CN104485112A - Audio decoding method and audio decoding device based on audio communication - Google Patents
Audio decoding method and audio decoding device based on audio communication Download PDFInfo
- Publication number
- CN104485112A CN104485112A CN201410742992.5A CN201410742992A CN104485112A CN 104485112 A CN104485112 A CN 104485112A CN 201410742992 A CN201410742992 A CN 201410742992A CN 104485112 A CN104485112 A CN 104485112A
- Authority
- CN
- China
- Prior art keywords
- pulsewidth
- wide
- coding
- audio
- sequence
- 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
Links
Landscapes
- Noise Elimination (AREA)
- Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)
Abstract
The invention provides an audio decoding method and an audio decoding device based on audio communication. The method comprises the following steps: S100, receiving an original encoding waveform transmitted from a transmitting terminal by a receiving terminal; S200, carrying out filtering processing on the original encoding waveform by the receiving terminal, so as to obtain a first encoding waveform; S300, calculating the first encoding waveform, so as to obtain a pulse-width sequence; S400, converting the pulse-width sequence into a binary sequence, and correcting an error by using fuzzy logic when the error appears in conversion; S500, converting the binary sequence into an encoding result according to an encoding specification. According to the audio decoding method, the processes of filtering, calculating to obtain a pulse-width value, converting into the binary sequence and converting into the encoding result are sequentially carried out after the original encoding waveform is received; meanwhile, the problems of large noise effect, unadapted baud rate, pulse-width conversion error and the like frequently appearing in the decoding process are solved.
Description
Technical field
The present invention relates to the codec domain of audio volume control, specifically be a kind of based on the audio-frequency decoding method in voice communication and device thereof.
Background technology
Voice communication between current mobile phone or panel computer adopts a left side (right side) sound channel that the waveform of coding is mail to equipment mostly; The mode of carrying out decoding after the coding that receiving end is sended over by microphones equipment realizes.Concrete, mobile phone by microphones to be one section of discrete waveform data, Wave data may use the coding of F2F or other types, by supposing known current reception baud rate, then take out the point in waveform in order, add up higher than high threshold with lower than counting between high threshold, then add up and count lower than Low threshold with higher than between Low threshold, so repeatedly, Wave data is divided into pulsewidth one by one; Pulsewidth comparatively can be converted thereof into 01 logic with reference to peak pulse duration, then calculate coding result according to coding criterion; Equipment is by comparer and catch the data that timer receives left (right side) sound channel, and often carry out a pulsewidth, equipment just produces and once interrupts, equipment by pulsewidth and benchmark peak pulse duration compared with thus convert 01 logic to, then draw coding result according to coding criterion.
But, adopt aforesaid way to carry out voice communication, there is following shortcoming:
1, the mobile phone quantity on market and kind numerous, the audio scheme of different mobile phone application may have very big difference, and it is far short of what is expected that same input waveform records result, and the point adding up threshold value simply needs manually to adjust threshold value, and efficiency is low, and error is large.
2, the mobile phone had needs to use different baud rate, and statistics is counted to distinguish 01 logic and is only applicable to specific baud rate, and the just necessary update routine when could communicating with different baud rate from equipment, dirigibility is very low.
When 3, there is stronger noise, waveform distortions may be caused more severe, be difficult to get a fixing threshold value, noise intensity and real data may be very close, if there is crosstalk, the data that possible mobile phone sends all can be mixed in microphone, and existing method is more weak in anti-noise.
4, the actual pulsewidth obtained may the too short or long logical transition that causes be made mistakes, cannot according to the relation error correction between pulsewidth.
Therefore, be necessary to provide a kind of audio-frequency decoding method that can solve the problem very well and device.
Summary of the invention
Technical matters to be solved by this invention is: provide a kind of based on the audio-frequency decoding method in voice communication and device, the problem such as solve in audio decoding process that the waveform recording gap often occurred is large, baud rate inadaptable, noise jamming and pulse width conversion are made mistakes.
In order to solve the problems of the technologies described above, the technical solution used in the present invention is:
Based on the audio-frequency decoding method in voice communication, comprising:
S100: the original coding waveform that receiving end receiving end/sending end sends over;
S200: receiving end carries out filtering process to original coding waveform and obtains the first coding waveforms;
S300: calculate the first coding waveforms, obtains pulsewidth sequence;
S400: described pulsewidth sequence is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
S500: described binary sequence is converted to coding result according to coding criterion;
S300 comprises:
S301: be that described first coding waveforms is divided into lower regions by boundary with zero point line, get the X data point of closing on zero point line in region most, close on the Y data point of zero point line in lower area most;
S302: adopt linear interpolation method to calculate zero passage point value to described X data point and Y data point;
S303: calculate adjacent two zero passage point values, obtain pulsewidth.
