Conversation voice protection device
The conversational sound protection device effectively hides conversational voices by using voice recognition to generate masking sounds with matching consonants and different vowels, addressing the challenge of frequency spectrum mismatch in open spaces.
Patent Information
- Application Number
- JP2023083359
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-19
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing voice protection devices struggle to effectively hide conversational voices from third parties in open spaces with multiple noise sources due to difficulties in matching frequency spectra.
A conversational sound protection device that stores phoneme and syllable data, uses voice recognition to select phonemes and syllables with common consonants and different vowels, and outputs masking sounds to conceal the conversation.
Conceals conversational audio from third parties by generating masking sounds that mimic the speaker's phonemes and syllables, making it difficult for others to recognize the actual voice.
Smart Images

Figure 2024166925000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a conversation audio protection device that allows the contents of a conversation to be transmitted to a conversation partner without being leaked to a third party. [Background technology]
[0002] Patent Document 1 discloses an invention in which the frequency spectra of the voice of the person speaking detected by a microphone, the surrounding masking sounds, and background noise are made the same, so that a minimum amount of masking sound is emitted from a speaker so as not to disturb third parties around the person speaking. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2006-267174 A Summary of the Invention [Problem to be solved by the invention]
[0004] However, the invention described in Patent Document 1 had the risk of failing to reliably conceal conversation voices in open spaces or other locations where there are multiple noise sources, since it is difficult to match the frequency spectra of multiple noises from different sources with conversation voices.
[0005] SUMMARY OF THE PRESENT DISCLOSURE In consideration of the above, an object of the present invention is to provide a conversation voice protection device capable of concealing conversation voice from third parties. [Means for solving the problem]
[0006] In order to achieve the above object, the conversation voice protection device described in claim 1 includes a storage device that pre-stores source sound data that is information on phonemes and syllables of a speaker's voice, and a control unit that selects from the source sound data phonemes and syllables that have common consonants but different vowels with phonemes and syllables extracted by performing voice recognition on a voice signal of the speaker detected by a microphone, and outputs the selected phonemes and syllables from a speaker immediately after the phonemes and syllables extracted by voice recognition are spoken.
[0007] According to the conversation voice protection device described in claim 1, immediately after a phoneme and syllable spoken by a speaker, a phoneme and syllable having the same consonant as the phoneme and syllable but a different vowel are output, thereby concealing the conversation voice from third parties. Effect of the Invention
[0008] As described above, the conversation audio protecting device according to the present invention can conceal the conversation audio from third parties. [Brief description of the drawings]
[0009] [Figure 1] 1 is a schematic diagram showing a state in which a speaker wears a microphone, a control speaker, and a receiver which constitute a part of a conversation voice protection device according to an embodiment of the present invention. [Diagram 2] 1 is a block diagram showing an example of a specific configuration of a conversation voice protection device according to an embodiment of the present invention; [Diagram 3] (A) is an explanatory diagram showing an example of the result of voice recognition by a control computer, (B) is an explanatory diagram showing an example of a masking sound generated by the control computer, and (C) is an explanatory diagram showing an example of a composite waveform of the voice spoken by a speaker and the masking sound output from a control speaker. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0010] The conversation voice protecting device 100 according to this embodiment will be described below with reference to Fig. 1. Fig. 1 is a schematic diagram showing a state in which a speaker 10 wears a microphone 12, a control speaker 26, and a receiver 28 which constitute a part of the conversation voice protecting device 100 according to this embodiment. The microphone 12, the control speaker 26, and the receiver 28 are also worn by the other party who is speaking with the speaker 10 in the same way.
[0011] The microphone 12 is, for example, held by both ears of the speaker 10, and the voice uttered by the speaker 10 is inputted near the mouth of the speaker 10. The voice uttered by the speaker 10 is outputted from a receiver 28, which is a highly sound-insulating headphone or earphone worn on both ears by the person with whom the conversation is taking place.
[0012] Specifically, the microphone 12 is a dynamic microphone that converts sound into an electric signal by electromagnetic induction, or a condenser microphone that converts sound into an electric signal by a change in capacitance.
[0013] As described below, the control speaker 26 outputs a masking sound that conceals the voice uttered by the speaker 10. The control speaker 26 has directionality, and outputs the masking sound centrifugal from the mouth of the speaker 10 to prevent the microphone 12 from detecting the masking sound.
