JP4915575B2 - Audio transmission system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice transmission system capable of improving a reproducibility of voice, while suppressing generation of abnormal sound caused by an transmission error. <P>SOLUTION: The voice transmission system is equipped with a plurality of terminal devices 2 including: a first difference calculation means 36 for calculating an absolute value A<SB>k</SB>of difference between a quantization value Q<SB>k</SB>of a sample point S<SB>k</SB>of a digital voice signal, and a quantization value Q<SB>k-1</SB>of the last sample point S<SB>k-1</SB>; a compensation reference value calculation means 37 for calculating a maximum value M of the absolute value A<SB>k</SB>for one voice frame in which a voice data made by a coded digital voice signal is divided on a time axis; a transmission means 34 for transmitting a packet including the voice frame and the maximum value M; a receiving means 42 for receiving the packet; an abnormal sound coping processing means 45 for compensating the quantization value Q<SB>k</SB>of the sample point S<SB>k</SB>in which the absolute value A<SB>k</SB>calculated by the digital voice signal in which the voice frame of the received packet is decoded, exceeds the maximum value M, to an average value of the quantization value Q<SB>k-1</SB>and Q<SB>k+1</SB>of the before and the after sample points S<SB>k-1</SB>and S<SB>k+1</SB>. <P>COPYRIGHT: (C)2009,JPO&amp;INPIT

Description

  The present invention relates to an audio transmission system that transmits an audio signal via a signal line.

  2. Description of the Related Art Conventionally, an audio transmission system including an audio transmission terminal device that transmits an audio signal by a packet and an audio reception terminal device that outputs audio based on the received audio signal is provided, for example, installed at the entrance of a dwelling unit. It is used in door phone systems that transmit and receive audio signals via signal lines between a door phone slave unit and a base unit installed in a dwelling unit, voice call systems that use cellular phones, PHS, and the like.

  In such an audio transmission system, a transmission error may occur due to transmission loss of a transmission line (for example, a signal line) or external noise. When a transmission error occurs, the audio signal transmitted by the audio transmission terminal device is different from the audio signal received by the audio reception terminal device, which causes abnormal sound in the audio output by the audio reception terminal device. There was a possibility that the problem of mixing would occur.

Therefore, when transmitting an audio signal from the audio transmitting terminal device to the audio receiving terminal device, for example, when transmitting the audio data obtained by encoding the audio signal by information source into a plurality of packets, Is provided with an error detection code such as cyclic redundancy check (CRC), so that the voice reception terminal device detects the error in the voice data of the packet and decodes the voice data in which the error is detected. So-called muting, in which the audio signal obtained by suppressing the audio signal, for example, significantly reducing the amplitude of the audio signal or setting it to 0, does not output the audio corresponding to the audio signal from the audio receiving terminal device. It has been proposed to perform processing (see, for example, Patent Document 1).
Japanese Patent No. 3183490

  In patent document 1, it can reduce that an abnormal sound mixes with the audio | voice which an audio | voice receiving terminal device outputs. However, in the muting process as described above, not only a wrong place in the sound signal but also a sound by a certain range of sound signals including a place not wrong in the sound signal is suppressed. For this reason, in the muting process as in Patent Document 1, although the generation of abnormal noise can be suppressed, there is a possibility that the reproducibility of the voice in the voice transmission system is deteriorated.

  The present invention has been made in view of the above points, and an object of the present invention is to provide an audio transmission system capable of improving the reproducibility of audio while suppressing the generation of abnormal noise due to a transmission error.

In order to solve the above problems, the invention of claim 1, a voice transmission terminal apparatus for transmitting an audio signal by the packet, an audio receiving apparatus for outputting sound based on the received audio signal includes a plurality The plurality of voice transmitting terminal devices and the voice receiving terminal device are connected via the same signal line, and any one of the plurality of voice transmitting terminal devices and the voice receiving terminal devices is connected to the voice transmitting terminal device. And an audio transmission system including an arbitrary audio reception terminal device different from the audio transmission terminal device so that audio transmission is possible, and the audio transmission terminal device converts the received sound wave into an analog audio signal composed of an electrical signal. A / D conversion means for converting an analog voice signal obtained from voice input means for conversion into a signal into a digital voice signal having a predetermined sampling frequency and quantization width; and each sampling point of the digital voice signal Encoding means for generating speech data by encoding the child values in a predetermined format, and the absolute value of the difference between the quantized value of the sample point of the digital audio signal and the quantized value of the previous sample point A first difference calculation means for calculating the reference value and a correction reference value calculation for calculating the maximum absolute value calculated by the first difference calculation means for one voice frame obtained by dividing the voice data created by the encoding means on the time axis Means and a transmission means for creating a packet having the one voice frame of the voice data and the maximum value calculated by the correction reference value calculation means for the one voice frame and transmitting the packet to the voice receiving terminal device. The terminal device includes: a receiving unit that receives the packet transmitted by the transmitting unit; and a decoding unit that decodes one audio frame of audio data included in the packet received by the receiving unit to obtain a digital audio signal. A noise countermeasure processing means for performing noise countermeasure processing on the digital voice signal obtained from the decoding means, and converting the digital voice signal subjected to the noise countermeasure processing by the noise countermeasure processing means into an analog voice signal. And a D / A conversion means that performs second noise calculation processing means for calculating an absolute value of a difference between the quantized value of the sample point of the digital audio signal and the quantized value of the immediately preceding sample point. The difference calculation means, the comparison means for comparing the absolute value calculated by the second difference calculation means with the maximum value of the received packet, and the difference between the quantized value of the previous sample point by the comparison means The quantized value Q _k (k is an integer of 2 or more) of the sample point determined that the absolute value has exceeded the maximum value is quantized to the sample point immediately before the quantized value Q _k in the time series. sample points immediately after the value _{Q k-1} and the quantized value _{Q k 'Is} corrected to, the average value Q _k' average Q _k using the quantization value Q _{k + 1} absolute value A _k of the difference between the quantization value Q _k-1 sample point of the immediately preceding and ₌ | Q _k When '−Q _k−1 | exceeds the maximum value, the quantized value Q _k−2 immediately before the quantized value Q _k−1 of the immediately preceding sample point and the quantized value Q _k To the new average value Q _k ″ = (Q _k−2 + Q _k−1 + Q _{k + 1} + Q _{k + 2} ) / 4 using the quantized value Q _{k + 2} immediately after the quantized value Q _{k + 1} of the sample point immediately after The quantized value Q _k is corrected, and the absolute value of the difference between the new average value Q _k ″ and the quantized value Q _k−1 of the immediately preceding sample point = | Q _k ″ −Q _k−1 | When the maximum value is exceeded, the quantized value Q _k−3 that is just before the further previous quantized value Q _k−2 and the amount that is immediately after the quantized value Q _k−2 Correction for performing noise countermeasure processing for correcting the quantized value Q _k to a new average value using a total of six values obtained by adding the quantized value Q _{k + 3} immediately after the child value Q _{k + 2} Means.