Another technical scheme provided by the invention is:
Based on the audio decoding apparatus in voice communication, the voice communication applied between receiving end and transmitting terminal connects, and comprises receiver module, filtration module, computing module, the first modular converter and the second modular converter;
Described receiver module, for the original coding waveform that receiving end receiving end/sending end sends over;
Described filtration module, carries out filtering process for receiving end to original coding waveform and obtains the first coding waveforms;
Described computing module, for calculating the first coding waveforms, obtains a series of pwm value;
Described first modular converter, for described pwm value is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
Described second modular converter, for being converted to coding result by described binary sequence according to coding criterion;
Described computing module comprises chooses unit, linear interpolation computing unit and computing unit;
Describedly choosing unit, for being that described first coding waveforms is divided into lower regions by boundary with zero point line, getting the X data point closest to zero point line in region, closest to the Y data point of zero point line in lower area;
Described linear interpolation computing unit, for adopting linear interpolation method to calculate zero passage point value to described X data point and Y data point;
Described computing unit, for calculating adjacent two zero passage point values, obtains pulsewidth.
Beneficial effect of the present invention is: the problems such as the present invention is different from that the waveform recording gap existed based on the audio-frequency decoding method in voice communication and device of prior art is large, baud rate inadaptable, noise jamming and pulse width conversion are made mistakes.The invention provides a kind of based on the audio-frequency decoding method in voice communication and device, by first carrying out filtering to discrete original coding waveform, making filtered waveform closer to the input waveform of reality, thus reduce the impact of noise; By calculating pwm value accurately to the first coding waveforms, on the one hand, because pulsewidth is more accurate, therefore follow-up without the need to manually adjusting threshold values again by the high low valve valve than default setting accurately, reduces error, significantly improving the decoding efficiency of audio frequency simultaneously; On the other hand, obtain making subsequent calculations obtain benchmark pulsewidth after pwm value accurately and become feasible, thus realize receiving end and transmitting terminal can self-adaptation baud rate at audio decoding process, and no longer need the restriction of specific baud rate by Correspondent Node; Finally, by the process changed at pwm value, carry out intelligent correction by fuzzy logic, the relation between can be good at according to actual pulsewidth carries out error correction, ensures the accuracy in transfer process.A kind of audio-frequency decoding method provided by the invention and device, solve the problems such as the noise effect often occurred in decode procedure is large, baud rate inadaptable, pulse width conversion is made mistakes simultaneously, communication two party no longer causes by baud rate is not suitable with problem cannot normal communication, greatly improves accuracy rate and the decoding efficiency thereof of audio decoder result.
Accompanying drawing explanation
Fig. 1 is a kind of basic procedure block diagram based on the audio-frequency decoding method in voice communication of the present invention;
Fig. 2 is a kind of FB(flow block) based on the audio-frequency decoding method in voice communication of the present invention;
Fig. 3 is a kind of basic structure compositional block diagram based on the audio decoding apparatus in voice communication of the present invention;
Fig. 4 is a kind of structure compositional block diagram based on the audio decoding apparatus in voice communication of the present invention;
Fig. 5 is the hardware composition schematic diagram between equipment and intelligent terminal;
Fig. 6 is a kind of original coding waveform segment sample 1 that receiving end of the present invention receives;
Fig. 7 is a kind of original coding waveform segment sample 2 that receiving end of the present invention receives;
Fig. 8 is the filtered result of original coding waveform segment sample 1 of the present invention;
Fig. 9 is the filtered result of original coding waveform segment sample 2 of the present invention;
Figure 10 is the present invention first coding waveforms fragment;
Figure 11 is the schematic diagram adopting linear interpolation method in the present invention;
To be that the present invention is a kind of be all mistaken for the wide waveform segment of long pulse based on two short pulse durations in the audio-frequency decoding method in voice communication to Figure 12;
Figure 13 is a kind of waveform segment based on adopting fuzzy error correction LSL to be converted to LSS in the audio-frequency decoding method in voice communication of the present invention;
Figure 14 is the Data Comparison oscillogram after the process of the present invention one specific embodiment.
Label declaration:
3, receiver module; 4, filtration module; 5, computing module;
6, the first modular converter; 7, the second modular converter; 8, unit is chosen;
9, linear interpolation computing unit; 10, computing unit; 11, the first determining unit;
12, the second determining unit; 13, fuzzy logic error correction unit; 14, converting unit.
Embodiment
By describing technology contents of the present invention in detail, realized object and effect, accompanying drawing is coordinated to be explained below in conjunction with embodiment.
The design of most critical of the present invention is: carry out filtering successively after receiving original coding waveform, calculate pwm value, be converted to binary sequence and transfer to for coding result process, solves the problems such as the noise effect often occurred in decode procedure is large, baud rate inadaptable, pulse width conversion is made mistakes simultaneously.