[0014] Fig. 2 is a block diagram showing an example of a specific configuration of the conversation voice protection device 100 according to the present embodiment. As shown in Fig. 2, the conversation voice protection device 100 according to the present embodiment includes a microphone 12 that converts a voice uttered by a speaker 10 into an analog electric signal, an amplifier 14 that amplifies the electric signal output from the microphone 12, an AD converter 16 that converts the analog signal amplified by the amplifier 14 into a digital signal, a storage device 18 that stores source sound data, which is data on phonemes and syllables of the speaker 10's voice, in advance, and a device that extracts phonemes and syllables from the digital signal of the speaker's voice output from the AD converter 16 by voice recognition, and stores in the storage device 18 phonemes and syllables that belong to the same row as the extracted phonemes and syllables, i.e., phonemes and syllables that have the same consonant but different vowels. The control computer 20 includes a control computer that refers to the source sound data and selects the selected phoneme and syllable, adjusts the pitch and duration of the selected phoneme and syllable to be approximately the same as the phoneme and syllable of the voice uttered by the speaker 10, and outputs the selected phoneme and syllable as a digital signal of a masking sound at a timing before the speaker 10 utters the next phoneme and syllable, a DA converter 22 that converts the digital signal of the masking sound output by the control computer 20 into an analog signal, an amplifier 24 that amplifies the analog signal under the control of the control computer 20 so that the volume becomes approximately the same as the phoneme and syllable uttered by the speaker 10, and a control speaker 26 that outputs the amplified analog signal as a masking sound. The control computer 20 is an example of a control unit. Note that when the speaker 10 utters a phoneme and syllable of the "a" row, the phoneme and syllable of the "a" row do not include a consonant. In this embodiment, since the phoneme and syllable of the "a" row do not include a consonant, it is considered that the requirement that the phoneme and syllable of the "a" row have a common consonant is satisfied.
[0015] Fig. 3(A) is an explanatory diagram showing an example of the result of voice recognition by the control computer 20. Fig. 3(A) shows that the voices of the speaker 10 are "ku", "ru", and "ma".
[0016] FIG. 3(B) is an explanatory diagram showing an example of a masking sound generated by the control computer 20. When the speaker 10 utters "ku", the masking sound is selected from the source sound data as one of "ka", "ki", "ke", and "ro" which have a common consonant with "ku". In FIG. 3(B), "ka" is selected as an example, and after the speaker 10 utters "ku", the masking sound is output at a timing before the speaker 10 utters the next "ru". Similarly, when the speaker 10 utters "ru", the masking sound is selected from the source sound data as one of "ra", "ri", "re", and "ro" which have a common consonant with "ru". In FIG. 3(B), "ra" is selected as an example, and after the speaker 10 utters "ru", the masking sound is output at a timing before the speaker 10 utters the next "ma". Similarly, when the speaker 10 utters "ma", one of "mi", "mu", "me" or "mo", which has a common consonant with "ma", is selected from the source sound data. In FIG. 3(B), "mu" is selected as an example and output at the timing after the speaker 10 utters "ma". It is also possible to randomly select a phoneme or syllable from the source sound data and output it after the speaker 10 utters one phoneme or syllable.
[0017] FIG. 3(C) is an explanatory diagram showing an example of a composite waveform of the voice uttered by the speaker 10 and the masking sound output from the control speaker 26. As described above, the control computer 20 adjusts the volume, pitch, and duration of the phonemes and syllables of the masking sound, which have the same consonants as the phonemes and syllables of the voice uttered by the speaker 10 but different vowels, to approximately the same as those of the phonemes and syllables of the voice uttered by the speaker 10. In addition, the control computer 20 outputs the phonemes and syllables of the masking sound, the volume, pitch, and duration of which have been adjusted, at a timing before the speaker 10 utters the next phoneme and syllable. As a result, a third party other than the speaker 10 and the other party who is talking to the speaker 10 recognizes "Kukarramam," which is the voice shown in the composite waveform of FIG. 3(C), making it difficult to hear the actual voice of the conversation between the speaker 10 and the other party.
[0018] As described above, according to this embodiment, phonemes and syllables having characteristics of the speaker's own voice are stored in advance in the storage device 18, and the control computer 20 selects, as masking sounds, phonemes and syllables that have the same consonants as the phonemes and syllables extracted by speech recognition of the speech signal of the speaker 10 detected by the microphone 12, but different vowels, from the source sound data in the storage device 18. The control computer 20 outputs the selected phonemes and syllables of the masking sounds immediately after the speaker 10 utters the phonemes and syllables extracted by speech recognition and before the speaker 10 utters the next phonemes and syllables, thereby preventing the voice of the speaker 10 from being recognized by a third party.
[0019] In this embodiment, the volume of the masking sound output from the control speaker 26 is adjusted by the amplifier 24, but this is not limiting. For example, the control computer 20 may adjust the volume of the masking sound output from the control speaker 26 by modulating the digital signal before it is converted into an analog signal by the DA converter 22. [Explanation of symbols]
[0020] 10. Voicer 12 Microphone 18 Storage device 20 Control computer (control unit) 26 Control speaker 100 Conversational Audio Protection Device
Claims
[Claim 1] A storage device that stores in advance source sound data, which is information on the phonemes and syllables of a speaker's voice; a control unit that selects, from the source sound data, phonemes and syllables that have a common consonant with the phonemes and syllables extracted by performing voice recognition on the voice signal of the speaker detected by a microphone, but have a different vowel, and outputs the selected phonemes and syllables from a speaker immediately after the utterance of the phonemes and syllables extracted by the voice recognition; A conversation voice protection device including:
Citation Information
Patent Citations
Masking device
JP2005084645A
Device and method for changing voice, and confidential communication system for voice information
JP2012008391A
Multi-purposed audible apparatus for canceling user''s voice
KR101659515B1
Smart mask capable of voice transmission control and voice transmission control system using the same
KR102242838B1
Device for protection of speech privacy
JP2006267174A
