CN1368820A - Voice communication terminal and voice communication system - Google Patents

Voice communication terminal and voice communication system Download PDF

Info

Publication number
CN1368820A
CN1368820A CN01132642.5A CN01132642A CN1368820A CN 1368820 A CN1368820 A CN 1368820A CN 01132642 A CN01132642 A CN 01132642A CN 1368820 A CN1368820 A CN 1368820A
Authority
CN
China
Prior art keywords
voice
code
output
phonetic code
voice communication
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN01132642.5A
Other languages
Chinese (zh)
Other versions
CN1183734C (en
Inventor
田崎裕久
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Mitsubishi Electric Corp
Original Assignee
Mitsubishi Electric Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mitsubishi Electric Corp filed Critical Mitsubishi Electric Corp
Publication of CN1368820A publication Critical patent/CN1368820A/en
Application granted granted Critical
Publication of CN1183734C publication Critical patent/CN1183734C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Landscapes

  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Telephonic Communication Services (AREA)
  • Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)

Abstract

To solve the problem of the conventional voice communication terminal that the telephone call is made difficult by the echo of the self-terminal speech delayed greatly when the speech of the self-terminal is superimposed on the speech of other terminal, and the signal input, speech processing and outputting of the resultant speech are carried out. The voice communication terminal receives a speech and speech codes from separate terminals, respectively, encodes the inputted speech to obtain output speech codes, and decodes the speech codes to output the resultant speech from separate terminals. The voice communication terminal is provided with a correction means which corrects either inputted speech codes, decoded speech or the operation of a speech decoding means and reduces the amplitude of the portion ascribed to the inputted speech contained in the output speech on the basis of the similarity of the speech codes obtained by encoding the inputted speech and inputted speech codes.