Please refer to Fig. 1 to Figure 14, the invention provides a kind of based on the audio-frequency decoding method in voice communication, comprising:
S100: the original coding waveform that receiving end receiving end/sending end sends over;
S200: receiving end carries out filtering process to original coding waveform and obtains the first coding waveforms;
S300: calculate the first coding waveforms, obtains pulsewidth sequence;
S400: described pulsewidth sequence is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
S500: described binary sequence is converted to coding result according to coding criterion;
S300 comprises:
S301: be that described first coding waveforms is divided into lower regions by boundary with zero point line, get the X data point of closing on zero point line in region most, close on the Y data point of zero point line in lower area most;
S302: adopt linear interpolation method to calculate zero passage point value to described X data point and Y data point;
S303: calculate adjacent two zero passage point values, obtain pulsewidth.
It should be noted that, the original coding waveform received can well improve waveform after filtering process, contributes to follow-up process, therefore, first carries out filtering process to original coding waveform.
The first coding waveforms obtained after process after filtering remains some discrete datas, usual data point can not just drop in zero point line, as shown in Figure 10, medium line is zero point line, solid dot is the data of actual storage, in order to calculate pwm value, needs first to calculate the first discrete coding waveforms, find zero crossing, then obtain pulsewidth.After accurately the first coding waveforms being converted to pulsewidth sequence, need by pwm value preferably according to the coded system of F2F, each pwm value or be converted into logical zero, or become logical one with next pulse width conversion, be all converted to binary sequence.And use fuzzy logic method to carry out error correction when changing and occurring mistake, improve the accuracy of conversion.Finally, binary sequence is converted to character according to coding rule, obtains coding result.
As shown in Figure 10, be filtered first coding waveforms, as we know from the figure, Wave data is some discrete datas, and usual data point all can not just drop in zero point line, therefore, needs first to carry out interpolation calculation to discrete data, finds the value of zero crossing.Concrete, be worth at zero point of known reality between two data points in circular arc ring position in the drawings, and be fixing according to the sampling rate of microphone, therefore the interval between each data point is fixing, be set to T, and two the upper and lower data points in circular arc ring are respectively y1, y2, can according to t0=y1/ (y1-y2) * T, specifically consult Figure 11, the distance t0 of zero point from previous data point can be calculated, known in conjunction with Figure 10, pulsewidth A contains t0 part, and pulsewidth B contains T-t0 part, pass through the method, just the first coding waveforms can accurately be converted to pulsewidth sequence.
From foregoing description, beneficial effect of the present invention is: the invention provides a kind of based on the audio-frequency decoding method in voice communication, by first carrying out filtering to discrete original coding waveform, make filtered waveform closer to the input waveform of reality, thus reduce the impact of noise; By calculating pwm value accurately to the first coding waveforms, on the one hand, because pulsewidth is more accurate, therefore follow-up without the need to manually adjusting threshold values again by the high low valve valve than default setting accurately, reduces error, significantly improving the decoding efficiency of audio frequency simultaneously; On the other hand, obtain making subsequent calculations obtain benchmark pulsewidth after pwm value accurately and become feasible, thus realize receiving end and transmitting terminal can self-adaptation baud rate at audio decoding process, and no longer need the restriction of specific baud rate by Correspondent Node; Finally, by the process changed at pwm value, carry out intelligent correction by fuzzy logic, the relation between can be good at according to actual pulsewidth carries out error correction, ensures the accuracy in transfer process.A kind of audio-frequency decoding method provided by the invention, solve the problems such as the noise effect often occurred in decode procedure is large, baud rate inadaptable, pulse width conversion is made mistakes simultaneously, communication two party no longer causes by baud rate is not suitable with problem cannot normal communication, greatly improves accuracy rate and the decoding efficiency thereof of audio decoder result.
Further, in step S200, receiving end adopts auto adapted filtering mode to carry out filtering process to original coding waveform.
Seen from the above description, auto adapted filtering mode is adopted to process to original coding waveform, it is the quality based on getting the straightforward procedure such as average, weighted mean and can reduce normal waveform, and for the waveform be out of shape, if departing from that normal point is too large also cannot normally filtering, therefore adaptive filter method must be adopted, preferably, use least square method filtering, the original waveform fragment that the original waveform fragment that Fig. 6 is sample 1, Fig. 7 are sample 2, carrying out after adaptive filter method carries out filtering process, obtaining corresponding Fig. 8 and Fig. 9 result respectively.
Further, step S400 comprises:
S401: get the leading pulse width part in described pulsewidth sequence according to the coded system of transmitting terminal, determines benchmark pulsewidth from described leading pulsewidth;
S402: described pulsewidth and described benchmark pulsewidth are compared, determines the type of pulsewidth;
S403: the type conversion according to pulsewidth is corresponding binary sequence.
