Audio Processing Device

Abstract

To provide a voice processing apparatus capable of suppressing discomfort and clarifying voice.SOLUTION: A voice processing apparatus includes: an extraction part for acquiring guide voice including voice which a speaker speaks and extracting consonant parts; an acquisition part for acquiring ambient sound of a place where the guide voice is outputted; a determination part for determining whether or not the acquired ambient sound is a non-single-shot and whether or not frequencies of the ambient sound fall in a prescribed band; and a processing part for changing at least the consonant parts in the voice signal in a different manner according to a determination result and outputting the voice signal from a voice output part.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an audio processing device. [Background technology]

[0002] 2. Description of the Related Art There is known an audio processing device that corrects audio output from an audio output unit in accordance with ambient sound around the audio output unit (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2014-232245 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the technology of Patent Document 1, correction is performed in the same manner regardless of the frequency of the ambient sound, so depending on the frequency of the ambient sound, the corrected guidance voice may be emphasized, resulting in an unnatural sound.

[0005] The present invention has been made in view of the above, and has an object to provide a voice processing device that can reduce the sense of discomfort and clarify voice. [Means for solving the problem]

[0006] The voice processing device of the present invention includes an extraction unit that acquires a voice signal of a guidance voice including a speaker's voice and extracts a consonant portion, an acquisition unit that acquires ambient sounds at the location where the guidance voice is output, a judgment unit that judges whether the acquired ambient sounds are non-single and whether they are of a predetermined frequency band, and a processing unit that changes at least the consonant portion of the voice signal in a different manner depending on the judgment result and outputs the voice signal from a voice output unit. Effect of the Invention

[0007] According to the present invention, it is possible to provide a voice processing device that can reduce the sense of discomfort and make voice clear. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of a sound processing device according to the present embodiment. [Diagram 2] FIG. 2 is a functional block diagram illustrating an example of a voice processing device. [Diagram 3] FIG. 3 is a diagram illustrating an example of a guidance voice and a voice signal input to a voice input unit. [Figure 4] FIG. 4 is a diagram illustrating an example of the processing content of the processing unit in response to the ambient sound. [Diagram 5] FIG. 5 is a diagram illustrating an example of the processing content of the processing unit in response to the ambient sound. [Figure 6] FIG. 6 is a flowchart showing an example of the flow of processing in the voice processing device according to this embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of a voice processing device according to the present invention will be described with reference to the drawings. Note that the present invention is not limited to this embodiment. In addition, the components in the following embodiments include those that are replaceable and easy for a person skilled in the art, or those that are substantially the same.

[0010] Fig. 1 is a diagram illustrating an example of a voice processing device 100 according to the present embodiment. As illustrated in Fig. 1, the voice processing device 100 is used, for example, in public facilities, public transportation, etc., when a predetermined guidance voice is broadcast or the like.

[0011] The guidance voice is a voice including the voice of a speaker (person), such as the phrase "I will show you around." The voice of the speaker in the guidance voice includes vowels and consonants. The voice processing device 100 is disposed between a voice input unit 60 to which a voice signal, such as a voice signal of the guidance voice recorded in a memory or a voice signal obtained by collecting and converting the speaker's voice with a microphone, is input, and a voice output unit 70 such as a speaker that outputs a voice signal, and performs a predetermined process on the voice signal input from the voice input unit 60 and outputs the processed signal to the voice output unit 70.

[0012] The voice processing device 100 is a processing device including an arithmetic processing device such as a CPU (central processing unit) and a memory such as a ROM (read only memory) and a RAM (random access memory) or a storage device such as a storage. In the voice processing device 100, the arithmetic processing device performs arithmetic processing according to a computer program stored in the storage device.

[0013] 2 is a functional block diagram showing an example of the voice processing device 100. As shown in FIG. 2, the voice processing device 100 includes an acquisition unit 20, an extraction unit 10, a determination unit 30, a processing unit 40, and a storage unit 50.