Description

Voice communication terminal and voice communication system
Technical field
The present invention relates to a kind of voice communication terminal, it will import voice and code thereof as input, and will to described input voice encode and the output phonetic code that obtains and to described input phonetic code decode and the output voice that obtain as output.And, also relate to a kind of voice communication system of this voice communication terminal of application.Specially refer to, send on each voice communication terminal by 1 low speed line at the input voice that a plurality of voice communication terminal is imported, make each voice communication terminal can both hear 1 pair 1 the dialog context or the voice communication system of conference call, and the voice communication terminal that is fit to this voice communication system.
Prior art
In existing voice communication terminal and voice communication system, 2 voice compositions of stack transmit this overlapped signal by 1 side's circuit on analog voice signal.Such as, when between caller A and caller B, conversing, on the voice communication terminal A place of caller A side is added to the voice signal of the caller B that is received the voice signal of caller A, again the voice signal after the stack is transferred on the voice communication terminal B and other-end of caller B side.
Fig. 8 represents the general structure of existing voice communication system.
In the drawings, 1 for the used voice communication terminal A of caller A, 2 for the used voice communication terminal B, 3 of caller B used voice communication terminal C during for situations such as other relevant personnel monitor.
Below, voice communication terminal A, voice communication terminal B and voice communication terminal C abbreviate terminal A, terminal B and terminal C respectively as.
In terminal A, be provided with and the modulation signal B that comes self terminal B is carried out demodulation and the input voice A that is sent as the analog demodulator device 4 of output voice A, to caller A and described output voice A carry out overlappingly, and generate the voice overlapping device 5 of overlapping voice and the analog-modulated device 6 that overlapping voice are modulated.
In terminal B, be provided with analog-modulated device 7 that input voice B that caller B is sent modulates and the analog demodulator device 8 that the modulation signal A that comes self terminal A is carried out demodulation.
And, in terminal C, be provided with the analog demodulator device 9 that the modulation signal A that comes self terminal A is carried out demodulation.
In addition, 10 is the circuit A that the modulation signal A that comes self terminal A is transmitted, and 11 is the circuit B that the modulation signal B that comes self terminal B is transmitted.
Below in conjunction with figure, the operation of existing voice communication system is described.
In terminal A, the voice that caller A sends are transfused to as input voice A.And, be transfused to from the modulation signal B of the terminal B that hereinafter will introduce by circuit B.4 couples of described modulation signal B of analog demodulator device carry out demodulation, and resulting voice signal is exported as output voice A.In voice overlapping device 5, described input voice A and output voice A are carried out overlapping, thereby generate overlapping voice, and be entered in the analog-modulated device 6.6 pairs of overlapping voice of being imported of analog-modulated device are modulated, and resulting modulation signal is outputed among the circuit A as modulation signal A.
In terminal B, the voice that caller B is sent are transfused to as input voice B.And, by circuit A, come the modulation signal A of self terminal A to be transfused to.7 couples of described input voice B of analog demodulator device carry out demodulation, and resulting modulation signal is outputed among the circuit B as modulation signal B.8 couples of described modulation signal A of analog demodulator device carry out demodulation, and resulting voice signal is exported as output voice B.
In terminal C,, come the modulation signal A of self terminal A to be transfused to by circuit A.9 couples of described modulation signal A of analog demodulator device carry out demodulation, and resulting voice signal is exported as output voice C.
In above structure, when caller A and caller B utilize terminal A and terminal B to converse respectively, might overhear dialog context between caller A and the caller B at terminal C.
Fig. 9 is in described existing voice communication system, in order to improve the line efficiency of data communication importing etc., by improving the quality in the more weak zone of electric wave in radio communication with the combination of error-correcting code, guarantee confidentiality with password combination etc., seek the schematic diagram of other system digitalized structures.
In the drawings, 1 terminal A for caller A use, 2 terminal B for caller B use, 3 is the terminal C that relevant other staff adopt when monitoring.
Comprise among the terminal A carry out the demodulating equipment 12 of demodulation by the modulation signal B that comes self terminal B of circuit B11, the phonetic code B that obtains from demodulating equipment 12 is carried out decoded speech decoding device 13, the output of audio decoding apparatus 13 is carried out digital-to-analogue conversion (D-A conversion) and exported the digiverter 14 of the output voice A of analog signal, the voice overlapping device 15 of the output voice A of input voice A that overlapping caller sends and described digiverter 14, overlapping voice are carried out the analog-digital commutator 16 of analog-to-digital conversion (A-D conversion), the output of analog-digital commutator 16 is carried out the speech coding processing unit 17 of encoding process, the output of speech coding processing unit 17 is carried out Digital Modulation and to the modulating device 18 of circuit A10 output.
In terminal B, comprise to the modulation signal A from circuit A10 carry out demodulation demodulating equipment 20, to the output of demodulating equipment 20 carry out decoding processing audio decoding apparatus 21, the output of audio decoding apparatus 21 is carried out digital-to-analogue conversion (D-A conversion) and is exported the digiverter 22 of the output voice B of analog signal.
In terminal B, also comprise the input voice B that caller B is sent carry out the analog-digital commutator 23 of analog-to-digital conversion (A-D conversion), to the output of analog-digital commutator 23 carry out encoding process phonetic code processing unit 24, voice digital makeup is put that digital modulation is carried out in 24 output and to the modulating device 25 of circuit B11 output.
In addition, in terminal C, comprise to the modulation signal A from circuit A10 carry out demodulation demodulating equipment 26, to the output of demodulating equipment 26 carry out decoding processing audio decoding apparatus 27, the output of audio decoding apparatus 27 is carried out digital-to-analogue conversion (D-A conversion) and is exported the digiverter 28 of the output voice B of analog signal.
Below, the operation of existing another kind of voice communication system is explained.
In terminal A, the voice that caller A is sent are transfused to as the input voice, and, by circuit B, be transfused to from the modulation signal B of terminal B described later.12 couples of described modulation signal B of demodulating equipment carry out demodulation, and resulting phonetic code B is outputed in the audio decoding apparatus 13.13 couples of described phonetic code B of audio decoding apparatus carry out decoding processing, and resulting numeral output voice A is outputed in the digiverter 14.14 couples of described numeral output voice A of digiverter carry out digital-to-analogue conversion (D-A conversion), and the output voice A of output analog signal.
In terminal A, carry out overlapping processing by 15 couples of described input voice A of voice overlapping device and output voice A, thereby generate overlapping voice, and be entered in the analog-digital commutator 16.16 pairs of overlapping voice of being imported of analog-digital commutator carry out analog-to-digital conversion (A-D conversion), thereby obtain the numeral input voice A of digital signal.17 couples of described numeral input voice A of sound encoding device carry out encoding process, and resulting phonetic code A is outputed in the modulating device 18.18 couples of phonetic code A that imported of modulating device carry out digital modulation to be handled, and resulting modulation signal is outputed among the circuit A as modulation signal A.
In terminal B, the voice that caller B is sent are transfused to as input voice B.And, by circuit A, come the modulation signal A of self terminal A to be transfused to.23 couples of described input voice B of analog-digital commutator carry out analog-to-digital conversion (A-D conversion), thereby obtain the numeral input voice B of digital signal.24 couples of described numeral input voice B of sound encoding device carry out encoding process, and resulting phonetic code B is outputed in the modulating device 25.The phonetic code A of 25 pairs of inputs of modulating device carries out Digital Modulation, and resulting modulation signal is outputed on the circuit B as modulation signal B.
20 couples of described modulation signal A of demodulating equipment in the terminal B carry out demodulation, and resulting phonetic code A is outputed in the audio decoding apparatus 21.21 couples of described phonetic code A of audio decoding apparatus decode, and resulting numeral output voice B is outputed in the digiverter 22.22 couples of described numeral output voice B of digiverter carry out digital-to-analogue conversion (D-A conversion), and the output voice B of output analog signal.
In terminal C, the modulation signal A that comes self terminal A by circuit A is transfused to.26 couples of described modulation signal A of demodulating equipment carry out demodulation, and resulting phonetic code A is outputed on the audio decoding apparatus 27.27 couples of described phonetic code A of audio decoding apparatus carry out decoding processing, and resulting numeral output voice C is outputed on the digiverter 28.28 couples of described numeral output voice C of digiverter carry out digital-to-analogue conversion (D-A conversion), and the output voice C of output analog signal.
Described existing digitized voice communication system and the voice communication terminal that constitutes this system all have the following problem that solves of needing.
The regulation time frame unit of the speech coding that bit rate is lower and the enforcement of tone decoding mostly is about 10～40ms.Input voice B returns terminal B by terminal A, the needed time to be comprised in export voice B and be output till, be generally more than 6 times of this time frame, promptly can produce the time of delay of hundreds of ms.
Like this, caller B is in when conversation, because speech quality reduces, thereby while must listen as oneself the voice of the hundreds of ms of delay of the big echo of amplitude and converse, becomes very difficult thereby make to converse.In addition, the limit listens that oneself the voice limit conversation that postpones about 500ms is arranged is unusual difficulty, has many callers to have to temporarily stop conversation.
In addition, though on all terminals, all can hear the voice of caller A and caller B, but the input voice B that caller B is sent arrives from terminal B and terminal C output through terminal A, to handle through 2 speech codings of sound encoding device 24 and 17, so, speech quality will descend greatly.
Attempting to improve under the situation of line efficiency, because when speech coding, bit rate is set lowlyer by digitlization, so whenever carrying out a speech coding handles, quality all will have bigger decline, if carry out encoding process 2 times, then tonequality will be reduced to the stage that is difficult to not hear.
In addition; in general double talking; often have caller A and caller B gives orders or instructions simultaneously; the situation that is so-called double talking takes place; because the voice coding modes of low bit rate is to utilize independent sounding module to carry out effective Information Compression; so when double talking, the coding quality of the sound encoding device 17 in the terminal A reduces, the talk both sides can't not hear the other side at all and are saying something when serious.
The objective of the invention is to solve the above problems, avoid occurring following phenomenon: owing to postpone conversation difficulty that long echo causes, because the quality that 2 times codings cause descends, owing to the double talking voice being carried out the quality decline that encoding process causes.
Summary of the invention
Voice communication terminal of the present invention from each terminal input voice and phonetic code, carries out encoding process to the input voice, makes it to become the output phonetic code, and simultaneously to described phonetic code decoding, and as exporting voice from each terminal output, it has with lower device:
Sound encoding device carries out encoding process to the input voice, and resulting phonetic code is exported as the output phonetic code;
Audio decoding apparatus, to the input phonetic code carry out decoding processing, and with resulting decoded speech as output voice output;
Storage device is stored by the coded output phonetic code of described sound encoding device by some;
The similarity assessment device is assessed the output phonetic code stored in this storage device and the similarity of described input phonetic code, and is exported resulting similarity;
Means for correcting, similarity according to described similarity assessment device, the operation of described input phonetic code or described decoded speech or described audio decoding apparatus is carried out some proofread and correct, to dwindle the amplitude that results from described input phonological component that is contained in the output voice by described terminal output.
In addition, in voice communication terminal of the present invention, described means for correcting comprise the similarity based on described similarity assessment device decide the yield value of taking on the described decoded speech the yield value control device and
The yield value that the yield value control device is exported is taken on the described decoded speech, and the multiplying device of resulting result as the output of output voice signal.
And in voice communication terminal of the present invention, described means for correcting also comprises
Judgment means, based on the similarity of described similarity assessment device, determine whether to described input phonetic code proofread and correct and
The code displacement apparatus, make the decision of proofreading and correct in judgment means after, output has been converted to described input phonetic code the phonetic code of code designation, if judgment means is made the decision of not carrying out proofreading and correct, then former state is exported described input phonetic code,
Described audio decoding apparatus is decoded to the phonetic code that the code displacement apparatus is exported, and resulting decoded speech is exported as the output phonetic code.
And in voice communication terminal of the present invention, the phonetic code of being changed by described code displacement apparatus is a kind of fixing phonetic code of decodable code short arc decoded speech.