From the above, when pwm value is converted to 01 sequence according to coded system, be divided into two parts.Part I is the leading pulse width part chosen from the coded system made in advance according to transmitting terminal pwm value, and leads in pulse width part in the past and determine benchmark pulsewidth.The data that preferred transmitting terminal sends over are divided into according to type: leading short pulse duration+real data+after to lead long pulse wide, the existence of leading short pulse duration is that mobile phone Mike often understands attenuates high frequency signals on the one hand, if do not have the existence of leading short pulse duration, real data may be attenuated and cause decoding unsuccessfully; On the other hand, be self-adaptation baud rate, need to determine benchmark pulsewidth according to lead data.Specifically, when taking out 10 several pulsewidths continuously, their numerical value is close, thinks to have found benchmark pulsewidth, and pulsewidth below should be made comparisons with this benchmark pulsewidth.In fact, the noise ratio of a lot of mobile phone is comparatively strong, and being mistakenly considered is pulsewidth, and therefrom also can take out the close pulsewidth of continuous 10 numerical value.
Part II, compares the described pulsewidth drawn in the benchmark pulsewidth determined and previous step, determines the type of described pulsewidth, is that long pulse is wide, short pulse duration is also non-pulsewidth etc., and is converted to corresponding binary sequence.Preferably, for F2F coded system, pulsewidth is divided into long pulse is wide, short pulse duration, burr, non-pulsewidth Four types, represent with L, S, P, N symbol.Suppose that benchmark pulsewidth is B, when pulsewidth compared with B [0.675,1.35) be L, [0.2,0.675) for S, [0,0.2) for P, [1.35 ,+infite) be N, here the setting of value range is a citing, in fact, the setting of this scope is more random, because a perfect boundary value in fact can not be found to distinguish the type of a pulsewidth.After determining the type of pulsewidth, be converted to 01 corresponding sequence, wherein single L converts logical zero to, and 2 continuous print S convert logical one to.
Further, step S400 also comprises:
S404: set the ratio of pulsewidth and benchmark pulsewidth as fac, long pulse wide for L, short pulse duration be S, then fac degree of membership when (1.5 ,+∞) is 100$, [0,1.2) time, degree of membership is 0, and other situation degree of membership are 1-(1.5-fac)/0.3;
S405: when pulse width conversion is binary sequence appearance mistake, adopt fuzzy logic to carry out error correction, concrete:
When the degree of membership of the ratio fac of the pulsewidth length after the wide LL of two long pulses are added and benchmark pulsewidth is less than 100$, convert two wide LL of long pulse to two short pulse duration SS;
When there is LSL situation, analyze further, a wide L of long pulse after the wide L of previous long pulse is less than, and the degree of membership of the ratio fac of the pulsewidth length that adds of the wide L of previous long pulse and short pulse duration S-phase and benchmark pulsewidth is between 90$ to 100$, then transfer SSL to; A wide L of long pulse after the wide L of previous long pulse is greater than, and a rear wide L of long pulse and short pulse duration S-phase add after pulsewidth length and the degree of membership of ratio fac of benchmark pulsewidth between 90$ to 100$, then transfer LSS to; Otherwise directly transfer LLL to;
When occurring that pulsewidth sequence LSS (2m-1) SL is total to 2m+1 S, further analysis, when previous long pulse wide L and first short pulse duration S sum is less than last short pulse duration S and the wide L sum of last long pulse, and the degree of membership of the ratio of first long pulse wide L and first short pulse duration S sum and benchmark pulsewidth is between 90$ to 100$, then transfer SSS (2m-1) SL to; Otherwise transfer LLS (2m-1) SL to; When a previous long pulse wide L and first short pulse duration sum be greater than or etc. understand last short pulse duration S and after a wide L sum of long pulse, and the degree of membership of the ratio of a rear wide L of long pulse and last short pulse duration S sum and benchmark pulsewidth B is less than between 90$ to 100$, then transfer LSS (2m-1) SS to; Otherwise convert LSS (2m-1) LL to.
From the above, be in the process of binary sequence, probably occur transcription error, such as occur odd number short pulse duration in pulse width conversion, two short pulse durations are mistaken for the situations such as long pulse is wide, at this time, just need to adopt fuzzy logic to solve the situation of transcription error.Concrete, the situation of transforming mistakes can be divided into 3 kinds:
1, two all misjudged long pulses of short pulse duration are wide, convert SS to by LL.
Consult shown in Figure 12, under normal circumstances, the total length of two wide additions of long pulse is compared with benchmark pulsewidth B, and the degree of membership of pulsewidth normally should close to 100$, if not, illustrate that these two long pulses are wide and be in fact likely two longer short pulse durations.If when therefore angelica degree is less than 100$, convert LL to SS.
2, there is single short pulse duration, namely occur the situation of LSL, in fact should be converted to any one situation of SSL, LSS or LLL.
As shown in figure 13, specifically convert which kind of situation to, need these 3 Pulse Width Analysis, a L after previous L is less than, and the total length that previous L and S-phase add is compared with B, the degree of membership of non-pulsewidth is less than 95$ (between 90$ to 100$), just converts SSL to; Otherwise a L after previous L is greater than, and a rear L adds total length compared with B with S-phase, pulsewidth degree of membership is less than 95$ (between 90$ to 100$), just changes into LSS, otherwise converts LLL to.Figure 11 is example LSL being converted to LSS.
