners. The first problem arises, which is the feeling of pressure and confusion. On the other hand, it is possible to use a loudspeaker instead of a telephone receiver or earphones, but in this case, there are drawbacks that are not seen as problems with telephone receivers or earphones.

In other words, when a teleconference is held in an environment where there are people (non-parties) who are unrelated to the conference, the second problem arises that the content of the conference being played back may be heard by the non-parties. ．． To address this issue, a playback method can be considered in which the loudspeaker has narrow directivity. This takes advantage of the fact that if only the parties involved are located within the listening range reduced by the narrow directivity, it becomes difficult for non-parties unrelated to the conference to hear the conference. However, with this conventional method, it is difficult to achieve directivity in the bass range, where the wavelength of sound is large, and if this is achieved forcibly, the speaker diameter becomes large, resulting in an increase in the size of the device. To solve this problem, the frequency bands that affect the intelligibility of conversation content (for example, the formant frequency band) are played back using narrow directional speakers, and the other frequency bands are played back using omnidirectional speakers. A method for reproducing the data has been proposed. This invention improves this method to further reduce the intelligibility of the conversation to surrounding non-parties, that is, to prevent the conversation from leaking to the surroundings. The purpose of the present invention is to eliminate the above-mentioned conventional drawbacks by using a narrow directional speaker to reproduce the frequency band that affects the intelligibility of conversation content, and by utilizing the masking effect to make it easier for surrounding non-parties to hear the sound. An object of the present invention is to provide an audio reproduction method for reproducing masking sound for reducing the intelligibility of conversation contents using an omnidirectional speaker.

"Means for Solving the Problem" This invention divides the frequency band into two according to the degree to which it affects the intelligibility of conversation content, and narrowly directs the sound in the frequency band that affects the intelligibility (directional frequency band). A masking sound that masks the sound played from the narrow directional speaker is added to the sound in a frequency band that does not affect the intelligibility (omnidirectional frequency band) that is played back by the speaker, and synthesized. The main feature is that it can be played back. Frequency bands that affect the intelligibility of conversation content include the frequency band of formants, which are component sounds that characterize the five vowels of Japanese, and the frequency band that allows identification of individuals. ``Embodiment'' FIG. 1 is a first example of a system showing the features of this invention. Monaural sound X transferred from input terminal 10 (
=A+B) frequency division circuit 1 for the audio signal
Frequency band division (dividing into A and B) is performed using l, and for example, audio A with the formant frequency of interest (including frequencies around it) is reproduced by directional speaker l2, and audio B with other frequencies is reproduced. It is transferred to the masking sound adding section 25.

The masking sound addition section 25 has a masking sound input terminal 1
The "masking sound C for A" transferred from 5 is added to sound B, and this synthesized sound (B + C) is omnidirectional.
It is played back from the speaker 13. Masking sound C is another sound similar to sound A, for example.

The directional speaker 12 allows the conversation content to be transmitted to surrounding non-parties 14.
The non-directional speaker 13 has the role of making the sound audible to the person concerned 15 without being heard (understood) by others, and the omnidirectional speaker 13 has the role of making the sound appear wider, louder, and thicker. That is, the party 15 listens to the audio X+C played from the speaker 12 and the speaker l3.
(=A+B+C), the non-party 14 uses the speaker l
By controlling the volume of the masking sound C appropriately, the intelligibility of the conversation content of the party 15 to the non-party 14 can be reduced. Figure 2 is a second system example showing the features of this invention. Regarding the stereo audio signals (L signal and R signal) transferred from the input terminal 10, the L signal and R signal are respectively processed according to the playback operation described in FIG. The former sound is transmitted through the L signal directional speaker 12. The L and R signal directional speakers 12R are respectively reproduced, and the latter sounds (L signal, R signal) are transferred to the masking sound adding section 25. The masking sound adding section synthesizes the masking stereo sound transferred from the masking sound input terminal 15 and the stereo sound transferred from the frequency division circuit 11 for each corresponding R signal and R signal, and generates this synthesized sound. Omnidirectional speaker 1 for each L signal
3L and the R signal omnidirectional speaker 13R. Participant 15 can listen to the content of the conversation by localizing the sound image. In conventional technology, the speaker tends to become larger because the entire frequency band has directivity, but with this invention, the directivity is increased in the frequency band that affects the intelligibility of conversation content. Furthermore, a major difference from conventional methods is that it utilizes masking effects to increase the effect of intelligibility. FIG. 3 is a block diagram showing the configuration of an embodiment of this invention. In response to a command from the control unit 24, the frequency band setting unit 23 extracts directional band setting data and non-directional band setting data that are predetermined in consideration of male voices, female voices, conversation sound effects, etc. The data is transferred to the playback section 21 and the omnidirectional band extraction and playback section 22. The directional band extraction and reproduction unit 2l performs initial settings based on the directional band setting data, and after completing the setting, sends a completion notification to the frequency band setting unit 2.
Transfer to 3. At the same time, the omnidirectional band extraction and reproducing section 22 performs initial setting based on the omnidirectional band setting data, and after completing the setting, transfers a completion notification to the frequency band setting section 23. After receiving the completion notification received from the directional band extraction and playback unit 2l and the completion notification received from the omnidirectional band extraction and playback unit 22, the frequency band setting unit 23 issues an activation start command to the directional band extraction and playback unit. 21 and the omnidirectional band extraction and reproduction section 22. After the notification is completed, the directional band extraction and playback unit 21 extracts only the initially set audio band for the audio signal sent from the input terminal 10 and plays the audio via the directional speaker 12. ．． On the other hand, the omnidirectional band extraction and playback unit 22 extracts only the initially set audio and transfers the audio to the masking sound addition unit 25. The masking sound addition unit 25 adds sound to the audio transferred from the omnidirectional band extraction and playback unit 22.
The masking sound input from the masking sound input terminal 15 is added and synthesized, and the synthesized sound is output to the omnidirectional speaker 13.
Play via . "Effects of the Invention" As explained above, according to the audio reproduction method according to the present invention, for example, the voice in the formant frequency band (including frequencies in the vicinity), which affects the intelligibility of conversation content, can be reproduced.
It is played back through a directional speaker, and the sound in other frequency bands is synthesized by adding masking sound, and the synthesized sound is played back through an omnidirectional speaker. It has the advantage of being free, and
The masking effect has the advantage that the conversation can only be heard by the person concerned, but not by those around them. Furthermore, instead of providing directivity for all frequencies in the human vocal range of 100Hz to 8000Hz,
Since it provides directivity only in some frequency bands from Hz to 2000Hz, the speaker (speaker aperture)
This has the advantage that it can be made smaller and more economical. Another advantage of using omnidirectional speakers is that you can create a sense of spaciousness and depth in the sound.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a first system example showing the features of this invention, FIG. 2 is a block diagram showing a second system example showing features of this invention, and FIG. 3 is a block diagram showing a second system example showing features of this invention. It is a block diagram showing the configuration. 1 Figure 1: Hardwood diagram

Claims (1)

[Claims] (1) An audio reproduction system that has a directional speaker with directivity and an omnidirectional speaker with omnidirectionality, which has a directional frequency that gives directionality when playing back an input audio signal. means for setting a band and an omnidirectional frequency band having no directivity; means for extracting an audio signal in the directional frequency band from the input audio signal and reproducing it using the directional speaker; means for extracting an audio signal in the omnidirectional frequency band from an input audio signal; and a means for adding an audio signal for masking the audio signal in the omnidirectional frequency band to the audio signal in the omnidirectional frequency band; 1. A sound reproduction method comprising means for reproducing using a speaker.