And, in voice communication terminal of the present invention, phonetic code is made of information code and its error-correcting code of expression voice messaging, described code displacement apparatus, after judgment means is made the correction decision, be transformed into described audio decoding apparatus decision error correcting code and surpass the wrong overlapping phonetic code of proofreading and correct boundary.
Described audio decoding apparatus when detecting when surpass proofreading and correct boundary wrong overlapping, diminishes continuously by the amplitude that makes decoded speech and decodes.
And in voice communication terminal of the present invention, described means for correcting also comprises
Judgment means is made the decision of whether operation of described audio decoding apparatus being proofreaied and correct based on described similarity,
Control device, make the decision of proofreading and correct when judgment means after, the signal that output is proofreaied and correct resulting yield value in the decode procedure of described audio decoding apparatus or error detection mark,
Described audio decoding apparatus according to the signal of described control device, is proofreaied and correct resulting yield value in decode procedure or error detection mark, carries out decoding processing then.
And in voice communication terminal of the present invention, described each phonetic code is made of information code and its error-correcting code of expression voice messaging,
Similarity in the described similarity assessment device, the information code in the output phonetic code that described storage device is stored and according to the error-correcting code in the same input phonetic code information code in the described input phonetic code is carried out after the error correction after the error correction relation between the information code and assess.
And, in voice communication terminal of the present invention, described similarity assessment device as similarity to have same number of bits or the identical rate of bit between the phonetic code that becomes comparison other.
And voice communication terminal of the present invention from each terminal input voice and phonetic code, carries out encoding process to the input voice, makes it to become the output phonetic code, and simultaneously to described phonetic code decoding, and as exporting voice from each terminal output, it comprises:
Sound encoding device carries out encoding process to the input voice, and resulting phonetic code is exported as the output phonetic code;
Audio decoding apparatus, to the input phonetic code decode, and with resulting decoded speech as output voice output;
Storage device, length are in accordance with regulations stored described input voice;
The similarity assessment device is assessed the input voice stored in this storage device and the similarity of described decoded speech, and is exported resulting similarity;
The yield value control device, based on the similarity of described similarity assessment device, the yield value on the described decoded speech is taken in decision;
The multiplying device is taken the yield value that the yield value control device is exported on the described decoded speech, and resulting result is exported as the output voice signal.
Voice communication system of the present invention comprises:
The 1st voice communication terminal, it has arbitrary structure of being put down in writing in～the 7 page of the 1st row of the 5th page of the 13rd row of this paper;
The 2nd voice communication terminal, the input voice, and in the time frame of each certain hour according to the certain standard and the phonetic code of described voice being encoded and obtaining by sound encoding device, select a kind of by choice device in the output phonetic code of exporting with described the 1st voice communication terminal and output, simultaneously, the output phonetic code that described the 1st voice communication terminal is exported is decoded by audio decoding apparatus and is exported.
And voice communication system of the present invention comprises
The 1st voice communication terminal, it has the structure of being put down in writing in～the 7 page of the 15th row of the 7th page of the 2nd row of this paper;
The 2nd voice communication terminal, it is imported voice and described voice is encoded by sound encoding device by the time frame of each regulation, in the output phonetic code that this phonetic code that is encoded and described the 1st voice communication terminal are exported, select a kind of by choice device and output according to certain standard, simultaneously, the output phonetic code of exporting from described the 1st voice communication terminal is decoded by audio decoding apparatus and export.
Description of drawings
Fig. 1 is the voice communication system structure chart in the embodiment of the present invention 1.
Fig. 2 is the voice communication terminal structure chart in the embodiment of the present invention 2.
Fig. 3 is the replacement Treatment example key diagram of code displacement apparatus in the embodiment of the present invention 2.
Fig. 4 is the example key diagram of code displacement apparatus in the embodiment of the present invention 3.
Fig. 5 is the voice communication terminal structure chart of embodiment of the present invention 4.
Fig. 6 is the voice communication terminal structure chart of embodiment of the present invention 5.
Fig. 7 is the voice communication system structure chart of embodiment of the present invention 6.
Fig. 8 is existing voice communication system overall construction drawing.
Fig. 9 is the existing voice communication system structure chart that system is carried out digitized processing.
Symbol description
1,2,3: voice communication terminal; 12: demodulating equipment; 13: audio decoding apparatus; 14: digiverter; 15: the voice overlapping device; 16: analog-digital commutator; 17: sound encoding device; 18: modulating device; 20: demodulating equipment; 21: audio decoding apparatus; 23: analog-digital commutator; 24: sound encoding device; 25: modulating device; 30: choice device; 31: storage device; 32: the similarity assessment device; 33: means for correcting; 34: the yield value control device; 35: the multiplying device; 36: judgment means; 37: the code displacement apparatus; 38: control device; 39: the information source coding device; 40: the error correction code device; 41: the error correcting/decoding device; 42: the information source decoding device.
Execution mode
Below in conjunction with accompanying drawing, embodiments of the present invention are explained.
Execution mode 1
Fig. 1 represents the structure of voice communication system of the present invention.In the drawings, voice communication terminal B is equivalent to the voice communication terminal among the present invention.
In the drawings, 1 is the employed terminal A of caller A, and 2 is the employed terminal B of caller B, the 3 terminal C that under the monitoring situation, adopted for other relevant personnel, and 10 is circuit A, 11 is circuit B.
In terminal A, 12 is demodulating equipment, and 13 is audio decoding apparatus, and 14 is digiverter, and 16 is analog-digital commutator, and 17 is sound encoding device, and 18 is modulating device, and these are identical with the existing voice communication system shown in Fig. 9.
30 by according to fixed choice criteria, selects one from the output of sound encoding device 17 and demodulating equipment 12, and the choice device of exporting to modulating device 18, and it has replaced the voice overlapping device 15 in the existing voice communication system shown in Figure 9.
In terminal B, 20 is demodulating equipment, and 21 is audio decoding apparatus, and 22 is digiverter, and 23 is analog-digital commutator, and 24 is sound encoding device, and 25 is modulating device, and these are identical with the existing voice communication system shown in Fig. 9.
But 31 storage devices for the memory block of N time frame with phonetic code B that storaged voice code device 24 imported, 32 is to comparing from the phonetic code A ' of demodulating equipment 20 and the phonetic code B of storage device 31, and the similarity assessment device that its similarity is assessed, 33 is means for correcting, it comprises according to coming the yield value control device 34 of output gain value from the similarity of similarity assessment device 18 and be multiplied by yield value from gain control 20 in the decoded speech that audio decoding apparatus 12 is exported, and with resulting result as the multiplying device 21 of digitlization output voice B to digiverter 13 outputs, these all are unexistent among the existing voice communication system terminal B shown in Figure 9.
Existing voice communication system with shown in Figure 9 among the terminal C is identical, and demodulating equipment 26, audio decoding apparatus 27 and digiverter 28 are also arranged.
Below in conjunction with accompanying drawing, its operation is described.
In terminal A, the voice that caller A sent are used as input voice A and import.Simultaneously, the modulation signal B from the terminal B that hereinafter will introduce is transfused to by circuit B.12 couples of described modulation signal B of demodulating equipment carry out demodulation, and resulting phonetic code B is outputed to audio decoding apparatus 13 and choice device 30.
13 couples of described phonetic code B of audio decoding apparatus decode, and resulting numeral output voice A is transported in the digiverter 14.14 couples of described numeral output voice A of digiverter carry out digital-to-analogue conversion (D-A conversion), and the output voice A of output analog signal.
By analog-digital commutator 16, described input voice A carried out analog-to-digital conversion (A-D conversion) thus obtain the numeral input voice A of digital signal.17 couples of described numeral input voice A of sound encoding device carry out encoding process, and resulting phonetic code A is outputed to choice device 30.Choice device 30 choice criteria is according to the rules selected between input phonetic code A and phonetic code B, and selected phonetic code is outputed to modulating device 18 as phonetic code A '.
A method of choice criteria can be to after phonetic code A and the actual decoding of phonetic code B, the amplitude of resulting 2 decoded speech is compared, and therefrom select 1 bigger side of numerical value.Then, the phonetic code A ' that 18 pairs of choice devices 30 of modulating device are imported carries out digital modulation, and resulting modulation signal is transported among the circuit A10 as modulation signal A.
In terminal B, the voice that caller B sent are used as input voice B and import.Simultaneously, by circuit A, come the modulation signal A of self terminal A to be transfused to.Can obtain described input voice B is carried out the numeral input voice B of the back digital signal that is generated of analog-to-digital conversion (A-D conversion) by analog-digital commutator 23.24 couples of described numeral input voice B of sound encoding device carry out encoding process, and resulting phonetic code B is transported in modulating device 25 and the storage device 31.The phonetic code B of 25 pairs of inputs of modulating device carries out digital modulation, and resulting modulation signal is transported to circuit B11 as modulation signal B.
20 couples of described modulation signal A of demodulating equipment in the terminal B carry out demodulation, and resulting phonetic code A ' is transported in audio decoding apparatus 21 and the similarity assessment device 32.The phonetic code B that 31 pairs of sound encoding devices 24 internally of storage device are imported carries out the storage of N time frame, and part or all is transported in the similarity assessment device 32 with it.
In addition, having in storage device 31 can be to the memory block of phonetic code B storage N time frame storage, carry out phonetic code output to similarity assessment device 32 after, the memory block of the described phonetic code B of storage N time frame will be topped by the phonetic code B of current time frame institute, with this memory contents be upgraded.If can the phonetic code B of N time frame be stored, then structure and the update method in the storage device 31 can not limit by this.
Carrying out N time frame when storage, turn back to the time-delay that terminal B is generated owing to will absorb phonetic code B again by terminal A, so when setting the N value, frame number when being greater than the time-delay of anticipation.
And, because based on the structure difference of circuit kind or terminal A, frame number is also inequality during time-delay, so need to export a plurality of phonetic code B (part or all) that store when the time comes.
The phonetic code B more than 1 that phonetic code A ' that 32 pairs of demodulating equipments 20 of similarity assessment device are imported and storage device 31 are imported compares, its similarity is assessed in order, and will be wherein maximum value be transported in the yield value control device 34 in the means for correcting 33.The similarity here can adopt the same number of bits or the identical bit of 2 phonetic codes.
If the similarity that similarity assessment device 32 is imported surpasses the threshold value of regulation, then can conclude that by yield value control device 34 phonetic code B returns by terminal A, simultaneously to the multiplying device 35 outputs yield value littler than 1.If the similarity that similarity assessment device 32 is imported is lower than defined threshold, then to multiplying device 35 output gain values 1.
In addition, if yield value is from 1 anxious downwards subtracting, perhaps conversely from suddenly increasing to 1 below 1, discontinuity will appear in the output result of the multiplying device 35 that then will introduce hereinafter, thereby for the time frame that yield value changes, can increase or reduce numerical value gradually to each sampling, to little value or 1 close.
Form though the threshold value of defined generally be can't help fixed value, the similarity of a current time frame surpass get the small value during fixed threshold value, the similarity of a current time frame fixed threshold value get big value when following, therefore can suppress the frequent variations of yield value.
In addition, also can suppress the frequent variations of the yield value of each time frame with reference to described control result.
The phonetic code A ' of 21 pairs of inputs of audio decoding apparatus decodes, and resulting decoded speech is transported in the multiplying device 35.
Each sample value of the described decoded speech that 35 pairs of audio decoding apparatus 21 of multiplying device are imported, multiply by each yield value of sampling that yield value control device 34 is imported, and resulting result is transported in the digiverter 22 as numeral output voice B.The output voice B of the analog signal after the digital-to-analogue conversion (D-A conversion) is carried out in digiverter 22 output to described numeral output voice B.
By circuit A, the modulation signal A of self terminal A is input among the voice communication terminal C in the future.26 couples of described modulation signal A of demodulating equipment carry out demodulation, and resulting phonetic code A ' is transported on the audio decoding apparatus 27.27 couples of described phonetic code A ' of audio decoding apparatus decode, and resulting numeral output voice C is transported on the digiverter 28.The output voice C of the analog signal after the digital-to-analogue conversion (D-A conversion) is carried out in digiverter 28 output to described numeral output voice C.
In said embodiment, though often carry out the computing of yield value, be under 1 the situation at yield value, because operation result does not change by multiplying device 35, so can end the computing of arithmetic unit 35, make described decoded speech keep the former state of numeral output voice B.