3, there is odd number (>=3) short pulse duration continuously, namely pulsewidth sequence LSS (2m-1) SL 2m+1 S is altogether had, so have 1 S to match, actual capabilities change into LSS (2m-1) SS, SSS (2m-1) SL, LLS (2m-1) SL and LSS (2m-1) LL.
If front 1 L and the 1st S pulsewidth sum is less than last 1 S and rear 1 L pulsewidth sum, then may change into SSS (2m-1) SL, LLS (2m-1) SL two kinds of situations.Now, if the 1st L with the 1st S pulsewidth sum compared with B, pulsewidth degree of membership is less than 95$, thinks and should change into SSS (2m-1) SL, otherwise converts LLS (2m-1) SL to.
If front 1 L and the 1st S pulsewidth sum is more than or equal to last 1 S and rear 1 L pulsewidth sum, then may change into LSS (2m-1) SS, LSS (2m-1) LL two kinds of situations, now, if rear 1 L and last 1 S pulsewidth sum are compared with B, non-pulsewidth degree of membership is less than 95$, think and should change into LSS (2m-1) SS, otherwise convert LSS (2m-1) LL to.
It should be noted that, binary sequence is being converted in the process of coded character, preferred each character 11 codings, form is: start bit (1)+real data (8)+even parity check (1)+position of rest (1).If run into mistake, such as certain character does not detect start bit or inspection dislocation, or does not have position of rest, then failure of this time decoding, and can only require that equipment retransmits for the failed situation of decoding.If not there is mistake, then this successfully decoded, flow process terminates.
And the decode procedure of equipment end comprises: equipment end obtains the just process of a pulsewidth, and what it in fact just only had receiving end to decode is converted to binary sequence and scale-of-two is transferred to the step of coding result at every turn.Processing mode has outside some nuances, and other are all the same.Because equipment end exports low and high level by comparer on hardware, filtering function own is not strong, when input signal have comparatively shake time, equipment end may receive burr, therefore equipment end can not using burr P as termination condition, on the contrary, to run in N or certain hour (such as 15ms), without interrupting, as termination condition, (this is necessary to equipment end, under normal circumstances, if valid data finish just not triggered interrupts, then equipment does not receive N always and thinks and do not terminate).Can not be ignored when equipment end receives burr P, certain process must be done.The array mode of L, S, P tri-kinds of pulsewidths is very complicated, in fact need not enumerate all situations, because original waveform is not inherently very feasible in will inferring from result waveform when there is burr P, therefore only considers the simplest situation here.When running into burr P, burr P and two, left and right pulsewidth are synthesized one.
Further, the original coding waveform that transmitting terminal sends in the step s 100 comprises identification character.
Preferably, in the data type that transmitting terminal sends over, also comprise identification character, identification character also known as Magic number, i.e. leading short pulse duration+Magic number+real data+after to lead long pulse wide; In the environment that noise ratio is stronger, noise is easy to be thought pulsewidth by mistake, and therefrom also can take out the close pulsewidth of continuous 10 numerical value.For coming from noise difference, after lead data, adding a Magic number, (such as (00001001) b), only just think there is real data when the sequence converting logical zero 1 at first to equals Magic number, otherwise just think that these pulsewidths are false datas, again attempt Calculation Basis pulsewidth from a rear pulsewidth.After finding Magic number, by pulsewidth compared with benchmark pulsewidth to convert 01 sequence to, only having the wide L of long pulse and short pulse duration S to be normal pulsewidth, when running into the pulsewidth of burr P or non-pulsewidth N, thinking EOC.Insert Magic number in the data, data-crosstalk that left/right sound channel sends can be prevented to MIC, had useful signal by handset identity.
Another technical scheme provided by the invention is:
Based on the audio decoding apparatus in voice communication, the voice communication applied between receiving end and transmitting terminal connects, and comprises receiver module 3, filtration module 4, computing module 5, first modular converter 6 and the second modular converter 7;
Described receiver module 3, for the original coding waveform that receiving end receiving end/sending end sends over;
Described filtration module 4, carries out filtering process for receiving end to original coding waveform and obtains the first coding waveforms;
Described computing module 5, for calculating the first coding waveforms, obtains pulsewidth sequence;
Described first modular converter 6, for described pulsewidth sequence is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
Described second modular converter 7, for being converted to coding result by described binary sequence according to coding criterion;
Described computing module 5 comprises chooses unit 8, linear interpolation computing unit 9 and computing unit 10;
Describedly choosing unit 8, for being that described first coding waveforms is divided into lower regions by boundary with zero point line, getting the X data point closest to zero point line in region, closest to the Y data point of zero point line in lower area;
Described linear interpolation computing unit 9, for adopting linear interpolation method to calculate zero passage point value to described X data point and Y data point;
Described computing unit 10, for calculating adjacent two zero passage point values, obtains pulsewidth.