  According to the first aspect of the present invention, the difference between the quantized value at each sample point of the digital audio signal decoded by the decoding means and the quantized value at the previous sample point is an abnormal sound due to a transmission error. This is corrected to a value that does not exceed the maximum absolute value of the difference between the quantized values allowed in the original digital audio signal in which no noise occurs, which may cause abnormal noise in the decoded digital audio signal. Such a sudden change in the quantization value can be suppressed, the occurrence of abnormal noise due to transmission errors can be suppressed, and the quantization value is corrected only for the sample point that is causing the abnormal noise. Compared with muting processing that does not output (mute) audio by greatly reducing the signal amplitude or reducing it to 0, the original digital audio signal waveform remains, so the reproducibility of audio can be improved.In addition, the correction means corrects the quantized value at the sample point to be corrected to a value interpolated using the quantized values of other sample points in the predetermined range including the sample point. Since the quantized value in can be brought close to the quantized value (correct quantized value) at the sample point of the original digital speech signal, the speech reproducibility can be further improved.

  The present invention has an effect that the reproducibility of voice can be improved while suppressing the generation of abnormal noise due to a transmission error.

  As shown in FIG. 1A, the audio transmission system according to the present embodiment includes a plurality of units that transmit and receive audio signals by packets (hereinafter referred to as “audio transmission packets”) via a signal line Ls serving as a transmission path. , And a time slot TSi (i = 1, 2,..., N, which will be described later), which is connected to each terminal device 2 by a signal line Ls and stores packets for call control and voice transmission between the terminal devices 2. And a main device 1 that transmits a synchronization signal SY for defining the above. Each terminal device 2 is given a unique identification code (ID), and each terminal device 2 can be identified by this identification code (selection of a signal taken out from the signal line Ls is possible). Further, as the signal line Ls, a balanced line such as a pair line can be used. However, the signal line Ls is not limited to this example, and an unbalanced line such as a coaxial line, a pair of enhanced category 5 or category 6 LAN cables, One pair of CPEV cables may be used.

  As shown in FIG. 1A, the terminal device 2 in the present embodiment includes a microphone 20a that is a voice input unit that converts a received sound wave into an analog voice signal (see FIG. 2A) composed of an electrical signal. , A microphone amplifier 20b that amplifies the output of the microphone 20a, a speaker 21a that is sound output means for transmitting a sound wave (outputs sound) based on the input sound signal, and amplifies the sound signal input to the speaker 21a Speaker amplifier 21b. Note that since the sound signal output from the microphone 20a includes sound other than sound, it is a sound signal (acoustic signal) in a strict sense. Then, it is called an audio signal.

  In addition, the terminal device 2 transmits the analog audio signal obtained from the microphone 20 a to the other terminal device 2 using the audio transmission packet, and the audio signal received from the other terminal device 2. A signal processing means 22 having a reception processing means 4 for performing processing for obtaining an analog voice signal from a transmission packet, and a packet signal (hereinafter referred to as “voice transmission signal”) output from the transmission processing means 3 is amplified. And a receiving amplifier 23b for amplifying an audio transmission signal input to the receiving processing means 4.

  As shown in FIG. 1B, the transmission processing means 3 includes an A / D conversion means 30 for converting the analog voice signal amplified by the microphone amplifier 20b into a digital voice signal having a predetermined sample frequency and quantization width, And an encoding means 31 for generating audio data by encoding the quantized value of each sample point of the digital audio signal in a predetermined format (method).

The A / D conversion means 30 samples, for example, an analog audio signal as shown in FIG. 2A at a predetermined sampling frequency (also referred to as sampling frequency), for example, 44.1 KHz, and FIG. As shown, the intensity of the analog speech signal at the sample point S _n (S _{1 to} S _{10 in the} illustrated example) is acquired, and then the intensity at each sample point S _n is quantized with a predetermined quantization width, pointwise _{S n,} for example, obtain the quantization values _{Q n} of 16 bits 0 through 65535 (where, n is an integer of 1 or more). The encoding unit 31 encodes the quantized value Q _n at the sample point S _n of the digital audio signal obtained from the A / D conversion unit 30, that is, generates audio data by representing it with a bit string. As an encoding method, ADPCM (Adaptive Differential Pulse Code Modulation) encoding is adopted. In addition to ADPCM encoding, encoding may be performed by PCM encoding, APCM encoding, DPCM encoding, or the like. That is, in the audio transmission system of the present embodiment, the A / D conversion means 30 and the encoding means 31 constitute information source encoding means for encoding an analog audio signal as an information source.

  Further, the transmission processing unit 3 performs transmission line coding (Channel Coding) on the voice data so as to create a voice transmission packet in order to transmit the voice data created by the coding unit 31 to another terminal device 2. The path encoding unit 32 and the voice transmission packet generated by the transmission path encoding unit 32 are modulated by, for example, BPSK (Binary Phase Shift Keying), QPSK (Quadrature Phase Shift Keying), 16QAM (Quadrature Amplitude Modulation), 64QAM, etc. And a modulation means 33 for creating a voice transmission signal to be transmitted to the signal line Ls. The transmission path coding means 32 and the modulation means 33 are used for the voice data created by the information source coding means. Transmission means 34 for transmitting to the terminal device 2 (that is, the terminal device acting as a voice receiving terminal device) 2.