In said embodiment, though the structure that adopts wireless communication line and analog-modulated demodulating equipment transmission line etc. is described, also can adopt the circuit of other kinds such as ATM as path.At this moment, the structure of modulating device and demodulating equipment becomes according to the kind of circuit.
Usually, voice communication terminal C is identical with the structure of voice communication terminal B, under the situation that voice communication terminal A and voice communication terminal B converse, because the storage device in the voice communication terminal B is the storaged voice sign indicating number not, so frequent input gain value 1 in multiplying device 35, the conversation between voice communication terminal A and the voice communication terminal C is as output voice B output.Certainly, voice communication terminal a plurality of and voice communication terminal B and voice communication terminal C same structure also can be set.
In this execution mode 1, adopt following 3 kinds of devices, i.e. quantity storage device that described phonetic code B is stored in accordance with regulations, the phonetic code B that stored in this storage device and the similarity of A ' are assessed, and export the similarity assessment device of resulting assessed value, at least based on described similarity decoded speech is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, like this, in voice communication terminal, just can avoid owing to the problem that can't converse that causes of the excessive echo of time-delay.
In addition, because in the described means for correcting, comprise the yield value control device that decides the yield value of taking on the described decoded speech at least based on described similarity, take multiplying device on the described decoded speech with the yield value that the yield value control device is exported, so can dwindle the amplitude of part that results from input voice B that in the output voice B of final output, is contained, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, because in described similarity assessment device, same number of bits or identical bit between the phonetic code that becomes comparison other have been adopted as similarity, even so there is a spot of bit mistake overlapping on the transmission line, whether the phonetic code B of the output of also can correctly detecting over returns with the form of phonetic code A ', just can avoid in voice communication terminal owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, (voice communication terminal B) has the storage device that described phonetic code B is stored with specified quantity in the 1st voice communication terminal, the phonetic code B that stores in this storage device and the similarity of phonetic code A ' are assessed, and with the similarity assessment device of resulting similarity output, at least based on described similarity decoded speech is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, the 2nd voice communication terminal (voice communication terminal A) is when the output phonetic code that the 1st voice communication terminal is exported is decoded and exported, every certain time frame, select it to export in the lump between the output phonetic code that resulting phonetic code and described the 1st voice communication terminal are exported after the 2nd input voice (input voice A) are carried out encoding process, because 2 such terminals are arranged, so just can in voice communication system, avoid because the problem that the quality that 2 encoding process are brought descends is avoided because the excessive echo of delaying time causes the problem that can't converse.
In addition, since only the phonetic code behind the coding that caller A or caller B are imported decode the quality decline problem that coding caused of overlapping call voice so voice communication system just can be avoided.
Execution mode 2
Fig. 2 represents the another kind of structure of voice communication terminal of the present invention.This voice communication terminal is during as the voice communication terminal B among Fig. 1, the general structure of expression voice communication system.Omitted the part identical among the figure with Fig. 1 symbol.33 is means for correcting, and it comprises whether surpass judgment means 36 that defined threshold judges and the code displacement apparatus 37 that the phonetic code A ' from demodulating equipment 20 is carried out replacement Treatment according to this judged result from the similarity of similarity assessment device 32.
Below in conjunction with accompanying drawing its operation is explained.
The voice that caller B is sent are imported among the voice communication terminal B as input voice B.And, be transfused to by circuit A from the modulation signal A of voice communication terminal A.Can obtain described input voice B is carried out the numeral input voice B of the back digital signal that is generated of analog-to-digital conversion (A-D conversion) by analog-digital commutator 23.24 couples of described numeral input voice B of sound encoding device carry out encoding process, and resulting phonetic code B is transported in modulating device 25 and the storage device 31.The phonetic code B of 25 pairs of inputs of modulating device carries out digital modulation, and resulting modulation signal is transported in the circuit B as modulation signal B.
20 couples of described modulation signal A of demodulating equipment in the voice communication terminal B carry out demodulation, and resulting phonetic code A ' is transported in the code displacement apparatus 37 in similarity assessment device 32 and the means for correcting 33.The phonetic code B that the sound encoding device 24 of 31 pairs of inside of storage device is imported carries out the storage of N time frame, and part or all is transported in the similarity assessment device 32 with it.
The phonetic code A ' that 32 pairs of demodulating equipments 20 of similarity assessment device are imported and compare from the phonetic code B more than 1 of storage device 31, in order its similarity is assessed, and wherein maximum is transported in the judgment means 36 in the means for correcting 33.The similarity here can adopt the same number of bits or the identical bit of 2 phonetic codes.
Whether the similarity of 36 pairs of similarity assessment devices 32 of judgment means surpasses defined threshold is judged, and judged result is transported on the code displacement apparatus 37.The same with the yield value control device 34 of execution mode 1, even this defined threshold fixed value also can be carried out suitable control.
Express similarity in judged result and surpass under the situation of defined threshold, 37 couples of phonetic code A ' from demodulating equipment 20 of code displacement apparatus carry out replacement Treatment, and with resulting phonetic code A " be transported on the audio decoding apparatus 21.If judged result represents similarity and be lower than setting, the phonetic code A ' that demodulating equipment 20 is imported keeps intact motionless, and it as phonetic code A " be transported in the audio decoding apparatus 21.
Fig. 3 is a displacement example key diagram in the code displacement apparatus 37.In the drawings, the phonetic code of (a) being exported for demodulating equipment 20 (phonetic code A ') (b) is the phonetic code of being replaced by code displacement apparatus 37 (phonetic code A ").In this embodiment, 37 of code displacement apparatus will represent that the part of the gain code I (0110011) of gain information in the phonetic code is replaced into fixing gain code II (0000000).Find out the gain code of the decoded speech that can generate little amplitude in advance, if, then can dwindle the amplitude of the decoded speech after the displacement the fixed gain sign indicating number of this gain code as displacement.Such as, be under the situation of general CELP mode at voice coding modes, use source of sound and be almost 0 if be replaced into, drive the also very little gain code of source of sound gain, then can make the amplitude of decoded speech very little.
In addition, if permutation code is not limited to gain code, and last decoded speech amplitude can be very little, and then phonetic code can partly or entirely be replaced.Under voice coding modes, owing to do not have a gain code sometimes, so can carry out at that time to displacement about the code relevant such as the code of power with amplitude.
The phonetic code A of 21 pairs of inputs of audio decoding apparatus " carry out decoding processing, and resulting decoded speech is transported in the digiverter 22 as numeral output voice B.Digiverter 22 is used for the analog signal of digital-to-analogue conversion (D-A conversion) is carried out in output to described numeral output voice B output voice B.
In addition, identical with execution mode 1, also can adopt the circuit of other kinds such as ATM.And, usually the structure of voice communication terminal C is identical with voice communication terminal B, when voice communication terminal A and voice communication terminal C conversation, because the storage device in the voice communication terminal B is the storaged voice sign indicating number not, so do not carry out the displacement of code, the conversation of voice communication terminal A and voice communication terminal C is with the form output of output voice B.The voice communication terminal of several and voice communication terminal B and voice communication terminal C same structure also can be set certainly.
In execution mode 2, adopt 3 kinds of devices, promptly to the described phonetic code B storage device stored of quantity according to the rules, to assessing at the phonetic code B of this storage device stored and the similarity of phonetic code A ', and export the apparatus for evaluating of the similarity of resulting assessed value, at least based on described similarity phonetic code A ' is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
And, comprise in the described means for correcting at least based on described similarity, the judgment means whether decision proofreaies and correct described input phonetic code A ', after making the decision of proofreading and correct by judgment means, output is replaced as the phonetic code A of code designation with part or all of described phonetic code A ' "; if judgment means is made the decision of not carrying out proofreading and correct; then make described input phonetic code with phonetic code A " the code displacement apparatus of form former state output, therefore, can dwindle in the output voice B of final output, contained result from input voice B part amplitude and decode, and carry out code as principle and replace, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, because in described similarity assessment device, same number of bits or identical bit between the phonetic code that becomes comparison other have been adopted as similarity, even so there is a spot of bit mistake overlapping on the transmission line, whether the voice signal B of the output of also can correctly detecting over returns with the form of phonetic code A ', just can avoid in voice communication terminal owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, (voice communication terminal B) comprises the storage device that described phonetic code B is stored with specified quantity in the 1st voice communication terminal, the phonetic code B that stores in this storage device and the similarity of A ' are assessed, and with the similarity assessment device of resulting similarity output, at least based on described similarity phonetic code A ' is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, the output phonetic code that the 2nd voice communication terminal (voice communication terminal A) is exported the 1st voice communication terminal is decoded and is exported, simultaneously every certain time frame, select it to export in the lump between the phonetic code that resulting phonetic code and described the 1st voice communication terminal are exported after the 2nd input voice (input voice A) are carried out encoding process, because 2 such terminals are arranged, so can avoid because the problem that 2 quality of being brought of coding descend just can be avoided in voice communication system because the excessive echo of delaying time causes the problem that can't converse.In addition, because only the phonetic code that the input voice of caller A or caller B are encoded is decoded, so can avoid the quality decline problem of the voice communication system under the effect of overlapping call voice coding.
Execution mode 3
Execution mode 3 is applicable to the situation that phonetic code is made up of information code and its error-correcting code of simple expression voice messaging.
General radio link communication speech sign indicating number all is made up of information code and its error-correcting code of simple expression voice messaging.In the information code of expression voice messaging because the significance level of per 1 bit all lays particular stress on to some extent, so only the higher bit of the importance degree of computational rules number with the corresponding error correcting code.
Though the structure of the voice communication terminal in the execution mode 3 is identical with the diagram 2 of execution mode 2, but owing to phonetic code is made up of the information code and the error-correcting code of simple expression phonetic code information, so the structure of the inside of sound encoding device 24, audio decoding apparatus 21 and code displacement apparatus 37 is different.
Fig. 4 is the key diagram of another example of the replacement Treatment of code displacement apparatus 37 in execution mode 3 voice communication terminals.In the drawings, the phonetic code of (a) being exported for demodulating equipment 20 (phonetic code A ') (b) is the phonetic code that cemented out by code displacement apparatus 37 (phonetic code A ").
Sound encoding device 24 generates the information code of simple expression voice messaging, only the higher bit of importance degree is wherein concentrated, and it is calculated error-correcting code, and information code and error-correcting code are lumped together, and exports with the phonetic code form.
Audio decoding apparatus 21 utilizes the error-correcting code in the phonetic code that the mistake in the information code is carried out treatment for correcting, the information code after the error correction is decoded again.
21 pairs of audio decoding apparatus surpass that the mistake of error recovery boundary is whether overlapping to be judged, generate the error detection mark of expression judged result, if this mark represents to surpass the correction boundary, then will be discarded as the information code (significant bits of specified quantity) of error correction object, the value of time frame etc. is carried out decoding processing before being replaced as.
If there is continuous a plurality of expression to surpass the wrong overlapping mark of proofreading and correct boundary, the amplitude of decoded speech is slowly dwindled, proofread and correct control.Like this, just can in limited transmission amount of information, realize effective fault-tolerant improvement.
37 of code displacement apparatus show the similarity of self similarity degree apparatus for evaluating 32 above under the situation of defined threshold in judgment means 36, just the phonetic code A ' from demodulating equipment 20 is carried out replacement Treatment, and with resulting phonetic code A " be transported in the audio decoding apparatus 21.Concrete method of replacing is: the information code III of error-correcting code in the phonetic code shown in Figure 4 and error correction object is replaced as the error-correcting code in the phonetic code shown in Fig. 4 (b) and the information code IV of error correction object.