From foregoing description, beneficial effect of the present invention is: the invention provides a kind of based on the audio decoding apparatus in voice communication, first filtering is carried out to discrete original coding waveform by filtration module 4, make filtered waveform closer to the input waveform of reality, thus reduce the impact of noise; Pwm value is accurately calculated by computing module 5 first coding waveforms, on the one hand, because pulsewidth is more accurate, therefore follow-up without the need to manually adjusting threshold values again by the high low valve valve than default setting accurately, reduce error, significantly improve the decoding efficiency of audio frequency simultaneously; On the other hand, obtain making subsequent calculations obtain benchmark pulsewidth after pwm value accurately and become feasible, thus realize receiving end and transmitting terminal can self-adaptation baud rate at audio decoding process, and no longer need the restriction of specific baud rate by Correspondent Node; Finally, in the process changed at pwm value, carry out intelligent correction by fuzzy logic by the first modular converter 6, the relation between can be good at according to actual pulsewidth carries out error correction, ensures the accuracy in transfer process.Provided by the invention a kind of based on the audio decoding apparatus in voice communication, solve the problems such as the noise effect often occurred in decode procedure is large, baud rate inadaptable, pulse width conversion is made mistakes simultaneously, communication two party no longer causes by baud rate is not suitable with problem cannot normal communication, greatly improves accuracy rate and the decoding efficiency thereof of audio decoder result.
Further, described filtration module 4 is sef-adapting filter.
Further, described first modular converter 6 comprises the first determining unit 11, second determining unit 12, fuzzy logic error correction unit 13 and converting unit 14;
Described first determining unit 11, for getting the leading pulse width part in described pulsewidth sequence according to the coded system of transmitting terminal, determines benchmark pulsewidth from described leading pulsewidth;
Described second determining unit 12, for described pulsewidth and described benchmark pulsewidth being compared, determines the type of pulsewidth;
Described fuzzy logic error correction unit 13, when being binary sequence appearance mistake for pulse width conversion, adopts fuzzy logic to carry out error correction;
Described converting unit 14, for according to the type conversion of pulsewidth being corresponding binary sequence.
Please refer to Figure 14, embodiments of the invention one are:
S101: equipment receives the data that left (right side) sound channel sends, and data are sent to microphone;
Equipment receives the data that L channel sends, and is character 0x0f after decoding.Microphone is sent to again by after this character code.
S102: mobile phone (panel computer etc.) passes through left (right side) sound channel toward equipment sending data, the data sended over by microphones equipment.
In sum, provided by the invention a kind of based on the audio-frequency decoding method in voice communication and device, not only improve anti-noise ability; And pulsewidth accurately can be obtained, realizing self-adaptation baud rate between equipment, simultaneously without the need to manually adjusting threshold values again, significantly improving audio decoder efficiency; Further, pulse width conversion can adopt fuzzy logic error correction when makeing mistakes, further improve the accuracy of audio decoder result.
The foregoing is only embodiments of the invention; not thereby the scope of the claims of the present invention is limited; every equivalents utilizing instructions of the present invention and accompanying drawing content to do, or be directly or indirectly used in relevant technical field, be all in like manner included in scope of patent protection of the present invention.
Claims (8)
1. based on the audio-frequency decoding method in voice communication, it is characterized in that, comprising:
S100: the original coding waveform that receiving end receiving end/sending end sends over;
S200: receiving end carries out filtering process to original coding waveform and obtains the first coding waveforms;
S300: calculate the first coding waveforms, obtains pulsewidth sequence;
S400: described pulsewidth sequence is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
S500: described binary sequence is converted to coding result according to coding criterion;
S300 comprises:
S301: be that described first coding waveforms is divided into lower regions by boundary with zero point line, get the X data point of closing on zero point line in region most, close on the Y data point of zero point line in lower area most;
S302: adopt linear interpolation method to calculate zero passage point value to described X data point and Y data point;
S303: calculate adjacent two zero passage point values, obtain pulsewidth.
2. according to claim 1ly a kind ofly to it is characterized in that based on the audio-frequency decoding method in voice communication, in step S200, receiving end adopts auto adapted filtering mode to carry out filtering process to original coding waveform.
3. according to claim 1ly a kind ofly to it is characterized in that based on the audio-frequency decoding method in voice communication, step S400 comprises:
S401: get the leading pulse width part in described pulsewidth sequence according to the coded system of transmitting terminal, determines benchmark pulsewidth from described leading pulsewidth;
S402: described pulsewidth and described benchmark pulsewidth are compared, determines the type of pulsewidth;
S403: the type conversion according to pulsewidth is corresponding binary sequence.