Furthermore, the transmission processing unit 3, A / D digital audio signal of the digital audio signal delayed by sample points one minute of time obtained by the conversion means 30, that is the quantized value point at the sampling points S _k is Q _k And a digital signal obtained by the A / D conversion means 30 and a delay signal creating means 35 for obtaining a delay signal composed of a digital signal having a quantized value Q _{k−1 at} the sample point S _k . quantization value Q _k sampling points S _k-1 of the previous quantization value Q _k sampling points S _k of the digital audio signal with a delay signal obtained from the audio signal and the delayed signal generating means 35 The first difference calculation means 36 for calculating the absolute value A _k (= | Δ _k |) of the difference Δ _k (= Q _k −Q _k−1 ) with respect to ₋₁ and the audio data created by the encoding means 31 One voice frame divided by time axis (In the example shown in FIG. 2 (a), a part of the audio data created from the audio signal in a predetermined period T) beam to calculate the maximum value M of the absolute value A _k calculated by the first difference calculation means 36 for And a correction reference value calculation means 37 (where k is an integer of 2 or more). For example, when the sample point S _{2 to} S ₁₀ is included in the one audio frame of the audio data, the correction reference value calculation unit 37 selects the maximum value M from the absolute values A _{2 to} A ₁₀ .

  The transmission path encoding unit 32 creates a voice transmission packet having the audio data obtained from the encoding unit 31 and sends it to the modulation unit 33. For example, the transmission path encoding unit 32 includes a header section and a header section behind the header section. The voice transmission packet having the arranged data portion is created. The header part includes, for example, a preamble, an identification code (ID) indicating the transmission source terminal device 2, a packet type, a modulation method (a modulation method in the modulation means 33, that is, any one of BPSK, QPSK, 16QAM, and 64QAM in this embodiment). And the data length (packet data length) are arranged in the order of preamble, identification code, packet type, modulation scheme, and data length, and for the data sequence arranged after the data sequence. It consists of an error detection code. On the other hand, the data section includes the one voice frame of the voice data obtained from the encoding means 31, the maximum value M calculated by the correction reference value calculation means 37 for the one voice frame, the voice data and the maximum value M. Error detection codes. In the terminal device 2 in the present embodiment, CRC is used as an error detection method, and the error detection code in each of the header part and the data part is an error detection code (CRC code) corresponding to the CRC. Although other error detection methods may be employed, it is desirable to use an error detection code with as low redundancy as possible as long as there is a demand for real-time communication.

  As described above, the transmission unit 34 according to the present embodiment creates a packet having the one voice frame of the voice data and the maximum value M calculated by the correction reference value calculation unit 37 for the one voice frame, so that other terminal apparatuses 2 to send.

  As shown in FIG. 1 (c), the reception processing means 4 demodulates the voice transmission signal amplified by the reception amplifier 23b to obtain a voice transmission packet, and the voice transmission signal obtained by the demodulation means 40. Transmission path decoding means 41 for extracting voice data and the like by decoding the transmission path of the packet. The demodulation means 40 and the transmission path decoding means 41 act as other terminal devices (that is, as voice transmission terminal devices). The receiving means 42 for receiving the voice transmission packet transmitted by the transmitting means of the terminal device 2).

  The reception processing means 4 also includes a decoding means 43 for decoding one voice frame of the voice data extracted from the voice transmission packet by the transmission path decoding means 41 to obtain a digital voice signal, and one of the received voice data. An error detection unit 44 that performs error detection on a voice frame, an abnormal noise countermeasure processing unit 45 that performs an abnormal noise countermeasure process on a digital audio signal obtained from the decoding unit 43, and an abnormal noise countermeasure processing unit 45. D / A conversion means 46 for converting the digital audio signal subjected to the sound countermeasure processing into an analog audio signal.

  The error detection means 44 performs error detection of the header part and the data part using the error detection code of the header part of the voice transmission packet obtained by the demodulation means 40 and the error detection code of the data part, respectively. Is notified to the noise countermeasure processing means 45. In the present embodiment, the error detection code corresponding to the CRC is used as described above, so that the error detection means 44 performs error detection by CRC. Further, when an error is detected in the header part, the voice transmission packet is discarded.

Similar to the delay signal creating means 35, the noise countermeasure processing means 45 delays the digital sound signal obtained from the decoding means 43 by a time corresponding to one sample point, that is, at the sample point _Sk . A delay signal generating means 45a for obtaining a delay signal composed of a digital signal having a quantized value Q _{k-1 at} the sample point S _k for a digital speech signal having a quantized value point Q _k , and a decoding means 43 more resulting before one quantization value Q _k sampling points S _k of the digital audio signal with a delay signal obtained from a digital audio signal and the delayed signal generating means 45a sample points S _k-1 of Second difference calculation means 45b for calculating an absolute value A _k (= | Δ _k |) of a difference Δ _k (= Q _k −Q _k−1 ) from the quantized value Q _k−1, and transmission path decoding means 41 for voice transmission packet And a correction reference value holding means 45c for storing the obtained maximum value M.

Moreover, abnormal noise countermeasure processing unit 45, when an error is detected in the data portion of the audio transport packet by the error detection means 44, an absolute value A _k calculated by the second difference calculating means 45b, the received packet Is compared with the maximum value M, and if the absolute value _Ak exceeds the maximum value M, the comparison means 45d that outputs a correction signal, and the correction signal of the comparison means 45d is input, the digital sound signal is corrected. Correction means 45e for performing the above. That is, the noise countermeasure processing means 45 corrects the digital audio signal when an error is detected in the data portion of the received audio transmission packet (when a transmission error occurs), and the audio transmission packet is corrected. When no error is detected in the data portion, the digital audio signal is not corrected.

Correcting means 45e, when the correction signal of the comparison means 45d is input, the absolute value _{A k} of the difference delta _k of the quantization value _{Q k-1} of the previous sample point _{S k-1} by comparison means 45d The sample point S _k determined to exceed the maximum value M, that is, the quantized value Q _k of the correction-target sample point S _k is used as another sample point S _l (l ≠ in the predetermined range including the sample point S _k. It is corrected to a value obtained by interpolation using the quantized value Q _l of k). The predetermined range is, for example, a range centering on the sample point S _k to be corrected, for example, a range including the sample point S _k−m to the sample point S _{k + m} (where m is an integer).