The difference of the method for replacing of this displacement and described audio decoding apparatus 21 is that it judges the mistake of surpassing in the audio decoding apparatus 21 proofreading and correct boundary is whether overlapping.Specifically, as the displacement code, utilize whether the overlapping fixed code that detects is replaced as the displacement code to the mistake above the error correction boundary.
In replacement process, audio decoding apparatus 21 is discarded the information code (bit of important specified quantity) of error correction object, replaces preceding 1 time frame value etc. and carries out decoding processing.If the displacement of code displacement apparatus 37 is continuous, then audio decoding apparatus 21 will make the decoded speech amplitude slowly reduce, and carry out treatment for correcting.
The structure of this execution mode 3 as shown in Figure 2, because error-correcting code surpasses that the mistake of proofreading and correct boundary is whether overlapping to be judged among 37 couples of phonetic code A ' of code displacement apparatus, replace in this way, so the effect that in execution mode 2, is had, also having amplitude, that possibility jumpy takes place is little, and decoded speech tonequality is effect preferably.
Execution mode 4
Fig. 5 represents the structure of the voice communication terminal of execution mode 4.This voice communication terminal is then represented the general structure of voice communication system during as the voice communication terminal B among Fig. 1.Omitted the part or all of explanation under the same-sign with Fig. 1 and Fig. 2 in this example.33 is means for correcting, and it comprises whether surpassing judgment means 36 that defined threshold judges from the similarity of similarity assessment device 32 and according to the control device of judging when similarity surpasses defined threshold the little correction coefficient of audio decoding apparatus 21 conveying ratios that resulting yield value is taken advantage of in the decode procedure of audio decoding apparatus 21 1 38.
Below in conjunction with accompanying drawing its operation is explained.
The voice that caller B sends are transported among the voice communication terminal B as input voice B.And by circuit A, input is from the modulation signal A of voice communication terminal A.Can obtain described input voice B is carried out the numeral input voice B of the back digital signal that is generated of analog-to-digital conversion (A-D conversion) by analog-digital commutator 23.24 couples of described numeral input voice B of sound encoding device carry out encoding process, and resulting phonetic code B is transported in modulating device 25 and the storage device 31.25 couples of phonetic code B that imported of modulating device carry out digital modulation, and resulting modulation signal is transported to circuit B as modulation signal B.
20 couples of described modulation signal A of demodulating equipment in the voice communication terminal B carry out demodulation, and resulting phonetic code A ' is transported in audio decoding apparatus 21 and the similarity assessment device 32.31 couples of phonetic code B from the sound encoding device 24 of inside of storage device store the N time frames, and part or all is transported in the similarity assessment device 32 it.
32 pairs in similarity assessment device compares from the phonetic code A ' of demodulating equipment 20 with from the phonetic code B more than 1 of storage device 31, in order its similarity is assessed, and wherein maximum is transported in the judgment means 36 in the means for correcting 33.
Whether 36 pairs of similarities from similarity assessment device 32 of judgment means surpass defined threshold is judged, and judged result is transported on the control device 38.
Control device 38 is being judged similarity above under the situation of defined threshold, and the ratio 1 little correction coefficient that resulting yield value in the decode procedure is taken advantage of is transported in the audio decoding apparatus 21.At the consecutive numbers that surpasses the judged result of setting according to similarity, the control correction coefficient is carried out interim dwindling, the discontinuity that can suppress to control thus.
21 couples of phonetic code A ' that imported of audio decoding apparatus decode, and resulting decoded speech is transported on the digiverter 22 as numeral output voice B.But, from 38 inputs of described control device during, the yield value of resulting time frame in decode procedure or period of the day from 11 p.m. to 1 a.m frame (partial section of time frame) be multiply by correction coefficient, and then carries out decoding processing the correction coefficient of yield value.
Digiverter 22 is used for described numeral is exported the output voice B that voice B carries out digital-to-analogue conversion (D-A conversion) and exports analog signal.
In addition, described control device 38 also can replace the correction coefficient of yield value, the indication that output is proofreaied and correct the error detection mark in the audio decoding apparatus 21 of explanation in the execution mode 3.
The error detection mark is that expression is input to the information that whether overlapping phonetic code in the audio decoding apparatus 21 surpass to proofread and correct the mistake of boundary judged result as mentioned above, 38 pairs of these error detection marks of control device are proofreaied and correct, and the value when set to surpass proofreading and correct boundary wrong overlapping, in audio decoding apparatus 21, the same with explanation in the execution mode 3 handled the situation that the mistake that surpasses the correction boundary overlaps.
Be exactly the information code (bit of important specified quantity) of discarded error correction object specifically, replace described time frame value etc. and carry out decoding processing.If the correction of error detection mark is continuous, then audio decoding apparatus 21 will reduce at leisure to the amplitude of decoded speech, the decoding while proofreading and correct.
In execution mode 4, adopt 3 kinds of devices, promptly to the described phonetic code B storage device stored of quantity according to the rules, to assessing at the phonetic code B of this storage device stored and the similarity of phonetic code A ', and export the similarity assessment device of resulting assessed value, based on described similarity the operation of audio decoding apparatus is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, owing to have at least the judgment means of whether operation of audio decoding apparatus being proofreaied and correct in the described means for correcting based on described similarity decision, with make the control device that sends the indication that resulting yield value in described decode procedure or error detection mark are proofreaied and correct after the decision of proofreading and correct in judgment means, audio decoding apparatus is according to proofreading and correct from the indication of described control device or to resulting yield value in the decode procedure, or the error detection mark proofreaied and correct gradually and decode, so can dwindle the amplitude of part that results from input voice B that in the output voice B of final output, is contained, carry out the correction of decoding processing operation as principle, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, (voice communication terminal B) has the storage device that described phonetic code B is stored with specified quantity in the 1st voice communication terminal, the phonetic code B that stores in this storage device and the similarity of phonetic code A ' are assessed, and with the similarity assessment device of resulting similarity output, at least based on described similarity the operation of audio decoding apparatus is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, the 2nd voice communication terminal (voice communication terminal A) is when voice communication terminal is to the back output of decoding from the output phonetic code of the 1st voice communication terminal, every certain time frame, after being carried out encoding process, the 2nd input voice (input voice A) select between the phonetic code of resulting phonetic code and the output of described the 1st voice communication terminal, because 2 such terminals are arranged, so can avoid because the problem that 2 quality of being brought of coding descend just can be avoided in voice communication system because the excessive echo of delaying time causes the problem that can't converse.
In addition, since only the phonetic code behind the coding that caller A or caller B are imported decode the quality decline problem that overlapping call voice coding is caused so voice communication system just can be avoided.
In execution mode 2,, under this execution mode, can proofread and correct each time frame in the audio decoding apparatus 21 or the yield value of period of the day from 11 p.m. to 1 a.m frame though each output sampling of audio decoding apparatus 21 all be multiply by yield value.Though execution mode 2 need steadily be handled each sampling in large quantities in the yield value multiplying, because the independence of device is higher, audio decoding apparatus 21 can not changed, and this is its strong point.Though present embodiment independence is lower, gain calibration is handled simple, and this also is a strong point.
In addition, (voice communication terminal B) comprises the storage device that described phonetic code B is stored with specified quantity in the 1st voice communication terminal, the phonetic code B that stores in the storage device and the similarity of A ' are assessed, and with the similarity assessment device of resulting similarity output, at least based on described similarity the operation of audio decoding apparatus is proofreaied and correct, in order to dwindle the means for correcting of amplitude of part that results from input voice B that in the output voice B of final output, is contained, the 2nd voice communication terminal (voice communication terminal A) is when voice communication terminal is to the back output of decoding from the output phonetic code of the 1st voice communication terminal, every certain time frame, after being carried out encoding process, the 2nd input voice (input voice A) select between the phonetic code of resulting phonetic code and the output of described the 1st voice communication terminal, because 2 such terminals are arranged, so can avoid because the problem that 2 quality of being brought of coding descend just can be avoided in voice communication system because the excessive echo of delaying time causes the problem that can't converse.
In addition, since only the phonetic code behind the coding that caller A or caller B are imported decode the quality decline problem that overlapping call voice coding is caused so voice communication system just can be avoided.
Execution mode 5
Though the similarity in described execution mode 1～4 has adopted same number of bits or identical bit between the phonetic code that becomes comparison other, but under the situation that information code and error-correcting code thereof constituted of each phonetic code by the expression voice messaging, can also carry out information code after the error correction to phonetic code A ' for the information code among the phonetic code B with according to error-correcting code, obtain its same number of bits or identical bit as similarity.
Fig. 6 represents to have the voice communication terminal structure of this effect.And this voice communication terminal is during as the voice communication terminal B among Fig. 1, the general structure of expression voice communication system.Omitted in this example with Fig. 1 same-sign under part or all of explanation.In the drawings, sound encoding device 24 constitutes by the code device 39 that the information source of simple expression voice messaging is encoded with to the error correction code device 40 that its error recovery information is carried out encoding process, and audio decoding apparatus 21 constitutes by the error correcting/decoding device 41 that error-correcting code is carried out decoding processing with to the information source decoding device 42 that the information source code carries out decoding processing.
Below in conjunction with accompanying drawing operation is explained.
The voice that caller B is sent are transported among the voice communication terminal B as input voice B.And by circuit A, input is from the modulation signal A of voice communication terminal A.Can obtain described input voice B is carried out the numeral input voice B of the back digital signal that is generated of analog-to-digital conversion (A-D conversion) by analog-digital commutator 23.
39 couples of described numeral input voice B of information source coding device in the sound encoding device 24 are encoded into a spot of information, and resulting information code B is transported in the error correction code device 40 and storage device 31 in the sound encoding device 24.Applied representative manner is the CELP mode in this information source coding device 39.
Error correction code device 40 calculates all or part of error-correcting code B at described information code B, and it is added on the information code B, more resulting result is transported on the modulating device 25 as phonetic code B.Here used representational error correction coded system has convolutional encoding and CRC coding etc.Adopting under the CELP information source coding mode, collect the significant bits among a plurality of information code B usually, on them, add corresponding error-detecting code (CRC), a plurality of bits of CRC object and CRC sign indicating number are carried out convolutional encoding handle.
25 couples of phonetic code B that imported of modulating device carry out digital modulation, and resulting modulation signal is transported on the circuit B11 as modulation signal B.
20 couples of described modulation signal A of demodulating equipment in the voice communication terminal B carry out demodulation, and resulting phonetic code A ' is transported on the error correcting/decoding device 41 in the audio decoding apparatus 21.
Error correcting/decoding device 41 is being included in the error-correcting code among the phonetic code A ' and being released by the bit groups of this correcting code protection; carry out error correcting/decoding and handle, again bit groups after proofreading and correct and the combination of the bit groups outside the object of protection are transported on information source decoding device 42 and the similarity assessment device 32 as information code A '.If in error-correcting code, contain the error-detecting code of CRC and so on, then combine and be transported on information source decoding device 42 and the similarity assessment device 32 with the error detection result.
31 couples of information code B from the information source coding device 39 of inside of storage device carry out N time frame storage, and it part or all is transported on the similarity assessment device 32.The memory block of the information code B of storage N time frame is arranged in storage device 31, and behind similarity assessment device 32 mail message sign indicating numbers, the information code B before the N time frame will be replaced by the information code B of current time frame, thus the updated stored content.As long as can the information code B of N time frame be stored, then structure and the update method in the storage device 31 can not limit by this.