4. according to claim 3ly a kind ofly to it is characterized in that based on the audio-frequency decoding method in voice communication, step S400 also comprises:
S404: set the ratio of pulsewidth and benchmark pulsewidth as fac, long pulse wide for L, short pulse duration be S, then fac degree of membership when (1.5 ,+∞) is 100%, [0,1.2) time, degree of membership is 0, and other situation degree of membership are 1-(1.5-fac)/0.3;
S405: when described pulsewidth sequence is converted to binary sequence appearance mistake, adopt fuzzy logic to carry out error correction, concrete:
When the degree of membership of the ratio fac of the pulsewidth length after the wide LL of two long pulses are added and benchmark pulsewidth is less than 100%, convert two wide LL of long pulse to two short pulse duration SS;
When there is LSL situation, analyze further, a wide L of long pulse after the wide L of previous long pulse is less than, and the degree of membership of the ratio fac of the pulsewidth length that adds of the wide L of previous long pulse and short pulse duration S-phase and benchmark pulsewidth is between 90% to 100%, then transfer SSL to; A wide L of long pulse after the wide L of previous long pulse is greater than, and a rear wide L of long pulse and short pulse duration S-phase add after pulsewidth length and the degree of membership of ratio fac of benchmark pulsewidth between 90% to 100%, then transfer LSS to; Otherwise directly transfer LLL to;
When occurring that pulsewidth sequence LSS (2m-1) SL is total to 2m+1 S, further analysis, when previous long pulse wide L and first short pulse duration S sum is less than last short pulse duration S and the wide L sum of last long pulse, and the degree of membership of the ratio of first long pulse wide L and first short pulse duration S sum and benchmark pulsewidth is between 90% to 100%, then transfer SSS (2m-1) SL to; Otherwise transfer LLS (2m-1) SL to; When a previous long pulse wide L and first short pulse duration sum be greater than or etc. understand last short pulse duration S and after a wide L sum of long pulse, and the degree of membership of the ratio of a rear wide L of long pulse and last short pulse duration S sum and benchmark pulsewidth is less than between 90% to 100%, then transfer LSS (2m-1) SS to; Otherwise convert LSS (2m-1) LL to.
5. according to claim 1ly a kind ofly to it is characterized in that based on the audio-frequency decoding method in voice communication, the original coding waveform that transmitting terminal sends in the step s 100 comprises identification character.
6., based on the audio decoding apparatus in voice communication, the voice communication applied between receiving end and transmitting terminal connects, and it is characterized in that, comprises receiver module, filtration module, computing module, the first modular converter and the second modular converter;
Described receiver module, for the original coding waveform that receiving end receiving end/sending end sends over;
Described filtration module, carries out filtering process for receiving end to original coding waveform and obtains the first coding waveforms;
Described computing module, for calculating the first coding waveforms, obtains pulsewidth sequence;
Described first modular converter, for described pulsewidth sequence is converted to binary sequence, uses fuzzy logic to carry out error correction when transforming mistakes;
Described second modular converter, for being converted to coding result by described binary sequence according to coding criterion;
Described computing module comprises chooses unit, linear interpolation computing unit and computing unit;
Describedly choosing unit, for being that described first coding waveforms is divided into lower regions by boundary with zero point line, getting the X data point closest to zero point line in region, closest to the Y data point of zero point line in lower area;
Described linear interpolation computing unit, for adopting linear interpolation method to calculate zero passage point value to described X data point and Y data point;
Described computing unit, for calculating adjacent two zero passage point values, obtains pulsewidth.
7. according to claim 6ly a kind ofly to it is characterized in that based on the audio decoding apparatus in voice communication, described filtration module is sef-adapting filter.