[0014] The extraction unit 10 acquires a voice signal of a guidance voice input from the voice input unit 60, and extracts a consonant part from the acquired voice signal. For example, the extraction unit 10 performs frequency analysis on the voice signal by changing the time width for vowels and consonants, and detects the change in the power signal for each frequency, thereby making it possible to distinguish and detect the consonant part and the vowel part of the voice signal. For example, when the extraction unit 10 detects a part where the rate of change of the power signal is equal to or greater than a predetermined value under predetermined conditions, it can determine that the detected part is the start position of a consonant. When the consonant part in the voice signal is specified in advance based on time or the like, the extraction unit 10 can extract the consonant part of the voice signal based on the specified information. The extraction unit 10 may detect consonants using a known technique.

[0015] The acquisition unit 20 acquires ambient sounds at a location where the guidance voice is output, for example, at a location where the voice output unit 70 is located. The acquisition unit 20 can acquire ambient sounds collected by a sound collection unit 80 such as a sound collection microphone. The sound collection unit 80 is located at a location where the guidance voice from the voice output unit 70 can be heard. The sound collection unit 80 may be located at one location or at multiple locations.

[0016] The determination unit 30 determines whether the acquired ambient sound is non-single and whether the frequency is in a predetermined band. The determination unit 30 performs frequency analysis on the acquired ambient sound by Fourier transform or the like. The determination unit 30 determines whether the ambient sound is non-single based on the result of this frequency analysis. In this embodiment, the ambient sound being non-single means, for example, that the ambient sound has periodicity, continuity, or a time length of a predetermined period or more. When the ambient sound has periodicity, continuity, or a time length of a predetermined period or more, the guidance voice output from the voice output unit 70 is more likely to be difficult to hear than when the ambient sound does not have periodicity, continuity, and a time length of a predetermined period or more, that is, when the ambient sound is short and sporadic. The determination unit 30 can determine that the ambient sound is non-single when, for example, the ambient sound has a period, occurs multiple times in succession, or continues for a certain percentage or more of the target time.

[0017] When the determination unit 30 determines that the ambient sound is not a single sound, it determines whether the ambient sound has a frequency of a predetermined band. In this embodiment, the predetermined band can be, for example, a frequency band of human speech. In human speech, for example, for vowels, a frequency band of 300 Hz to 2000 Hz can be set as a vowel band. In addition, for consonants, a frequency band of 3500 Hz to 8000 Hz can be set as a consonant band. In addition, a frequency band of 2000 Hz to 3500 Hz can be set as a mixed band of vowels and consonants. Therefore, the frequency band of human speech can be set as a band of 300 Hz to 8000 Hz including the vowel band, the consonant band, and the mixed band. That is, in this embodiment, the predetermined band can be, for example, a band of 300 Hz to 8000 Hz.

[0018] The processing unit 40 changes at least the consonant portion of the voice signal of the guidance voice in a different manner according to the result of the determination, and outputs the voice signal with the changed consonant portion from the voice output unit 70.

[0019] When the determination unit 30 determines that the ambient sound is not single and has a frequency in a predetermined band, the processing unit 40 performs at least one of the following processes: amplifying the level of the consonant part in the voice signal of the guidance voice during the period in which the ambient sound is generated, extending the length of the consonant part, and inserting a silent period before the consonant part. In this process, when the voice is a prepared voice, and vowel period information, consonant period information, etc. of the voice are added, and the consonant part of the voice signal is specified in advance, the processing unit 40 can detect the start position of the consonant part based on the specific information and increase the amplification rate at the start position. In addition, when the voice is not prepared in advance, when vowel period information, consonant period information, etc. of the voice are not added, or when the consonant part of the voice signal is not specified in advance, the processing unit 40 can detect a part where the relative change rate of the voice signal is equal to or higher than a certain level from the output of multiple filtering processes, and determine that the detected part is the start position of the consonant part, and increase the amplification rate at that position.

[0020] When the determination unit 30 determines that the ambient sound is not single and has a frequency wider than the frequency of the predetermined band, the processing unit 40 performs at least one of a process of amplifying the start position of the consonant part in the voice signal of the guidance voice during the period when the ambient sound is generated and a process of amplifying the level of the entire voice signal including the consonant part. In this process, the processing unit 40 can detect the start position of the consonant part based on the specification information when the voice is a prepared voice, the voice is provided with vowel period information, consonant period information, etc., and the consonant part of the voice signal is specified in advance, and can increase the amplification factor at the start position. In addition, when the voice is not prepared in advance, the voice is not provided with vowel period information, consonant period information, etc., and the consonant part of the voice signal is not specified in advance, the processing unit 40 can detect a part where the relative change rate of the voice signal is equal to or higher than a certain level from the output of multiple filtering processes, determine the detected part to be the start position of the consonant part, and increase the amplification factor at that position.