The phonetic code A ' that 32 pairs of error correction demodulating equipments 41 of similarity assessment device are imported and compare from the phonetic code B more than 1 of storage device 31, in order its similarity is assessed, and wherein maximum is transported in the gain control 34 in the means for correcting 33.Here used similarity can adopt the same number of bits or the identical bit of 2 phonetic codes.
If the similarity from similarity assessment device 32 surpasses the threshold value of stipulating, then can conclude that phonetic code B returns by voice communication terminal A, simultaneously to the multiplying device 35 outputs yield value littler than 1 by yield value control device 34.If the similarity from similarity assessment device 32 is lower than defined threshold, then to multiplying device 35 output gain values 1.
42 couples of phonetic code A ' that imported of information source decoding device in the audio decoding apparatus 21 carry out decoding processing, and resulting decoded speech is transported in the multiplying device 35.And if the error detection result shows mistake, the decoded speech of the information code A ' of time frame or preceding time frame is carried out correcting process and decoding gradually before then utilizing, and resulting decoded speech is transported in the multiplying device 35.
35 pairs of each sample values from the described decoded speech of audio decoding apparatus 42 of multiplying device multiply by the yield value from each sampling of gain control 34, and resulting result is transported in the digiverter 22 as numeral output voice B.The output voice B of the analog signal after the digital-to-analogue conversion (D-A conversion) is carried out in digiverter 22 output to described numeral output voice B.
Certainly, under the situation that adopts this voice communication system, the sound encoding device of other voice communication terminals also all should be identical with the structure of sound encoding device 24 and audio decoding apparatus 21 with audio decoding apparatus.
Here, sound encoding device in the execution mode 1 shown in Figure 1 is made of information source coding device and error correction code device, audio decoding apparatus is made of error correcting/decoding device and information source decoding device, when the similarity of carrying out information code is calculated, execution mode 2 shown in Fig. 2 and Fig. 5 to the sound encoding device in the execution mode 4 can carry out identical change with the structure of audio decoding apparatus, can carry out the calculation of similarity degree of information code under these conditions.
In addition, specifically, in each figure, part in the sound encoding device is replaced by the information source coding device, between it and modulating device 25, insert the error correction code device, insert the error correcting/decoding device in the back of demodulating equipment 20, the part of audio decoding apparatus 21 is replaced by the information source decoding device.
In the similarity assessment device of execution mode 5, as similarity, in the information code of described storage device stored with the information code among the phonetic code A ' is carried out according to the error-correcting code among the same phonetic code A ' assess between the information code of error correction, even on transmission line, there is a spot of mistake overlapping like this, whether the phonetic code B that is exported of also can correctly detecting over returns as phonetic code A ', thereby just can make voice communication terminal avoid taking place because the conversation difficulty that bigger time-delay echo is caused.
Execution mode 6
Fig. 7 represents the another kind of structure of voice communication system of the present invention.In the drawings, voice communication terminal B is equivalent to voice communication terminal of the present invention.Be omitted with the same section explanation of Fig. 1 and Fig. 9 same-sign.31 for carrying out the storage device that the N time frame is stored to the numeral input voice B from analog- digital commutator 23,32 for to comparing with numeral input voice B from storage device 31 from the decoded speech of audio decoding apparatus 21, and the similarity assessment device that its similarity is assessed.
Be explained below in conjunction with the operation of accompanying drawing it.
The voice that caller A is sent are input among the voice communication terminal A as input voice A.And, by circuit B, be transfused to from the modulation signal B of voice communication terminal B described later.12 couples of described modulation signal B of demodulating equipment carry out demodulation, and resulting phonetic code B is transported on the audio decoding apparatus 13.13 couples of described phonetic code B of audio decoding apparatus decode, and resulting numeral output voice A is transported on the digiverter 14.14 couples of described numeral output voice A of digiverter carry out digital-to-analogue conversion (D-A conversion), and the output voice A of resulting analog signal is exported.
In voice communication terminal A, generate described input voice A and the overlapping overlapping voice of output voice A, and it is transported in the analog-digital commutator 16.The overlapping voice of 16 pairs of inputs of analog-digital commutator carry out analog-to-digital conversion (A-D conversion), thereby obtain the numeral input voice A of digital signal.17 couples of described numeral input voice A of sound encoding device carry out encoding process, and resulting phonetic code A is transported on the modulating device 18.18 couples of phonetic code A that imported of modulating device carry out digital modulation, and resulting modulation signal is transported on the circuit 10 as modulation signal A.
The voice that caller B is sent are input among the voice communication terminal B as input voice B.And, by circuit A10, be transfused to from the modulation signal A of voice communication terminal A.Can obtain described input voice B is carried out the numeral input voice B of the digital signal that analog-to-digital conversion (A-D conversion) back generated by analog-digital commutator 23, and it is transported on sound encoding device 24 and the storage device 31.24 couples of described numeral input voice B of sound encoding device carry out encoding process, and resulting phonetic code B is transported in the modulating device 25.The phonetic code B of 25 pairs of inputs of modulating device carries out digital modulation, and resulting modulation signal is transported in the circuit B11 as modulation signal B.
31 couples of numeral input phonetic code B from the analog-digital commutator 23 of inside of storage device carry out the storage of N time frame, and part or all is transported on the similarity assessment device 32 with it by time frame unit.And, the memory block that the numeral input voice B of storage N time frame is arranged in storage device 31, after carrying numeral input voice to similarity assessment device 32, the zone that numeral input voice B before the N time frame is stored will be covered by the numeral of current time frame input voice B, and then the updated stored content.In addition, as long as can the numeral input voice B of N time frame be stored, then structure and the update method in the storage device 31 can not limit by this.
20 couples of described modulation signal A of demodulating equipment in the voice communication terminal B carry out demodulation, and resulting phonetic code A ' is transported in the audio decoding apparatus 21.21 couples of phonetic code A ' that imported of audio decoding apparatus decode, and resulting decoded speech is transported in similarity assessment device 32 and the multiplying device 35.
32 pairs in similarity assessment device is from the decoded speech of 1 time frame of audio decoding apparatus 21 and be that the numeral input phonetic code B more than 1 that unit imported compares with the time frame from storage device 31, its similarity is assessed in order, and will be wherein maximum be transported in the yield value control device 34 in the means for correcting 33.And used here similarity can adopt the inverse of the vector distance of 2 voice.
If the similarity from similarity assessment device 32 surpasses the threshold value of stipulating, then can conclude that by yield value control device 34 phonetic code B returns by voice communication terminal A, simultaneously to the multiplying device 35 outputs yield value littler than 1.If the similarity from similarity assessment device 32 is lower than defined threshold, then to multiplying device 35 output gain values 1.
35 pairs of each sample values from the described decoded speech of audio decoding apparatus 21 of multiplying device multiply by the yield value from each sampling of gain control 34, and resulting result is transported in the digiverter 22 as numeral output voice B.Digiverter 22 is exported the output voice B that described numeral output voice B is carried out the analog signal after the digital-to-analogue conversion (D-A conversion).
In addition, in said embodiment, though often carry out the multiplying of yield value by multiplying device 35, but be under 1 the situation at yield value, because operation result does not change, so can end the computing of multiplying device 35, described decoded speech is kept intact, and as numeral output voice B.
In addition, in said embodiment,, also can adopt the circuit of other kinds such as ATM though be that the structure that adopts radiolink and analog-modulated demodulate transmitted circuit etc. is described.At this moment, the structure of modulating device and demodulating equipment will become according to the kind of circuit.
Usually, voice communication terminal C is identical with the structure of voice communication terminal B, under the situation that voice communication terminal A and voice communication terminal C converse, because the storage device in the voice communication terminal B is the storaged voice sign indicating number not, so multiplying device 35 frequent input gain values 1, the conversation between voice communication terminal A and the voice communication terminal C will be used as output voice B output.Certainly, voice communication terminal a plurality of and voice communication terminal B and voice communication terminal C same structure also can be set.
In this execution mode 6, adopt following apparatus, to the described numeral input voice B storage device stored of length according to the rules, to assessing at the numeral input voice B of this storage device stored and the similarity of decoded speech, and export the similarity assessment device of resulting similarity, at least take the yield value control device of the yield value on the described decoded speech based on described similarity decision, take on the described decoded speech with the yield value that the yield value control device is exported, and with the multiplying device of resulting result output, like this, in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
In addition, because when having this voice communication terminal (voice communication terminal B), also having the phonetic code that this voice communication terminal B is exported decodes also as output voice A output, the voice signal that simultaneously this output voice A and the 2nd input voice (input voice A) is added and obtain is encoded, and the 2nd voice communication terminal (voice communication terminal A) that resulting phonetic code is exported, so in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes the problem that can't converse.
The invention effect
Described voice communication terminal of the present invention is from each terminal input voice and phonetic code, to input The voice processing of encoding makes it to become the output phonetic code, and to described phonetic code decoding, , fail owing to it has to dwindle at voice communication terminal from each terminal output as the output voice The means for correcting of the amplitude of the contained part that results from the input voice in the output voice that go out, So in voice communication terminal, just can avoid owing to the excessive echo of time-delay causes and can't lead to The problem of words.
And described means for correcting also comprises based on exporting phonetic code and importing phonetic code mutually Like degree, determine yield value control device and the handle of taking the yield value on the described decoded speech to increase The yield value that benefit value control device is exported is taken the multiplying device on the described decoded speech, Therefore can dwindle the contained input voice that result from the output voice of voice communication terminal The amplitude of part, in voice communication terminal, just can avoid like this because excessive the returning of time-delay Sound and cause the problem that to converse.
And, have the output phonetic code and import similar that the similarity of phonetic code assesses The degree apparatus for evaluating is as its similarity, owing to adopt between the phonetic code that becomes comparison other Same number of bits or identical bit, so even have on the transmission line a small amount of wrong overlapping, Whether the phonetic code B that exports of also can detecting over returns as input phonetic code A ', So just can make voice communication terminal avoid taking place owing to leading to that bigger time-delay echo is caused The words difficulty.
In the voice communication terminal of the present invention, phonetic code is by the expression information code of voice messaging and its Error-correcting code consists of, as the similarity of similarity assessment device, to the letter of output phonetic code Breath code and according to the error-correcting code in the same input phonetic code to the information code in the input phonetic code Carry out after the error correction after the error correction relation between the information code and assess, even therefore Have a small amount of wrong overlappingly on the transmission line, also can correctly detect over the voice of exporting Whether code returns as the input phonetic code, so just can make voice communication terminal avoid generation by The conversation difficulty that causes in bigger time-delay echo.
Voice communication terminal of the present invention is because can be to the tone decoding of exporting after the phonetic code decoding Each time frame in the device or the yield value of period of the day from 11 p.m. to 1 a.m frame are proofreaied and correct, so the Corrections Division of yield value Reason is just very simple, and this is its strong point.
The 1st voice communication terminal of described voice communication terminal of the present invention is from each terminal input language Sound and phonetic code, the voice processing of encoding to input makes it to become the output phonetic code, and And described phonetic code is decoded, and from each terminal output, it has and can dwindle with normal voice The amplitude of the contained part that results from the input voice in the output voice of voice communication terminal Means for correcting, the 2nd terminal can be from each terminal input voice and phonetic code, to the language of importing The sound processing of encoding, to the processing of decoding of described phonetic code, then logical as the output voice Cross each terminal output, so just can avoid the Quality Down that causes owing to 2 coding processing Problem, in voice communication terminal, just can avoid owing to the time-delay excessive echo cause nothing The problem of method conversation.
Because only to the caller A of the 1st voice communication terminal or leading to of the 2nd voice communication terminal Carried out the part of coding processing in the voice that words person B imports and decoded, so overlapping Under the effect of call voice coding, just can avoid the showing of Quality Down in the voice communication system Resemble.

Claims (11)

1. voice communication terminal, it carries out encoding process from each different terminals input voice and phonetic code to the input voice, make it to become the output phonetic code, and the described phonetic code of decoding, and, it is characterized in that: comprise as exporting voice respectively from each terminal output
Sound encoding device, it carries out encoding process to the input voice, and resulting phonetic code is exported as the output phonetic code;
Audio decoding apparatus, it is decoded to the input phonetic code, and resulting decoded speech is exported as the output voice;
Storage device, it is stored by certain quantity the output phonetic code of encoding at described sound encoding device;
The similarity assessment device, it is assessed the output phonetic code stored in this storage device and the similarity of described input phonetic code, and resulting similarity is exported;
Means for correcting, it is according to the similarity of described similarity assessment device, operation to described input phonetic code or described decoded speech or described audio decoding apparatus is proofreaied and correct, thereby dwindles the amplitude of part that results from described input voice that is contained in the output voice by the output of described terminal.
2. the voice communication terminal of claim 1 record, its feature also is:
Described means for correcting comprises
The yield value control device, it is based on the similarity of described similarity assessment device, and the yield value on the described decoded speech is taken in decision;
The multiplying device, the yield value that its yield value control device is exported is taken on the described decoded speech, and resulting result is exported as the output voice signal.
3. the voice communication terminal of claim 1 record, its feature also is:
Described means for correcting comprises
Judgment means, it is based on the similarity of described similarity assessment device, and whether decision proofreaies and correct described input phonetic code;
The code displacement apparatus, after it made the decision of proofreading and correct when judgment means, output had been replaced into described input phonetic code the phonetic code of code designation, if judgment means is made the decision of not carrying out proofreading and correct, then made described input phonetic code former state output,
Described audio decoding apparatus is decoded to the phonetic code that the code displacement apparatus is exported, and resulting decoded speech is exported as the output phonetic code.
4. the voice communication terminal of claim 3 record, its feature also is:
The phonetic code that described code displacement apparatus is replaced is the fixing phonetic code of the decoded speech of decodable code short arc.
5. the voice communication terminal of claim 3 record, its feature also is:
Phonetic code is made of the expression information code of voice messaging and its error-correcting code, described code displacement apparatus, make the decision of correction when judgment means after, replace described audio decoding apparatus and judge and surpass the wrong overlapping phonetic code that error-correcting code is proofreaied and correct boundary,
Described audio decoding apparatus detecting when surpass proofreading and correct boundary wrong overlapping, is decoded according to the principle that the amplitude that makes decoded speech reduces in proper order.
6. the voice communication terminal of claim 1 record, its feature also is:
Described means for correcting comprises
Judgment means, it is based on described similarity, makes the decision of whether operation of described audio decoding apparatus being proofreaied and correct;
Control device, after it makes the decision of proofreading and correct in judgment means, the signal that resulting yield value or error detection mark are proofreaied and correct in the decoding process of output to described audio decoding apparatus,
Described audio decoding apparatus according to the signal of described control device, is proofreaied and correct resulting yield value in the decode procedure or error detection mark, and is carried out decoding processing.
7. claim 1 is to the voice communication terminal of one of claim 6 record, and its feature also is:
Described phonetic code is made of the information code of representing voice messaging and its error-correcting code,
Described similarity assessment device, information code in the output phonetic code that described storage device is stored, and the information code in the described input phonetic code is carried out between the information code after the error correction similarity being assessed according to the error-correcting code in the same input phonetic code.
8. claim 1 is to the voice communication terminal of one of claim 6 record, and its feature also is:
Described similarity assessment device as similarity, utilizes same number of bits or identical bit between the phonetic code that becomes comparison other.
9. voice communication terminal, it carries out encoding process from each terminal input voice and phonetic code to the input voice, makes it to become the output phonetic code, and the described phonetic code of decoding, and as the output voice respectively from each terminal output, it is characterized in that: comprise
Sound encoding device, it carries out encoding process to the input voice, and resulting phonetic code is exported as the output phonetic code;
Audio decoding apparatus, phonetic code is imported in its decoding, and resulting decoded speech is exported as the output voice;
Storage device, it stores described input voice by certain length;
The similarity assessment device, it assesses the input voice stored in this storage device and the similarity of described decoded speech, and resulting similarity output;
The yield value control device, based on the similarity of described similarity assessment device, the yield value on the described decoded speech is taken in decision;
The multiplying device, the yield value that its yield value control device is exported is taken on the described decoded speech, and resulting result is exported as the output voice signal.
10. a voice communication system comprises
The 1st voice communication terminal has the structure that claim 1 is put down in writing to the arbitrary claim in the claim 6;
The 2nd voice communication terminal, it is at the input voice and utilize choice device to select it according to standard in the output phonetic code exported of by sound encoding device described voice are encoded by the time frame of each regulation phonetic code that obtains and described the 1st voice communication terminal and export in the lump, simultaneously, the output phonetic code that utilizes audio decoding apparatus that described the 1st voice communication terminal is exported is decoded and is exported.
11. a voice communication system comprises
The 1st voice communication terminal has the structure that claim 9 is put down in writing;
The 2nd voice communication terminal, it is imported voice and described voice is encoded by sound encoding device by the time frame of each regulation, utilizing choice device to select it according to standard in the output phonetic code that this phonetic code that is encoded and described the 1st voice communication terminal are exported exports in the lump, simultaneously, the output phonetic code that described the 1st voice communication terminal is exported is decoded by audio decoding apparatus and is exported.
CN01132642.5A 2001-02-06 2001-09-05 Voice communication terminal and voice communication system Expired - Fee Related CN1183734C (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
JP29757/2001 2001-02-06
JP2001029757A JP3942831B2 (en) 2001-02-06 2001-02-06 Voice communication terminal and voice communication system
JP29757/01 2001-02-06