In the correction unit 45e in the present embodiment, a range where m = 1 is adopted as the predetermined range. Thus, the correction means 45e in the present embodiment, the quantization value Q _k sampling points S _k to be corrected, the other sample points of the sampling points S _k in the predetermined range, i.e. before and after the sampling of the sample points S _k A value interpolated using the quantized values Q _k-1 and Q _{k + 1} of the points S _k−1 and S _{k + 1} , for example, a quantized value Q _k ′ = (Q _k−1 + Q) which is a value obtained by linear interpolation. _{k + 1} ) / 2.

Note that m may be 2 or more, and the value of m is the number of sample points included in one cycle (about 300 μm to 5 msec) when the frequency band of speech is considered to be 200 Hz to 3.4 KHz. It may be a value corresponding to. That is, the time between sample points, for equal to the inverse of the sampling frequency of the A / D converter 30 (i.e. the sampling period), included in the two-fold period of the sampling period around the sampling point S _k The number of sample points is 3, and m is 1 in this case.

The second difference calculation means 45b, when the quantization value _{Q k} sampling points _{S k} by the correction of the correcting means 45e is corrected to _{Q k} 'is an absolute value _{A k + 1} at the next sampling point _{S k + 1} In the calculation, the quantized value Q _k ′ corrected by the correcting unit 45e is employed instead of the quantized value Q _k of the delayed signal obtained from the delayed signal generating unit 45a. absolute value _{a k + 1} for the next sample point _{S k + 1 is, | Q k + 1 -Q k} '| , and the result is output to the comparison result.

In such abnormal noise countermeasure process section 45, for example, by the decoding means 43 from the audio frame of the audio transport packets received in the digital audio signal (FIG. 3 consisting of sampled points S ₂₀ to S ₃₀ shown in FIG. 3 If you are also illustrated that interpolates the analog audio signal) is obtained is compared with the originally calculated absolute value a ₂₀ for sampling points S ₂₀ is the maximum value M, then sequentially, the absolute value A ₂₁ , A ₂₂ ,... Are calculated and compared with the maximum value M. Here, if all of the absolute values A _{20 to} A ₂₄ are equal to or less than M, the correction means 45e does not correct the quantized values Q _{20 to} Q _{24 at the} respective sample points S _{20 to} S ₂₄ . It should be noted that at the beginning of when calculating the absolute value A ₂₀ for sampling points S ₂₀ in the audio frame, quantization values Q ₁₉ sampling points S ₁₉ but is required, such quantized value Q ₁₉ is the previous From the voice frame (voice frame of the voice transmission packet received last time).

On the other hand, has become rapidly the quantization value Q ₂₅ sampling points S ₂₅ increases by a transmission error, if it exceeds the absolute value A ₂₅ is a M, the sample points S ₂₅ is a correction target, the quantized value Q _25, as shown in FIG. 3, sample points _S k-1 before and after the sample points _{S k} to be _{corrected, S} quantized values of _{k + 1 Q k-1,} Q k is the average value of _{+ 1} quantized values _{Q k} It is corrected to '= (Q _k-1 + Q _{k + 1} ) / 2.

In this case, the absolute value _{A 26} for the next sampling point _{S 26 is,} | Q 26 _{-Q 25} '| becomes, when where the absolute value _{A 26} <M, the quantization value _{Q 26} at the sampling points _{S 26} No correction is made for it. Absolute value _{A 27} for the next sampling point _{S 27,} to the quantized value _{Q 26} at the immediately preceding sample point _{S 26} has not been _corrected, | Q 27 _{-Q 26} | next is the absolute value _{A 27} <M When, the correction is not performed on the quantized value _{Q 27} at the sampling points _{S 27,} hereinafter, similarly, the absolute value _a 28 _{to a 30} for sampling points _S 28 _{to S 30} each is calculated, either absolute If the values A _{28 to} A ₃₀ are also equal to or smaller than M, the quantizing values Q _{28 to} Q _{30 at the} respective sample points S _{28 to} S ₃₀ are not corrected by the correcting unit 45e.

Therefore, from the correction means 45e of noise countermeasure process section 45, shown partly in broken lines in the digital audio signal (FIG. 3 as the quantization value at the sampling points _S 21 _{to S 30} each is _Q 21 _{to Q 30} Digital speech signals whose quantized values at sample points S _{21 to} S ₃₀ are Q _{21 to} Q ₂₄ , Q ₂₅ ′, Q _{26 to} Q ₃₀ (shown by solid lines in FIG. 3) Is output. The digital audio signal outputted from the abnormal noise investigation process unit 45, the quantized values at the sampling points S ₂₅ is corrected to Q _{25 'from} Q _25, thereby, a quantum at the sampling points S ₂₅ that may be causing the abnormal noise Sudden changes in chemical value are suppressed.

The abnormal noise countermeasure processing means 45 described above is a digital audio signal after decoding including a sampling point S _k that is considered to cause a sudden change in the quantized value Q _k (that is, a terminal device on the receiving side). the digital audio signal) created in the second decoding means 43, the change of the quantization value Q _k (the difference between the quantized value) the original digital audio signal (i.e. a / D conversion terminal 2 as the transmission side The digital voice signal is corrected so as not to exceed the maximum value M of the absolute value A _k of the difference Δ _k of the quantized value Q _k in the digital voice signal generated by the means 30. In other words, abnormal noise countermeasure processing means 45, the difference delta _k quantized values Q _k at each sample point S _k of the digital audio signal after decoded, the original digital audio signal noise due transmission error has not occurred The digital audio signal after decoding is corrected so as not to exceed the maximum value M of the absolute value A _k of the difference Δ _k of the quantized value Q _k that is allowed (that is, it is considered that no abnormal noise is generated). in suppressing an abrupt change in the quantization value Q _k which cause the generation of noise, thereby suppressing the occurrence of abnormal noise due to transmission errors. In this case, since the quantized value Q _k is applied only to the sampling point S _k causing the abnormal sound, the sound is not output by greatly reducing the amplitude of the sound signal for a certain period or setting it to 0 (mute). Compared with muting processing, the waveform of the original digital audio signal remains, so that the reproducibility of audio can be improved.