[0021] The storage unit 50 stores various information. The storage unit 50 has storage such as a hard disk drive or a solid state drive. An external storage medium such as a removable disk may be used as the storage unit 50. The storage unit 50 can store programs, data, and the like for performing the process of acquiring ambient sounds in the acquisition unit 20, the process of extracting consonant parts in the extraction unit 10, the judgment process in the judgment unit 30, and the conversion process in the processing unit 40.

[0022] The memory unit 50 also stores a voice processing program that causes the computer to execute the following processes: acquiring ambient sounds at the location where the guidance voice is output; extracting consonant portions from the voice signal of the guidance voice; determining whether the acquired ambient sounds are non-single sounds and whether they are in a predetermined frequency band; and modifying at least the consonant portions of the voice signal in a different manner depending on the determination result, and outputting the voice signal from the voice output unit 70.

[0023] Next, an example of the operation of the voice processing device 100 configured as above will be described. The extraction unit 10 acquires a voice signal of a guidance voice and extracts a consonant part from the voice signal. The acquisition unit 20 acquires surrounding sounds collected by a sound collection unit 80 such as a sound collection microphone.

[0024] FIG. 3 is a diagram showing an example of a guidance voice and a voice signal input to the voice input unit 60. FIG. 3 shows an example of a voice in which a speaker speaks "I will guide you." The guidance voice is not limited to Japanese, and may be a foreign language voice such as English. In the guidance voice shown in FIG. 3, "go," "na," "shi," "ma," and "su" are combinations of consonants and vowels, "n" is composed of only consonants (nasal consonants), and "a" and "i" are composed of only vowels. In the following FIG. 3 to FIG. 5, oval marks indicate consonants, and rectangular marks with rounded corners indicate vowels.

[0025] The determination unit 30 determines whether the acquired ambient sound is non-single and whether the sound is in a predetermined frequency band. The processing unit 40 changes at least the consonant part of the audio signal in a different manner according to the determination result, and outputs the audio signal with the changed consonant part from the audio output unit 70. Figures 4 and 5 are diagrams that typically show examples of the processing contents of the processing unit according to the ambient sound.

[0026] The example shown in FIG. 4 shows a case where the ambient sound is non-single and has a frequency in a predetermined band. As shown in FIG. 4, the processing unit 40 calculates, for example, a temporal change in the amplitude of the ambient sound, and calculates a period in which the amplitude is equal to or greater than a predetermined value. In the example shown in FIG. 4, the processing unit 40 can calculate that, for example, the entire period of the voice signal of the guidance voice for "n", the entire period including consonants for "na" and "shi", and the period corresponding to the consonant part for "ma" are periods in which the ambient sound is non-single and has a frequency in a predetermined band. The processing unit 40 can perform a process of amplifying the level (amplitude) of the consonant part of the voice signal and extending its length during the calculated periods.

[0027] In FIG. 4, the level of the entire "n" and the consonant parts "na" and "shi" in the period calculated as the ambient sound being non-single and having a frequency in the predetermined band is amplified and extended. The processing unit 40 can also perform a process of inserting a silent time in the period preceding the consonant part in the calculated period. In FIG. 4, a silent time is inserted in the period preceding the consonant part of "shi". The processing unit 40 may perform only one or two of the process of amplifying the level of the consonant part of the audio signal, the process of extending the length of the consonant part, and the process of inserting a silent time in the period preceding the consonant part. Note that the process of extending the length of the entire "n" and the consonant parts "na" and "shi" and the process of inserting a silent time in the period preceding the consonant part of "shi" causes the consonant part of "ma" to fall outside the period in which the ambient sound is non-single and has a frequency in the predetermined band. For this reason, the processing unit 40 can prevent the change process from being performed on the consonant part of "ma".