Publications (2)

Publication Number Publication Date
CN1368820A true CN1368820A (en) 2002-09-11
CN1183734C CN1183734C (en) 2005-01-05

Family

ID=18894063

Family Applications (1)

Application Number Title Priority Date Filing Date
CN01132642.5A Expired - Fee Related CN1183734C (en) 2001-02-06 2001-09-05 Voice communication terminal and voice communication system

Country Status (3)

Country Link
JP (1) JP3942831B2 (en)
CN (1) CN1183734C (en)
TW (1) TW515190B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102158917A (en) * 2010-02-03 2011-08-17 通用电气公司 Handoffs between different voice encoder systems
CN102300240A (en) * 2011-08-26 2011-12-28 北京邮电大学 Output performance parameter-based method for evaluating similarity of two systems
CN104508718A (en) * 2012-08-01 2015-04-08 大陆汽车有限责任公司 Method for outputting information by means of synthetic speech
WO2020253054A1 (en) * 2019-06-20 2020-12-24 平安科技（深圳）有限公司 Method and apparatus for evaluating audio signal loss, and storage medium

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102158917A (en) * 2010-02-03 2011-08-17 通用电气公司 Handoffs between different voice encoder systems
CN102300240A (en) * 2011-08-26 2011-12-28 北京邮电大学 Output performance parameter-based method for evaluating similarity of two systems
CN104508718A (en) * 2012-08-01 2015-04-08 大陆汽车有限责任公司 Method for outputting information by means of synthetic speech
WO2020253054A1 (en) * 2019-06-20 2020-12-24 平安科技（深圳）有限公司 Method and apparatus for evaluating audio signal loss, and storage medium

Also Published As

Publication number Publication date
JP3942831B2 (en) 2007-07-11
JP2002229595A (en) 2002-08-16
CN1183734C (en) 2005-01-05
TW515190B (en) 2002-12-21

Similar Documents

Publication Publication Date Title
CN1138347C (en) Error correcting/decoding apparatus and error correcting/decoding method
CN1220177C (en) Audio decoder and coding error compensating method
CN1267891C (en) Voice communication system and method for processing drop-out fram
CN1172294C (en) Audio-frequency coding apapratus, method, decoding apparatus and audio-frequency decoding method
CN1201288C (en) Decoding method and equipment and program facility medium
CN1922660A (en) Communication device, signal encoding/decoding method
CN1526196A (en) Reduced soft output information packet selection
CN1254936C (en) Multicarrier CDMA communication device
CN101040535A (en) Two step arithmetic decoding with conversion into an intermediate format
CN1276669C (en) Receipt signal synthesis method and system, radio receiving station and receiving station
CN1341996A (en) Radio transmitting/receiving device and method, system and storage medium
CN1235348C (en) Sound communication device and echo processor
CN1158852C (en) Data communication device and data communication method
CN1671149A (en) Method of recovery from abnormal state of handsfree phone speech, recovery system, and handsfree phone
CN1617115A (en) Memory expansion pack for providing content to portable terminal
CN1183734C (en) Voice communication terminal and voice communication system
CN1697472A (en) Method and device for switching speeches
CN1164050C (en) Speech decording appts
CN1287347A (en) Voice coding apparatus and voice decoding apparatus
CN1525738A (en) Apparatus and program for image processing
CN101034956A (en) Pulse modulation decoding and coding method, device and system
CN1215663C (en) Speech sound transmitting device
CN1607759A (en) Method of decoding a data word
CN1375980A (en) Signal identifying device, signal identifying method and transmission device using said identifying device
CN1366412A (en) Sound code cut-over method and device and sound communication terminal

Legal Events

Date Code Title Description
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20050105

Termination date: 20160905