Moreover, the correction means 45e is the absolute value A _k of the difference delta _k is a quantized value Q _k at the sampling points S _k of the correction target that exceeds the maximum value M, the sample points in a predetermined range including the corresponding sample point S _k is corrected to the average value of the quantization values, the quantization value Q _{k 'at} the sample points S _k corrected, quantized values at the sampling points S _k of the original digital audio signal Q _{k (correct} quantization value Q _k ), The reproducibility of voice can be further improved.

Here, the correction unit 45e in this embodiment, has a range of the predetermined range and m = 1, the quantization value Q _k sampling points S _k to be corrected, before and after the sample points S _k to be corrected Although it is corrected to a quantized value Q _k ′ = (Q _k−1 + Q _{k + 1} ) / 2 which is an average value of the quantized values Q _k−1 and Q _{k + 1} of the sampling points S _k−1 and S _{k + 1} , Depending on the magnitude of the quantization value Q _{k + 1} , the absolute value A _{k of the} difference Δ _k = Q _k ′ −Q _k−1 may exceed the maximum value M even at the corrected sample point S _k . In such a case, the value of k may be increased, for example, increased by 1, and correction may be performed again. For example, when A _k = | Q _k ′ −Q _k−1 |> M, the quantized value Q _k ′ of the corrected sample point S _k is included in the m = 2 range. Quantized value Q _k ″ = (Q _k− which is an average value of quantized values Q _k−2 , Q _k−1 , Q _{k + 1} , and Q _{k + 2} of S _k−2 , S _k−1 , S _{k + 1} , and S _{k + 2 2} + Q _k−1 + Q _{k + 1} + Q _{k + 2} ) / 4 If this absolute value A _k exceeds the maximum value M, correction may be performed by further increasing the value of m.

Further, the correction means 45e in the present embodiment, a method for the quantization value Q _k sampling points S _k to be corrected, using the quantization values of other sample points in a predetermined range including the sample point S _k interpolation As an example, linear interpolation is illustrated, but the present invention is not limited to this. For example, interpolation may be performed using an n-order polynomial such as a second-order polynomial, a third-order polynomial, or an eighth-order polynomial as a predetermined function. Etc. can be used.

  By the way, the terminal device 2 transmits / receives a control signal between the main device 1 and the control unit 24 that controls the signal processing unit 22 and the like using the CPU as a main component in addition to the above-described configuration via the signal line Ls. A function (multiplexing function) and a signal line for frequency-division multiplexing the control signal transmission / reception means 25, the audio transmission signal output from the transmission amplifier 23a and the control signal output from the control signal transmission / reception means 25, and sending them to the signal line Ls A demultiplexing / multiplexing unit 26 having a function (separation function) for separating the audio transmission signal and the control signal from the signal taken from Ls and outputting the audio transmission signal to the receiving amplifier 23b and the control signal to the control signal transmitting / receiving unit 25, respectively. I have.

  The terminal device 2 includes a calling means (not shown) including a call button for calling another terminal device 2 and a response means (not illustrated) including a response button for responding to a call by the other terminal device 2. And a general call button (not shown) for calling all the terminal devices 2 connected to the signal line Ls. The control means 24 and the control signal transmission / reception means 25 are configured so as to be able to perform the operations described later (the operations shown in FIGS. 5 and 6). Since it is well known, detailed description is omitted.

  The main device 1 includes a control unit 10 having a CPU as a main component, a control signal transmission / reception unit 11 that transmits / receives a control signal to / from the terminal device 2, and a packet that includes synchronization data (described later) supplied from the control unit 10 ( Hereinafter, a modulation unit 12 that generates a synchronization signal SY by modulating (for example, modulation by BPSK, QPSK, 16QAM, 64QAM, etc.) and a synchronization signal SY generated by the modulation unit 12 A function of frequency-division-multiplexing the control signal output from the control signal transmission / reception means 11 and sending it to the signal line Ls, and a separation for separating the control signal from the signal taken in from the signal line Ls and outputting it to the control signal transmission / reception means 11 Multiplexing means 13 is provided. The control means 10 and the control signal transmission / reception means 11 are configured so as to be able to perform the operations described later (the operations shown in FIGS. 5 and 6). Since it is well known, detailed description is omitted.

  Incidentally, the synchronization packet has, for example, a header part and a data part arranged behind the header part. The header part includes, for example, a preamble, a packet type, a modulation method (a modulation method in the modulation means 12, that is, any one of BPSK, QPSK, 16QAM, and 64QAM in this embodiment), and a data length (data length of a synchronization packet). Is composed of a data sequence arranged in the order of preamble, packet type, modulation scheme, and data length, and an error detection code for the data sequence arranged after the data sequence. The data part is composed of the above-described actually empty synchronization data and the error detection code for the synchronization data.

  In the main device 1 and the terminal device 2, as shown in FIG. 4, the time point when a predetermined time has elapsed from the falling edge of the synchronization signal SY is set as the start time point of the first time slot TS <b> 1 and the next synchronization signal SY rises. A plurality (n) of time slots TS1, TS2,..., TSn are arranged within a period (signal transmission period). That is, voice signals are transmitted and received between the terminal devices 2 by time division multiplexing access (TDMA). In the audio transmission system of this embodiment, n time slots TSi are arranged in the signal transmission period, so that a maximum of n / 2 terminal devices 2 can transmit and receive audio signals on a one-to-one basis. It has become.

  Next, an operation of the transmission system according to the present embodiment, for example, an operation when a one-to-one call (individual call) is performed between two terminal apparatuses 2 will be described with reference to FIG. In the following description, the terminal device 2 is represented by a terminal device A and a terminal device B in order to distinguish the two terminal devices 2 as necessary.

  First, when a call button for calling the terminal device B is operated in the terminal device A, the control means 24 of the terminal device A calls the call (call establishment) request data for calling the partner terminal device B, Control data (call request control data) including the other identification code (identification code of terminal device A) and the other party's identification code (identification code of terminal device B) are generated and output to control signal transmitting / receiving means 25 The control signal transmitting / receiving means 25 modulates the call request control data obtained from the control means 24 (for example, modulates by BPSK, QPSK, 16QAM, 64QAM, etc.) to generate a control signal (call request control signal), The signal is sent to the signal line Ls through the demultiplexing means 26 (arrow F1 in FIG. 5).

  When the control signal transmission / reception means 11 of the main apparatus 1 receives the call request control signal via the signal line Ls, the call request data included in the call request control signal and the identification code of the call source (request source). The identification code of the terminal device A and the identification code of the terminal device B, which is the identification code of the call destination (request destination), are sent to the control means 10, and the control means 10 uses the call identification code as the destination. Control data including the data (call request notification control data) is generated and output to the control signal transmission / reception means 11. The control signal transmission / reception means 11 modulates the call request notification control data obtained from the control means 10 to generate a control signal (call request notification control signal) and sends it to the signal line Ls via the demultiplexing means 13. (Arrow F2 in FIG. 5).

  When receiving the call request notification control signal through the signal line Ls, the control signal transmission / reception means 25 of the terminal device B sends the call request data included in the call request notification control signal to the control means 24. The control means 24 Then, a notification sound (ringing sound) is generated from the speaker 21a.

  When the response button is operated in the terminal device B after the notification sound (ringing sound) is sounded in this way and before a predetermined time has elapsed, the control means 24 of the terminal device B causes the calling terminal device A to Control data (acknowledgment control data) including acknowledgment data indicating that the individual call has been accepted and the identification code (identification code of the terminal device B) is generated and output to the control signal transmission / reception means 25 The control signal transmission / reception means 25 modulates the acknowledgment control data obtained from the control means 24 to generate a control signal (acknowledgment control signal) and sends it to the signal line Ls via the demultiplexing means 26 ( Arrow F3 in FIG. 5).

  When the control signal transmission / reception means 11 of the main apparatus 1 receives the acknowledgment control signal via the signal line Ls, the acknowledgment data included in the acknowledgment control signal and the identification code of the call destination (request destination) The terminal device B identification code is sent to the control means 10, and the control means 10 uses the caller identification code (terminal device A identification code) as the destination and indicates call establishment notification indicating that acknowledgment data has been obtained from the call destination. Control data including the data (call establishment notification control data) is generated and output to the control signal transmission / reception means 11. The control signal transmission / reception means 11 modulates the call establishment notification control data obtained from the control means 10 to generate a control signal (call establishment notification control signal) and sends it to the signal line Ls via the demultiplexing means 13. (Arrow F4 in FIG. 5).

  In addition, after creating the control data (call establishment notification control data), the control means 10 assigns the time slot TSi to each of the calling terminal device A and the called terminal device B, and notifies the result. Control data (control data for time slot allocation) is generated and output to the control signal transmission / reception means 11. The control signal transmission / reception means 11 modulates the time slot allocation control data obtained from the control means 10 to generate a control signal (time slot allocation control signal) and sends it to the signal line Ls via the demultiplexing means 13. (Arrow F5 in FIG. 5).

  When the control signal transmitting / receiving unit 25 of the terminal device 2 receives the time slot allocation control signal through the signal line Ls, the control signal transmitting / receiving unit 25 sends the time slot allocation control data to the control unit 24. Based on the control data, the demultiplexing means 26 is controlled. The control unit 24 activates the microphone 20a, the microphone amplifier 20b, the speaker 21a, the speaker amplifier 21b, the signal processing unit 22, the transmission amplifier 23a, the reception amplifier 23b, and the like.

  Thereafter, for example, when voice is input to the microphone 20a of the calling terminal device A, the microphone 20a creates an analog voice signal based on the voice and outputs it to the microphone amplifier 20b, and the microphone amplifier 20b receives the analog voice. The signal is amplified and output to the transmission processing means 3 of the signal processing means 22. The transmission processing means 3 creates an audio transmission signal as described above and outputs it to the transmission amplifier 23a. The transmission amplifier 23 a amplifies the audio transmission signal output from the transmission processing unit 3 and outputs it to the demultiplexing unit 26.

  When the demultiplexing means 26 receives the synchronization signal SY (arrow F6 in FIG. 5), the demultiplexing means 26 sends an audio transmission signal to the signal line Ls in accordance with the time slot TSi assigned in advance.

  On the other hand, when receiving the voice transmission signal whose transmission source is the terminal device A, the demultiplexing means 26 of the called terminal device B outputs the voice transmission signal to the reception amplifier 23b, and the reception amplifier 23b receives the voice transmission signal. Amplified and output to the reception processing means 4 of the signal processing means 22, and the reception processing means 4 outputs an analog audio signal as described above, and the analog audio signal output by the reception processing means 4 is amplified by the speaker amplifier 21b. Is then input to the speaker 21a, and the speaker 21a outputs sound based on the input analog audio signal.

  The above-described operation is the same even when a voice signal is transmitted from the called terminal device B to the calling terminal device A, and the two terminal devices A and B use different time slots TSi. Then, voice signals are transmitted and received, whereby individual calls between the terminal devices A and B are performed.

  Next, an operation of the transmission system according to the present embodiment, for example, an operation when performing a many-to-many call (simultaneous call) between the three terminal devices 2 will be described with reference to FIG. In the following description, the terminal device 2 is represented by a terminal device A, a terminal device B, and a terminal device C in order to distinguish the three terminal devices 2 as necessary.

  First, in the terminal device A, when the simultaneous call button for calling all the terminal devices 2 connected to the signal line Ls is operated, the control means 24 of the terminal device A transmits the other terminal devices B and C all at once. Control data (simultaneous call request control data) including the general call request data for calling and the identification code of the terminal device A (identification code of the terminal device A) is generated and output to the control signal transmission / reception means 25, and the control signal The transmission / reception means 25 modulates the general call request control data obtained from the control means 24 to generate a control signal (general call request control signal) and sends it to the signal line Ls via the demultiplexing means 26 (FIG. 6 middle arrow G1).

  When the control signal transmission / reception means 11 of the main device 1 receives the general call request control signal through the signal line Ls, the general call request data included in the general call request control signal and identification of the call source (request source) The terminal device A identification code, which is a code, is sent to the control means 10, and the control means 10 has control data (simultaneous call request notification control data) including the above-mentioned general call request data, each addressed to the call destination identification code, That is, the control data for general call request notification destined for the terminal device B and the control data for general call request notification destined for the terminal device C are generated and output to the control signal transmitting / receiving means 11. The control signal transmission / reception means 11 modulates each of the general call request notification control data obtained from the control means 10 to generate a control signal (general call request notification control signal), and the signal line Ls via the demultiplexing means 13. (Arrows G2 and G3 in FIG. 6).

  When the control signal transmission / reception means 25 of each of the terminal devices B and C that are the general call destinations receives the general call request notification control signal through the signal line Ls, the general call request data included in the general call request notification control signal is received. The information is sent to the control means 24, and the control means 24 sounds a notification sound (calling sound) from the speaker 21a. When the response button is operated in the terminal devices B and C after the notification sound (ringing tone) is sounded in this way and before the predetermined time has elapsed, the control means 24 of the terminal devices B and C Create and control control data (acknowledgment control data) including acknowledgment data indicating that the individual terminal device A has been individually communicated and its own identification code (identification codes of the terminal devices B and C) The control signal transmitting / receiving unit 25 generates a control signal (acknowledgment control signal) by modulating the acknowledgment control data obtained from the control unit 24, and outputs the control signal via the demultiplexing / multiplexing unit 26. The signal is sent to the signal line Ls (arrows G4 and G5 in FIG. 6).

  When the control signal transmission / reception means 11 of the main device 1 receives the acknowledgment control signal from each of the terminal devices B and C through the signal line Ls, the acknowledgment data included in the acknowledgment control signal and the general call destination ( The identification codes of the terminal devices B and C, which are the identification codes of the request destination), are sent to the control means 10, and the control means 10 affirms the identification code of the simultaneous call source (identification code of the terminal apparatus A) as the destination from the call destination Control data (call establishment notification control data) including call establishment notification data indicating that the response data has been obtained is generated and output to the control signal transmission / reception means 11. The control signal transmission / reception means 11 modulates the call establishment notification control data obtained from the control means 10 to generate a control signal (call establishment notification control signal) and sends it to the signal line Ls via the demultiplexing means 13. (Arrow G6 in FIG. 6).

  Further, after creating the control data (call establishment notification control data), the control means 10 assigns a time slot TSi to each of the terminal devices A, B, and C, and provides control data for notifying the result ( Time slot allocation control data) is generated and output to the control signal transmission / reception means 11. The control signal transmission / reception means 11 modulates the time slot allocation control data obtained from the control means 10 to generate a control signal (time slot allocation control signal) and sends it to the signal line Ls via the demultiplexing means 13. (Arrow G7 in FIG. 6).

  When the control signal transmitting / receiving unit 25 of the terminal device 2 receives the time slot allocation control signal through the signal line Ls, the control signal transmitting / receiving unit 25 sends the time slot allocation control data to the control unit 24. Based on the control data, the demultiplexing means 26 is controlled. The control unit 24 activates the microphone 20a, the microphone amplifier 20b, the speaker 21a, the speaker amplifier 21b, the signal processing unit 22, the transmission amplifier 23a, the reception amplifier 23b, and the like.

  Thereafter, for example, when voice is input to the microphone 20a of the calling terminal device A, the microphone 20a creates an analog voice signal based on the voice and outputs it to the microphone amplifier 20b, and the microphone amplifier 20b receives the analog voice. The signal is amplified and output to the transmission processing means 3 of the signal processing means 22. The transmission processing means 3 creates an audio transmission signal as described above and outputs it to the transmission amplifier 23a. The transmission amplifier 23 a amplifies the audio transmission signal output from the transmission processing unit 3 and outputs it to the demultiplexing unit 26.

  When the demultiplexing means 26 receives the synchronization signal SY (arrow G8 in FIG. 6), the demultiplexing means 26 sends an audio transmission signal to the signal line Ls in accordance with a pre-assigned time slot TSi.

  On the other hand, when the demultiplexing means 26 of each of the terminal devices B and C, which are the general call destinations, receives a voice transmission signal whose source is the terminal device A, the voice transmission signal is received by the reception processing means via the reception amplifier 23b. 4, the reception processing unit 4 outputs an analog audio signal as described above. The analog audio signal output by the reception processing unit 4 is amplified by the speaker amplifier 21 b and then input to the speaker 21 a, and the speaker 21 a Audio is output based on the input analog audio signal.

  The operation described above is the same when the terminal device B transmits an audio signal or when the terminal device C transmits an audio signal, and the three terminal devices A, B, and C are in different time slots. A voice signal is transmitted and received using TSi, whereby a simultaneous call between terminal apparatuses A, B, and C is performed.

  By the way, in the above description, when each of the terminal devices B and C receives the general call request notification control signal, the terminal device B and C sound a notification sound, and then the response button is operated, so that the acknowledgment control signal is transmitted. A transmission is made to be transmitted to the signal line Ls, and a call between the terminal devices A, B, and C is possible by operating the response button, but the operation of such a response button is omitted. It is good also as the structure which carried out. In the above example, each of the terminal devices B and C receives the simultaneous call request notification control signal so that the signal processing means can input and output the sound for a certain time without sounding the notification sound. 22 may be activated. The certain time is extended when the audio transmission signal is received.

  In this case, the acknowledgment indicated by arrows G4 and G5 in FIG. 6 and the call establishment notification indicated by arrow G6 are omitted, and after the general call request notification indicated by arrows G2 and G3 is performed from the main apparatus 1, the arrow G7 The time slot allocation notification shown in FIG. Note that the configuration in which the operation of the response button is omitted as described above can also be adopted in the example shown in FIG. 5, and in this case, the affirmative response indicated by the arrow F3 and the call establishment notification indicated by the arrow F4 are omitted.

According to the audio transmission system of the present embodiment described above, the difference Δ _k indicates the quantized value Q _k at each sample point S _k of the decoded digital audio signal, and no abnormal noise due to a transmission error has occurred. By correcting so as not to exceed the maximum value M of the absolute value A _k of the difference Δ _k of the quantized value Q _k allowed in the digital audio signal, it is possible to cause abnormal noise in the decoded digital audio signal. since become such inhibition an abrupt change of the quantization value Q _k, it is possible to suppress the generation of abnormal noise due to transmission errors and, moreover, the quantization value Q _k only the sample points S _k is causing the abnormal noise Therefore, the waveform of the original digital audio signal remains compared to the muting process that does not output (mute) audio by significantly reducing the amplitude of the audio signal for a certain period or reducing it to 0. Improve voice reproducibility.

Moreover, the correction means 45e is the absolute value A _k of the difference delta _k is a quantized value Q _k at the sampling points S _k of the correction target that exceeds the maximum value M, the sample points in a predetermined range including the corresponding sample point S _k is corrected to the average value of the quantization values, the quantization value Q _{k 'at} the sample points S _k corrected, quantized values at the sampling points S _k of the original digital audio signal Q _{k (correct} quantization value Q _k ), The reproducibility of voice can be further improved.

  By the way, in the terminal device 2 in the present embodiment, since the signal processing unit 22 includes the transmission processing unit 3 and the reception processing unit 4, an audio signal to be output based on the audio input from the microphone 20 a is output as audio. A function as a voice transmission terminal device that converts data into a voice transmission packet and a voice reception terminal that outputs voice to the speaker 21a based on a voice signal obtained from voice data included in the received voice transmission packet It has both functions as a device. The configuration of the terminal device 2 is not limited to the above configuration, and may have only a function as a voice transmission terminal device, or only a function as a voice reception terminal device. In short, in addition to the terminal device 2 having a function as a voice transmitting terminal device among the terminal devices 2 connected by the signal line Ls, the voice receiving terminal device It is only necessary that the terminal device 2 having a function exists.

(A) is a block diagram of an audio transmission system according to an embodiment of the present invention, (b) is a block diagram of transmission processing means, and (c) is a block diagram of reception processing means. (A) is explanatory drawing of an analog audio | voice signal, (b) is explanatory drawing of a digital audio | voice signal. It is explanatory drawing of the correction method in the same as the above. It is a time chart of a synchronous signal and time slot in the same as the above. It is operation | movement explanatory drawing in the same as the above. It is operation | movement explanatory drawing in the same as the above.

Explanation of symbols

2 Terminal device (voice transmitting terminal device, voice receiving terminal device)
20a Microphone (voice input means)
30 A / D conversion means 31 Encoding means 34 Transmission means 36 First difference calculation means 37 Correction reference value calculation means 42 Reception means 43 Decoding means 45 Abnormal noise countermeasure processing means 45b Second difference calculation means 45d Comparison means 45e Correction means 46 D / A conversion means

Claims (1)

A plurality of voice transmitting terminal apparatuses that transmit voice signals by packets and voice receiving terminal apparatuses that output voice based on the received voice signals, and the plurality of voice transmitting terminal apparatuses and voice receiving terminal apparatuses; Are connected via the same signal line, and the voice transmitting terminal device and any voice receiving terminal different from the voice transmitting terminal device among the plurality of voice transmitting terminal devices and voice receiving terminal devices. An audio transmission system comprising a device capable of audio transmission,
The voice transmitting terminal device converts an analog voice signal obtained by voice input means for converting a received sound wave into an analog voice signal composed of an electrical signal into a digital voice signal having a predetermined sample frequency and quantization width. Conversion means; encoding means for encoding the quantized value of each sample point of the digital speech signal in a predetermined format to create speech data; and the quantized value of the sample point of the digital speech signal and the previous one First difference calculation means for calculating an absolute value of a difference from the quantized value of the sample point, and a first difference calculation means for one voice frame obtained by dividing the voice data created by the coding means on the time axis A packet having a correction reference value calculation means for calculating the maximum absolute value, the one voice frame of the voice data, and the maximum value calculated by the correction reference value calculation means for the one voice frame And transmitting means for transmitting the audio receiving terminal to create,
An audio receiving terminal device includes: a receiving unit that receives a packet transmitted by a transmitting unit; a decoding unit that decodes one audio frame of audio data included in the packet received by the receiving unit to obtain a digital audio signal; and a decoding unit Noise countermeasure processing means for performing noise countermeasure processing on the obtained digital voice signal, and D / A conversion for converting the digital voice signal subjected to noise countermeasure processing by the noise countermeasure processing means into an analog voice signal Means and
The abnormal noise countermeasure processing means includes a second difference calculating means for calculating an absolute value of a difference between a quantized value of the sample point of the digital audio signal and a quantized value of the immediately preceding sample point, and a second difference calculating means. Comparing means for comparing the absolute value calculated by the above and the maximum value of the received packet, and determining that the absolute value of the difference between the quantized value of the previous sample point exceeds the maximum value by the comparing means The quantized value Q _k (k is an integer greater than or equal to 2) of the sample points thus _{obtained is used} as the quantized value Q _k-1 and the quantized value Q of the sample point immediately before the quantized value Q _k in the time series. The absolute value of the difference between the average value Q _k ′ and the quantized value Q _k−1 of the immediately preceding sample point is corrected to the average value Q _k ′ using the quantized value Q _{k + 1} of the sample point immediately after _k. the value _{a k = | Q k '-Q} k-1 | if exceeds the maximum value, the sample point of the immediately preceding Coca value Q further quantized value of the previous _k-1 Q _k-2 and the quantized value Q further quantized value immediately after the sample points of the quantization value Q _{k + 1} immediately following the _k Q _{k + 2} and the new using The average value Q _k ″ = (Q _k−2 + Q _k−1 + Q _{k + 1} + Q _{k + 2} ) / 4, the quantized value Q _k is corrected, and the new average value Q _k ″ and the previous sample point When the absolute value of the difference from the quantized value Q _k−1 = | Q _k ″ −Q _k−1 | exceeds the maximum value, the further previous quantized value Q _k−2 is further increased. The quantized value is further added to a new average value using a total of six values obtained by adding the immediately preceding quantized value Q _k−3 and the immediately following quantized value Q _{k + 2 to} the immediately following quantized value Q _{k + 3} . voice transmission system, characterized by comprising a correction means for performing abnormal noise correction processing that corrects the value Q _k .

Images