[0028] The example shown in FIG. 5 shows a case where the ambient sound is non-single and has a frequency wider than the frequency of a predetermined band. As shown in FIG. 5, the processing unit 40 calculates, for example, a temporal change in the amplitude of the ambient sound, and calculates a period in which the amplitude is equal to or greater than a predetermined value. In the example shown in FIG. 5, the processing unit 40 can calculate that, for example, the entire period of the voice signal of the guidance voice, "n" and "i", the entire period including the consonants of "na" and "shi", and the period corresponding to the consonant part of "ma", are periods in which the ambient sound is non-single and has a frequency wider than the predetermined band. The processing unit 40 can perform a process of amplifying the level of the start point of the consonant part of the voice signal during the calculated period. The processing unit 40 can also perform a process of amplifying the level of the entire voice signal including the consonant part of the voice signal.

[0029] In FIG. 5, during the period in which the ambient sound is calculated to be non-single and to have a frequency wider than the predetermined frequency band, the level at the start of "n" and the levels at the start of the consonant parts of "na," "shi," and "ma" are amplified. Furthermore, the level is amplified over the entire period of "n," and the level is amplified over the entire period including the consonant parts of "na" and "shi." Furthermore, the level of only the consonant part of "ma" is amplified. For "i," the entire period (vowel part) is amplified. The processing unit 40 may perform only one of the process of amplifying the level at the start of the consonant part of the audio signal and the process of amplifying the level of the entire audio signal including the consonant part of the audio signal.

[0030] 4 and 5, the processing unit 40 performs a smoothing process on the audio signal, and outputs the audio signal from the audio output unit 70. The audience listening to the guidance audio output from the audio output unit 70 can hear the guidance audio, which has been processed in a manner corresponding to the ambient sound, clearly and without any sense of discomfort, even in an environment where ambient sound is generated.

[0031] Next, an example of the operation of the voice processing device 100 according to this embodiment will be described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of the flow of processing in the voice processing device 100 according to this embodiment. As shown in Fig. 6, the extraction unit 10 acquires a voice signal of a guidance voice, and extracts a consonant part from the acquired voice signal (step S10). The acquisition unit 20 acquires ambient sounds collected by a sound collection unit 80 such as a sound collection microphone (step S20).

[0032] The determination unit 30 determines whether the acquired ambient sound is non-single (step S30). If it is determined in step S30 that the ambient sound is not non-single (No in step S30), the process proceeds to step S70 described below. If it is determined in step S30 that the ambient sound is non-single (Yes in step S30), the determination unit 30 determines whether the ambient sound has a frequency within a predetermined band (step S40).

[0033] In step S40, if it is determined that the ambient sound has a frequency within the range of the predetermined band (Yes in step S40), processing is performed in a first manner (step S50). The first manner processing in step S50 includes at least one of processing to amplify the level of the consonant part in the audio signal during the period when the ambient sound is generated, processing to extend the length of the consonant part, and processing to insert a silent period before the consonant part.

[0034] In step S40, if it is determined that the ambient sound has a frequency range wider than the frequencies within the range of the predetermined band (No in step S40), processing is performed in a second manner (step S60). The second manner processing in step S60 includes at least one of a process of amplifying the start position of a consonant part during the period when the ambient sound is generated and a process of amplifying the level of the entire audio signal including the consonant part.

[0035] After performing the first aspect of processing in step S50, after performing the second aspect of processing in step S60, or if it is determined in step S30 that the ambient sound is not non-single (No in step S30), the processing unit 40 outputs an audio signal from the audio output unit 70 (step S70) and terminates the processing.

[0036] As described above, the voice processing device 100 according to this embodiment includes an extraction unit 10 that acquires a guidance voice including a speaker's voice and extracts the consonant portion, an acquisition unit 20 that acquires ambient sounds at the location where the guidance voice is output, a judgment unit 30 that judges whether the acquired ambient sounds are non-single sounds and whether they are in a predetermined frequency band, and a processing unit 40 that changes at least the consonant portion of the voice signal in a different manner depending on the judgment result and outputs the voice signal from the voice output unit 70.

[0037] According to this configuration, at least the consonant portion of the audio signal is changed in a different manner depending on whether the ambient sound is non-single and whether it is within a specified frequency band. Therefore, in the processed audio guidance, the processed parts are clarified while suppressing the emphasis, resulting in a sound that reduces the sense of discomfort.

[0038] In the voice processing device 100 according to this embodiment, when the determination unit 30 determines that the ambient sound is non-single and has a frequency in a predetermined band, the processing unit 40 performs at least one of the following processes: amplifying the level of the consonant part in the audio signal during the period in which the ambient sound is generated, extending the length of the consonant part, and inserting a silent period before the consonant part. With this configuration, when the ambient sound is non-single and has a frequency in a predetermined band, it is possible to reduce discomfort and clarify the voice.

[0039] In the voice processing device 100 according to the present embodiment, when the determination unit 30 determines that the ambient sound is non-single and has a frequency range wider than the predetermined frequency band, the processing unit 40 performs at least one of a process of amplifying the start position of the consonant portion during the period in which the ambient sound is generated and a process of amplifying the level of the entire voice signal including the consonant portion. With this configuration, when the ambient sound is non-single and has a frequency range wider than the predetermined frequency band, it is possible to reduce discomfort and clarify the voice.

[0040] In the voice processing device 100 according to the present embodiment, the processing unit 40 increases the amplification factor at the start position of the consonant part when the consonant part of the voice signal is specified in advance, and increases the amplification factor at a position where the relative change rate of the voice signal is equal to or greater than a certain level when the consonant part of the voice signal is not specified in advance. With this configuration, the start position of the consonant part of the voice signal can be amplified with high accuracy.

[0041] In the voice processing device 100 according to the present embodiment, the predetermined band is a frequency band of human speech. With this configuration, ambient sounds caused by human speech are distinguished from other ambient sounds, and at least the consonant parts of the voice signal are changed in different ways, so that the guidance voice including the speaker's voice can be made to have a less unnatural sounding feel.

[0042] The technical scope of the present invention is not limited to the above-described embodiment, and appropriate modifications can be made without departing from the spirit of the present invention.

[0043] This disclosure includes matters that contribute to the realization of the SDGs' goal of "creating sustainable urban and urban areas" and contributing to the safety and security of public facilities. [Explanation of symbols]

[0044] 10...extraction unit, 20...acquisition unit, 30...determination unit, 40...processing unit, 50...storage unit, 60...audio input unit, 70...audio output unit, 80...audio collection unit, 100...audio processing device

Claims

1. an extraction unit that acquires a voice signal of a guidance voice including a voice of a speaker and extracts a consonant part; An acquisition unit that acquires ambient sound around a location where the guidance voice is output; a determination unit that determines whether the acquired ambient sound is non-single and whether the acquired ambient sound is in a predetermined frequency band; a processing unit that outputs, from an audio output unit, an audio signal in which the consonant portion has been changed in a different manner; Equipped with When the determination unit determines that the ambient sound is not a single sound and has a frequency in a predetermined band, the processing unit identifies a start position of the consonant portion during a period in which the ambient sound is generated, and increases an amplification factor when performing processing to amplify the consonant portion at the start position of the consonant portion. Audio processing device.

2. An extraction unit that acquires a voice signal of a guidance voice including a speaker's voice and extracts a consonant part; An acquisition unit that acquires ambient sound around a location where the guidance voice is output; a determination unit that determines whether the acquired ambient sound is non-single and whether the acquired ambient sound is in a predetermined frequency band; a processing unit that outputs, from an audio output unit, an audio signal in which the consonant portion has been changed in a different manner; Equipped with When the determination unit determines that the ambient sound is not a single sound and has a frequency band wider than a predetermined frequency band, the processing unit identifies a start position of the consonant portion during the period in which the ambient sound is generated, and increases an amplification factor when performing a process of amplifying the consonant portion at the start position of the consonant portion. Audio processing device.

3. The processing unit increases an amplification factor when performing processing to amplify the consonant portion at a start position of the consonant portion when the consonant portion of the audio signal is specified in advance, and increases an amplification factor when performing processing to amplify the consonant portion at a position where a relative change rate of the audio signal is equal to or greater than a certain value when the consonant portion of the audio signal is not specified in advance.

3. The audio processing device according to claim 1 or 2.

4. The predetermined band is a frequency band of human speech.

3. The audio processing device according to claim 1 or 2.

Images