8. according to claim 6ly a kind ofly to it is characterized in that based on the audio decoding apparatus in voice communication, described first modular converter comprises the first determining unit, the second determining unit, fuzzy logic error correction unit and converting unit;
Described first determining unit, for getting the leading pulse width part in described pulsewidth sequence according to the coded system of transmitting terminal, determines benchmark pulsewidth from described leading pulsewidth;
Described second determining unit, for described pulsewidth and described benchmark pulsewidth being compared, determines the type of pulsewidth;
Described fuzzy logic error correction unit, during for being binary sequence appearance mistake by described pulse width conversion, adopts fuzzy logic to carry out error correction;
Described converting unit, for according to the type conversion of pulsewidth being corresponding binary sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410742992.5A CN104485112B (en) | 2014-12-08 | 2014-12-08 | A kind of audio-frequency decoding method and its device based in voice communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410742992.5A CN104485112B (en) | 2014-12-08 | 2014-12-08 | A kind of audio-frequency decoding method and its device based in voice communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104485112A true CN104485112A (en) | 2015-04-01 |
| CN104485112B CN104485112B (en) | 2017-12-08 |
Family
ID=52759651
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410742992.5A Active CN104485112B (en) | 2014-12-08 | 2014-12-08 | A kind of audio-frequency decoding method and its device based in voice communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104485112B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI690920B (en) * | 2018-01-10 | 2020-04-11 | 盛微先進科技股份有限公司 | Audio processing method, audio processing device, and non-transitory computer-readable medium for audio processing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1502189A (en) * | 2000-05-17 | 2004-06-02 | ����˹�Ƽ��о�����˾ | Octave pulse data method and apparatus |
| CN1922660A (en) * | 2004-02-24 | 2007-02-28 | 松下电器产业株式会社 | Communication device, signal encoding/decoding method |
| CN101247183A (en) * | 2007-02-15 | 2008-08-20 | 凌阳科技股份有限公司 | Transmission device, receiving device, system and method for implementing communication based on ultrasonic wave |
| CN102128975A (en) * | 2010-12-22 | 2011-07-20 | 四川省电力公司 | Voltage stabilization online monitoring phasor data measurement device and phasor measurement method |
| EP2511907A1 (en) * | 2009-03-06 | 2012-10-17 | NTT DoCoMo, Inc. | Audio signal encoding method, audio signal decoding method, encoding device, decoding device, audio signal processing system, audio signal encoding program, and audio signal decoding program |
| CN103312419A (en) * | 2012-03-07 | 2013-09-18 | 苏州汉朗光电有限公司 | Method and system for data communication based on audio or ultrasonic |
-
2014
- 2014-12-08 CN CN201410742992.5A patent/CN104485112B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1502189A (en) * | 2000-05-17 | 2004-06-02 | ����˹�Ƽ��о�����˾ | Octave pulse data method and apparatus |
| CN1922660A (en) * | 2004-02-24 | 2007-02-28 | 松下电器产业株式会社 | Communication device, signal encoding/decoding method |
| CN101247183A (en) * | 2007-02-15 | 2008-08-20 | 凌阳科技股份有限公司 | Transmission device, receiving device, system and method for implementing communication based on ultrasonic wave |
| EP2511907A1 (en) * | 2009-03-06 | 2012-10-17 | NTT DoCoMo, Inc. | Audio signal encoding method, audio signal decoding method, encoding device, decoding device, audio signal processing system, audio signal encoding program, and audio signal decoding program |
| CN102128975A (en) * | 2010-12-22 | 2011-07-20 | 四川省电力公司 | Voltage stabilization online monitoring phasor data measurement device and phasor measurement method |
| CN103312419A (en) * | 2012-03-07 | 2013-09-18 | 苏州汉朗光电有限公司 | Method and system for data communication based on audio or ultrasonic |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI690920B (en) * | 2018-01-10 | 2020-04-11 | 盛微先進科技股份有限公司 | Audio processing method, audio processing device, and non-transitory computer-readable medium for audio processing |
| US10650834B2 (en) | 2018-01-10 | 2020-05-12 | Savitech Corp. | Audio processing method and non-transitory computer readable medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104485112B (en) | 2017-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JPS62169597A (en) | Dial pulse detection | |
| CN110011942B (en) | Morse message intelligent detection and identification method based on deep learning | |
| WO2005062469A1 (en) | Noise-tolerant signaling schemes supporting simplified timing and data recovery | |
| US20120027148A1 (en) | System and method for synchronization tracking in an in-band modem | |
| CN109861932B (en) | Short wave Morse message automatic identification method based on intelligent image analysis | |
| JPH08107429A (en) | Variable multithreshold detector and method for detecting a plurality of bits in baseband signal sampled using variable multithreshold | |
| WO2020029023A1 (en) | Baud rate calibration circuit and serial chip | |
| MX2010013332A (en) | System and method of an in-band modem for data communications over digital wireless communication networks. | |
| CN101989430A (en) | Audio mixing processing system and audio mixing processing method | |
| CN109412703B (en) | Time delay difference coding method utilizing short pulse short-time energy spectrum timing | |
| CN107515836B (en) | BMC decoding device and method with dynamic double thresholds | |
| CN108512552A (en) | A kind of the decoding system and coding/decoding method of Manchester code | |
| CN104485112A (en) | Audio decoding method and audio decoding device based on audio communication | |
| CN103379397A (en) | Data transmission method, data processing method, and signal processing method and module based on audio interface | |
| CN107004418B (en) | Digital communication method and device based on audio channel | |
| US4627073A (en) | Binary data transmission method | |
| CN109302239B (en) | Anti-noise and anti-distortion near field acoustic communication method | |
| US10616015B2 (en) | Signal processing system and method, and apparatus | |
| CN101543031B (en) | Bivariate histogram for impulse noise monitor | |
| US20060155816A1 (en) | Data transmission system and data transmission device | |
| WO2008038388A1 (en) | Signal receiver apparatus and waveform shaping method | |
| US20070100611A1 (en) | Speech codec apparatus with spike reduction | |
| CN211018863U (en) | Decoding system of BMC (baseboard management controller) code | |
| CN211406003U (en) | Decoding system of Manchester code | |
| CN109802761B (en) | Fault identification method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |