CN101287029A - Method and apparatus for automatically respond to detection - Google Patents
Method and apparatus for automatically respond to detection Download PDFInfo
- Publication number
- CN101287029A CN101287029A CNA2007100981197A CN200710098119A CN101287029A CN 101287029 A CN101287029 A CN 101287029A CN A2007100981197 A CNA2007100981197 A CN A2007100981197A CN 200710098119 A CN200710098119 A CN 200710098119A CN 101287029 A CN101287029 A CN 101287029A
- Authority
- CN
- China
- Prior art keywords
- threshold value
- timing
- module
- energy
- signal
- 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.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
- H04M3/51—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing
- H04M3/5158—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing in combination with automated outdialling systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M2203/00—Aspects of automatic or semi-automatic exchanges
- H04M2203/20—Aspects of automatic or semi-automatic exchanges related to features of supplementary services
- H04M2203/2027—Live party detection
Abstract
The invention provides a method and a device for automatically detecting answering, and in particular the method provided by the invention comprises the steps that: the energy of an input signal is detected; when a voice signal is detected, the timing of the voice signal is started; when pause of the voice signal is detected for the first time, the timing of the pause interval is started; if the pause interval reaches the maximum pause threshold value, the voice signal is determined to be the answering by the real person; otherwise, the timing of the pause interval is restarted on the sequent detection of the pause of the voice signal; when the pause interval in the restarted timing reaches the minimum pause threshold value, the voice signal is determined to be the answering by the real person; if the answering by the real person is not detected when the timing of the voice signal reaches the automatic answering threshold value, the voice signal is determined to be the answering by automatic answering equipment; the automatic answering threshold value is bigger than the maximum pause threshold value. By adopting the proposal of the invention, when the first pause detection fails, the minimum pause threshold value can be used for continuing to detect the sequent pause interval, thereby effectively increasing the accuracy of the answering detecting.
Description
Technical field
The present invention relates to the voice signal detection technique, particularly relate to the method and apparatus that a kind of automatic-answering back device detects.
Background technology
In the practical application, the call center may need to transfer to a large amount of telephone numbers in a period of time, promptly needs to realize a large amount of outbound call services.Such as: opinion poll, commercial propaganda, professional return visit or the like.
In traditional call business, the Outgoing Call number one by one of attending a banquet of call center is if access success then carries out communication on telephone with the user; If call failure is transferred to next telephone number again.According to statistics, about 50% is failed calls in traditional call business, as: ring unanswered, number error etc., this has not only wasted a large amount of resources, and has had a strong impact on the professional efficient of call center.
In order to overcome the defective of traditional call business, a kind of outer paging system of prediction type has been proposed, its principle roughly is: the background system of a call center is set, by the background system Outgoing Call, if access success is then given certain idle seating with lane assignment; If call failure then continues to transfer to other phones.Like this, the phone that background system just can be transferred to needs earlier screens, and only the phone of access success is distributed to and is attended a banquet, and the efficient of call center can be provided greatly.
In above-mentioned predictive outbound system, be a key technology of background system to the detection of call.Detection to call described here just to true man reply, the identification of call such as the replying of automatic answering equipment, ring-back tone, busy tone.Wherein, because replying of automatic answering equipment is to play the voice of recording in advance, easy and true man reply and obscure, and have influenced the accuracy rate that call detects.
In the practical application, people have the custom of answerphone usually, promptly can occur long dwell interval after greeting, and replying of automatic answering equipment then do not have long dwell interval usually.Fig. 1 and Fig. 2 are respectively that true man reply the oscillogram of replying with automatic answering equipment, and therefrom true man reply the difference of replying with automatic answering equipment as can be seen.According to above-mentioned principle, prior art proposes a kind of detection method of replying, and pause timer (PAUSE TIMER), delay timer (DELAY TIMER) and greeting timer (GREETING TIMER) are set.When detecting voice and begin, start delay timer and greeting timer.When delay timer surpasses the threshold value that is provided with in advance, restart the pause timer.After this, if when the pause timer surpasses the threshold value set in advance, then directly be judged as true man and reply; If when the greeting timer surpasses the threshold value that sets in advance, then be judged as replying of automatic answering equipment.
Mainly utilize a pause timer to judge that it still is the replying of automatic answering equipment that true man reply from above-mentioned detection method, if but the threshold value of pause timer is provided with smallerly, then easily the erroneous judgement of replying of automatic answering equipment is replied for true man, if and be provided with greatlyyer, easily true man being replied erroneous judgement again is replying of automatic answering equipment, the accuracy rate that influence detects.
Summary of the invention
The embodiment of the invention provides a kind of automatic answer method, can improve the accuracy rate that automatic-answering back device detects.
The method that a kind of automatic-answering back device detects, this method is:
Energy to input signal detects, when detecting voice signal, begin voice signal is carried out timing, and when detecting the voice signal pause first, begin dwell interval is carried out timing, if the dwell interval time arrives maximum pause threshold value, then voice signal is judged to be true man and replys; Otherwise, when subsequent detection is paused to voice signal, again dwell interval is carried out timing, and when the dwell interval time of reclocking arrives the minimum pause threshold value, voice signal is judged to be true man replys;
Do not reply if detect true man when described timing to voice signal arrives the automatic-answering back device threshold value, then directly voice signal is judged to be replying of automatic answering equipment, described automatic-answering back device threshold value is greater than the maximum threshold value of pausing.
The device that the embodiment of the invention also provides a kind of automatic-answering back device to detect, this device comprises:
The energy computing module calculates current energy value according to the input signal that receives;
The energy discrimination module, the speech energy threshold value of preserving in energy value that transmits according to the energy computing module and the memory module compares, when the energy value that transmits at the energy computing module arrives the speech energy threshold value first to timing module output voice timing initial signal, and when the follow-up energy value that transmits from the energy computing module is lower than the speech energy threshold value, export dwell interval timing initial signal to timing module, when arriving the speech energy threshold value, the follow-up energy value that transmits from the energy computing module exports dwell interval timing end signal to timing module;
Timing module begins voice signal is carried out timing when receiving voice timing initial signal, and clocking value is transferred to replys judging module; When receiving dwell interval timing initial signal each time, begin dwell interval is carried out timing, and clocking value is transferred to replys judging module; When receiving dwell interval timing end signal each time to dwell interval timing zero clearing;
Reply judging module, the threshold value of preserving clocking value that will transmit from timing module and the memory module compares, if the dwell interval timing arrives maximum pause threshold value for the first time, then judges and exports and reply testing result and reply for true man; If follow-up dwell interval timing arrives the minimum pause threshold value, then judge and export and reply testing result and reply for true man; If the voice signal timing arrives the automatic-answering back device threshold value, then judge and export that to reply testing result be replying of automatic answering equipment;
Memory module is used to preserve speech energy threshold value, maximum pause threshold value, minimum pause threshold value and automatic-answering back device threshold value.
The method and apparatus that above-mentioned automatic-answering back device detects when detecting failure to pausing for the first time, can utilize the minimum pause threshold value that follow-up dwell interval is continued to detect, and can improve the accuracy rate of replying detection effectively.
Description of drawings
Fig. 1 is the oscillogram that true man reply;
Fig. 2 is the oscillogram that automatic answering equipment is replied;
Fig. 3 is a kind of typical speech waveform figure;
Fig. 4 is the oscillogram among the inventive method embodiment;
Fig. 5 is the flow chart of the inventive method embodiment;
Fig. 6 is the structural representation of automatic-answering back device checkout gear of the present invention;
Fig. 7 is the structural representation of device embodiment.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with the accompanying drawings and the specific embodiments.
Show by a large amount of experiments and statistics, true man reply after each section voice finish, usually there is one section long dwell interval to wait for that the other side replys, such as: " feed " at first section voice and to have longer dwell interval afterwards, also have longer dwell interval afterwards here at second section voice " being * * company "; It is smaller that true man reply the probability that occurs shorter dwell interval after each section voice finish; The probability of replying the longer dwell interval of appearance after each section voice finish of automatic answering equipment is also smaller.
According to above-mentioned statistical law, in order effectively to prevent the generation of small probability event, improve the accuracy rate that detects, by two dwell interval threshold values are set, be maximum pause threshold value and minimum pause threshold value, utilize maximum dwell interval threshold value to the first time dwell interval detect, if detect failure, utilize minimum pause pocket door limit value that follow-up dwell interval is proceeded to detect again.Simultaneously, all fail in order to prevent all detections, also be provided with an automatic-answering back device threshold value, described automatic-answering back device threshold value is greater than maximum pause threshold value and minimum pause threshold value.
The concrete grammar that detects is:
Energy to input signal detects, when detecting voice signal, begin voice signal is carried out timing, and when detecting the voice signal pause first, begin dwell interval is carried out timing, if the dwell interval time arrives maximum pause threshold value, then voice signal is judged to be true man and replys; Otherwise, when subsequent detection is paused to voice signal, again dwell interval is carried out timing, and when the dwell interval time of reclocking arrives the minimum pause threshold value, voice signal is judged to be true man replys;
Do not reply if detect true man when the timing of voice signal is arrived described automatic-answering back device threshold value, then voice signal is judged to be replying of automatic answering equipment, described automatic-answering back device threshold value is greater than the maximum threshold value of pausing.
The input signal here is exactly the call answering that the background system of call center receives behind Outgoing Call.In order to narrate conveniently, the voice signal in the whole call answering can be divided into: the signal of greeting correspondence and the signal of greeting correspondence for the first time for the second time.The greeting first time here is to detect voice signal first to the signal that detects for the second time between the voice signal, detects voice signal signal afterwards for the second time and then can be referred to as greeting for the second time.
Fig. 3 is a typical speech waveform figure.As shown in Figure 3, A, C, E detect the moment that voice begin, and B, D, F detect the moment that voice finish, G is the moment that call answering finishes, and then A～C is the greeting first time, and C～G is the greeting second time, B～C is the dwell interval first time, and D～E is the dwell interval second time.
Detection to input signal can realize by the detection to the input signal energy.Such as: calculate the mean value of n sampling point energy of current input signal, and with the average energy value of calculating energy value as current demand signal.If the energy value of current demand signal rises than very fast, surpass the speech energy threshold value that sets in advance, just can determine that current is the moment that voice begin; Otherwise, be the moment of speech pause.
In the reality, in order to prevent that it is replying of automatic answering equipment that true man are replied erroneous judgement, can be provided with maximum dwell interval threshold value bigger.Like this, if for the first time dwell interval detects failure, i.e. the timing of dwell interval for the first time surpasses maximum dwell interval threshold value, and so, input signal may be the replying of automatic answering equipment, and also may reply for true man.In this case, can also utilize minimum pause pocket door limit value that follow-up dwell interval is effectively detected, its reason that can effectively detect is:
Suppose to be provided with minimum pause pocket door limit value smaller; if true man reply; for the second time dwell interval should be long, can surpass minimum pause pocket door limit value usually, automatic answering equipment reply that dwell interval is then very little above the probability of minimum pause pocket door limit value for the second time.According to this principle, can when surpassing minimum pause pocket door limit value, the dwell interval timing judge that directly input signal is that true man reply.
Certainly, if dwell interval detection has for the second time also been failed, follow-up dwell interval is detected with regard to also needing.In the reality, because that true man reply the smaller probability of dwell interval after each section voice is very little, so when total voice signal timing arrives the automatic-answering back device threshold value, just can judge directly that input signal is replying of automatic answering equipment.
In the practical application, the background noise that automatic answering equipment is replied is smaller usually, and people are when receiving calls in public places, and the background noise that true man reply is just bigger.Replying just bigger situation of background noise for true man, even there is long dwell interval, but owing to the energy of the input signal that noise is arranged is bigger, can't distinguish the moment that voice finish, can't accurately be the dwell interval timing promptly.
In this case, threshold value in the time of can setting in advance noise energy threshold value and noisemeter in the process that input signal is detected, also detects the noise in the input signal simultaneously, whether promptly detects noise energy greater than described noise energy threshold value.When detecting noise, just can begin noise is carried out timing, if threshold value when timing arrives described noisemeter then can be judged to be true man with voice signal and reply.Certainly, if threshold value during timing no show noisemeter is then carried out timing to noise again when detecting noise next time.
In the practical application, people may be single syllable words such as " feeding ", " " at the greeting first time that receives calls, and are the dwell interval first time afterwards; Described single syllable word then can not appear in automatic answering equipment reply usually.In this case, when detecting voice signal first, can be from detecting voice signal first to detecting between the speech pause first, determine the syllable number of voice according to the envelope of input signal, if be single syllable, the threshold value of then maximum being paused is revised as the maximum pause threshold value of single syllable correspondence, and the maximum pause threshold value of described single syllable correspondence is less than the maximum threshold value of pausing.
Here the threshold value of maximum being paused is revised as the maximum pause threshold value of single syllable correspondence, to detect for the first time in fact exactly, the threshold value at interval reduces, the reason that reduces is: owing to detect single syllable, increased the probability that input signal is replied for true man, be that dwell interval generally can be long, if the maximum threshold value of pausing is reduced, just can determine testing result quickly, improve detection efficiency.
In the practical application, if call answering is replying of automatic answering equipment, then can promptly play the prompt tone of a lasting fixed frequency for the other side after one section voice signal after greeting usually, promptly the BEEP sound points out the other side to leave a message.In this case, can be in the process that input signal is detected, whether be the signal of fixed frequency, if then be judged to be replying of automatic answering equipment if detecting input signal simultaneously.That is to say, in a period of time after detecting voice, can't judge that true man reply or the situation of replying of automatic answering equipment,, just can judge that this input signal is replying of automatic answering equipment if detect the signal that input signal is a fixed frequency.
In addition, the embodiment of the invention is when receiving the signaling of user's off-hook event, begins input signal is detected.If the situation of delay takes place the signaling of user's off-hook event, current input signal may not be the beginning of greeting for the first time so, and has passed through a part, causes easily and detects error.For this situation, can when receiving the signaling of user's off-hook event, earlier the input signal that receives be recalled, energy is initial as what detect less than the signal segment of the quiet threshold value that sets in advance.As for how input signal recalled and then to belong to prior art, no longer be described in detail herein.
Certainly, in the practical application, whether detect noise in the input signal, whether detect syllable number, whether detect the BEEP sound, whether the input signal that receives recalled etc. and can be determined voluntarily by the user who uses the present invention program, can use a certain method separately, also can be with above-mentioned several method combination in any, as for how to make up, enumerate no longer one by one herein.
The present invention program be described in detail below with embodiment in order to illustrate better.
Method embodiment
In the present embodiment, can carry out timing at interval to speech pause by the dwell interval timer, the automatic-answering back device timer carries out timing to whole voice signal, the noise timer carries out timing to detected noise, and threshold value when maximum pause threshold value, minimum pause threshold value, automatic-answering back device threshold value, noisemeter are set, and the maximum pause threshold value of single syllable correspondence.
Fig. 4 is the oscillogram of present embodiment input signal, and as shown in Figure 4, A is the moment that receives the off-hook event signaling, A ' is the starting point after recalling, and B and D are the moment that voice begin, and C and E are the moment of speech pause, B～D is the greeting first time, is the greeting second time after the D.
Fig. 5 is the flow chart of present embodiment.As shown in Figure 5, when receiving input signal, present embodiment detects the method for replying and may further comprise the steps:
Step 501: earlier input signal is recalled, with the starting point of A ' as detection; When the signaling of user's off-hook event takes place to postpone, can adopt this step to recall, the user can select this step as required.
Step 502: input signal is detected, calculate the energy value of input signal, and detect voice signal according to energy value first at the B point, promptly greeting begins for the first time.
In the present embodiment, when detecting voice signal first, can also detect noise and BEEP sound in the input signal as required simultaneously.If threshold value when the timing that detects noise is arrived noisemeter can directly be judged to be true man with input signal and reply.If detect BEEP, then can directly input signal be judged to be true man and reply.
Step 503: the energy value according to input signal detects greeting end for the first time at the C point, promptly detects speech pause, and determines that according to the envelope of input signal greeting is double-tone joint voice for the first time.
Step 504: the dwell interval timer picks up counting.
Step 505: detecting for the second time according to the energy value of input signal at the D point, greeting begins.
Step 506:, judge timing and do not exceed maximum pause threshold value according to dwell interval timer present timing result and maximum pause threshold value.
Step 507: with the zero clearing of dwell interval timer.
Step 508: the energy value according to input signal detects speech pause at the E point, the reclocking of dwell interval timer.
Step 509: when the F point is arrived in the timing of dwell interval timer, judge and arrive the minimum pause threshold value, then input signal is judged to be true man and replys.
In the practical application, omnidistance quiet timer and omnidistance quiet threshold value can also be set according to actual needs, and when receiving the signaling of off-hook event, pick up counting.Like this, no matter later whether detecting voice, can when arriving omnidistance quiet threshold value, the quiet timer of whole process directly provide result of determination, such as: directly be judged to be replying of automatic answering equipment, to prevent the situation of call answering without any effective voice signal.
In addition, in the practical application, for fear of the interference of noise, to the accurate timing of dwell interval, continuous speech timer and continuous speech timing thresholding can also be set according to actual needs, and described continuous speech timer is used for one section continuous real voice is carried out timing.If when timing arrives continuous speech timing thresholding, can judge that this signal is real voice signal, at this moment, can be with the zero clearing of dwell interval timer.And if timing does not arrive continuous speech timing thresholding, can think that this signal is not real voice signal, but one section noise signal, the dwell interval timer need not zero clearing, can continue timing.
At the automatic-answering back device detection method, the device that the embodiment of the invention also provides a kind of automatic-answering back device to detect.
Fig. 6 is the basic structure schematic diagram that the present invention replys checkout gear embodiment one.As shown in Figure 6, this device comprises:
Reply judging module 604, the threshold value of preserving clocking value that will transmit from timing module 603 and the memory module 605 compares, if the dwell interval timing arrives maximum pause threshold value for the first time, then judges and exports and reply testing result and reply for true man; If follow-up dwell interval timing arrives the minimum pause threshold value, then judge and export and reply testing result and reply for true man; If the voice signal timing arrives the automatic-answering back device threshold value, then judge and export that to reply testing result be replying of automatic answering equipment;
Fig. 7 is the apparatus structure schematic diagram of device present embodiment two.In the present embodiment, the device of replying detection not only comprise the initial judgement energy of energy computing module 601, voice discrimination module 602 among the embodiment one, timing control module 603 and, reply judging module 604 and memory module 605, in order to increase the accuracy of detection, can also detect noise, detection syllable number, detection BEEP sound in the input signal, and the input signal that receives recalled etc., this device also further comprises: noise measuring module 606, syllable tracking module 607, fixed frequency detection module 608, recall control module 609.
Wherein,
Reply judging module 604, the threshold value of preserving clocking value that will transmit from timing module 603 and the memory module 605 compares, if the dwell interval timing arrives maximum pause threshold value for the first time, then judges and exports and reply testing result and reply for true man; If follow-up dwell interval timing arrives the minimum pause threshold value, then judge and export and reply testing result and reply for true man; If the voice signal timing arrives the automatic-answering back device threshold value, then judge and export that to reply testing result be replying of automatic answering equipment; If threshold value when the noisemeter duration arrives noisemeter is then judged and is exported and replys testing result and reply for true man; If when receiving the single syllable marking signal from syllable tracking module 607, the maximum pause threshold value in the memory module 605 is revised as the maximum pause threshold value of single syllable correspondence;
Fixed frequency detection module 608 detects input signal, detecting when continuing fixed frequency, to replying judging module 604 output fixed frequency marking signals.
When input signal, recall 609 pairs of input signals that receive of control module and recall, and the signal after will recalling is exported to energy computing module 601, noise measuring module 606 and fixed frequency detection module 608.Afterwards, the signal of input will detect from three aspects:
First aspect, the noise that noise measuring module 606 detects in the input signal, and noise energy during greater than the energy threshold value to timing module 603 output noise timing control signals, timing module 603 begins noise is carried out timing, and will be transferred at the clocking value of noise and reply judging module 604, threshold value compares when replying judging module 604 with the noisemeter preserved in noisemeter duration and the memory module 605, if threshold value when arriving noisemeter is then judged and is exported and replys testing result and reply for true man;
Second aspect, whether fixed frequency detection module 608 detects input signal fixed frequency, if having, then to replying judging module 604 output fixed frequency marking signals, replys 604 of judging module and directly is judged to be replying of automatic answering equipment;
The 3rd aspect, energy computing module 601 bases are calculated current energy value from recalling control module 609 input signals, and energy value is exported to energy discrimination module 602 and syllable tracking module 607;
Energy discrimination module 602 is judged the speech energy threshold value of preserving in the energy value that transmits and the memory module 605, when the energy value that transmits at the energy computing module arrives the speech energy threshold value first to timing module 603 output voice timing initial signals, timing module 603 begins voice signal is carried out timing, and clocking value is transferred to replys judging module 604.When follow-up energy value is lower than the speech energy threshold value, energy discrimination module 602 need be exported dwell interval timing initial signals to timing module 603, and exports dwell interval timing end signals to timing module 603 when follow-up energy value arrives the speech energy threshold value.Wherein, when the first time, energy value was lower than the speech energy threshold value, syllable tracking module 607 will be according to the energy value of energy computing module 601 inputs, calculating is from detecting voice signal first to the syllable number that detects first between the voice signal pause, if be single syllable, then, reply the maximum pause threshold value that judging module 604 is revised as the maximum pause threshold value in the memory module 605 the single syllable correspondence again to replying judging module 604 output single syllable marking signals.Certainly, not single syllable if detect voice signal to the syllable number that detects first between the voice signal pause first, reply judging module 604 and will can not revise maximum pause threshold value.
In addition, when receiving dwell interval timing initial signal, timing module 603 will begin dwell interval is carried out timing, and clocking value is transferred to reply judging module 604; When receiving dwell interval timing end signal to dwell interval timing zero clearing.Compare and reply the threshold value of preserving in pause clocking value that judging module 604 can will receive and the memory module 605,, then judge and export and reply testing result and reply for true man if the dwell interval timing arrives maximum pause threshold value for the first time; If follow-up dwell interval timing arrives the minimum pause threshold value, then judge and export and reply testing result and reply for true man.
Use the present embodiment scheme, owing to be provided with two dwell interval threshold values, and utilize methods such as noise measuring, fixed frequency detection, can improve the accuracy rate of replying detection greatly; And when detecting when greeting for the first time to single syllable, the threshold value of also maximum being paused is revised as the maximum pause threshold value of single syllable correspondence, can improve the efficient of detection.
In sum, more than be preferred embodiment of the present invention only, be not to be used to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1, a kind of method of automatic-answering back device detection is characterized in that this method is:
Energy to input signal detects, when detecting voice signal, begin voice signal is carried out timing, and when detecting the voice signal pause first, begin dwell interval is carried out timing, if the dwell interval time arrives maximum pause threshold value, then voice signal is judged to be true man and replys; Otherwise, when subsequent detection is paused to voice signal, again dwell interval is carried out timing, and when the dwell interval time of reclocking arrives the minimum pause threshold value, voice signal is judged to be true man replys;
Do not reply if detect true man when the timing of voice signal is arrived described automatic-answering back device threshold value, then voice signal is judged to be replying of automatic answering equipment, described automatic-answering back device threshold value is greater than the maximum threshold value of pausing.
2, method according to claim 1 is characterized in that, in the process that the energy to input signal detects, this method further comprises:
When noise energy is greater than the noise energy threshold value in detecting input signal, begin noise is carried out timing, if threshold value when timing arrives noisemeter then is judged to be true man with voice signal and replys.
3, method according to claim 1 and 2 is characterized in that, when detecting voice signal first, this method further comprises:
From detecting voice signal first to detecting first between the voice signal pause, determine the syllable number of voice according to the envelope of signal, if be single syllable, then described maximum pause threshold value is revised as the maximum pause threshold value of single syllable correspondence, the maximum pause threshold value of described single syllable correspondence is less than described maximum pause threshold value.
4, method according to claim 1 and 2 is characterized in that, in the process that input signal is detected, this method further comprises:
Whether be the signal of fixed frequency, if the signal of fixed frequency then is judged to be replying of automatic answering equipment if detecting input signal.
5, method according to claim 1 and 2 is characterized in that, before input signal was detected, this method further comprised:
The input signal that receives is recalled, energy is initial as what detect less than the signal segment of the quiet threshold value that sets in advance.
6, a kind of device of automatic-answering back device detection is characterized in that this device comprises:
The energy computing module calculates current energy value according to the input signal that receives;
Memory module is used to preserve speech energy threshold value, maximum pause threshold value, minimum pause threshold value and automatic-answering back device threshold value;
The energy discrimination module, the speech energy threshold value of preserving in energy value that transmits according to the energy computing module and the memory module is judged, when the energy value that transmits at the energy computing module arrives the speech energy threshold value first to timing module output voice timing initial signal, and when the follow-up energy value that transmits from the energy computing module is lower than the speech energy threshold value, export dwell interval timing initial signal to timing module, when arriving the speech energy threshold value, the follow-up energy value that transmits from the energy computing module exports dwell interval timing end signal to timing module;
Timing module begins voice signal is carried out timing when receiving voice timing initial signal, and clocking value is transferred to replys judging module; When receiving dwell interval timing initial signal, begin dwell interval is carried out timing, and clocking value is transferred to replys judging module; When receiving dwell interval timing end signal to dwell interval timing zero clearing;
Reply judging module, the threshold value of preserving clocking value that will transmit from timing module and the memory module compares, if the dwell interval timing arrives maximum pause threshold value for the first time, then judges and exports and reply testing result and reply for true man; If follow-up dwell interval timing arrives the minimum pause threshold value, then judge and export and reply testing result and reply for true man; If the voice signal timing arrives the automatic-answering back device threshold value, then judge and export that to reply testing result be replying of automatic answering equipment.
7, device according to claim 6 is characterized in that, described device further comprises:
The noise measuring module is used for detecting the noise of input signal, and noise energy during greater than the noise energy threshold value to timing module output noise timing control signal;
Described timing module is further used for: during control signal, begin noise is carried out timing when receiving the noisemeter that the noise measuring module transmits, and clocking value is transferred to replys judging module;
Replying judging module is further used for: threshold value compares during the noisemeter preserved in noisemeter duration that timing module is transmitted and the memory module, if threshold value when the noisemeter duration arrives noisemeter is then judged and is exported and replys testing result and reply for true man.
8, device according to claim 6 is characterized in that, this device further comprises:
The syllable tracking module according to the energy value of energy computing module input, calculates from detecting voice signal first to the syllable number that detects first between the voice signal pause, if be single syllable, then to replying judging module output single syllable marking signal;
Reply judging module and be further used for,, will be kept at the maximum pause threshold value that maximum pause threshold value in the memory module is revised as the single syllable correspondence according to the single syllable marking signal of syllable tracking module input.
9, device according to claim 6 is characterized in that, this device further comprises:
The fixed frequency detection module detects input signal, detecting when continuing fixed frequency, to replying judging module output fixed frequency marking signal;
Reply judging module and be further used for, when receiving the fixed frequency marking signal that the fixed frequency detection module transmits, judge and output is replied testing result and replied for true man.
10, device according to claim 6 is characterized in that, this device further comprises:
Recall control module, be used for the input signal that receives is recalled, and will date back to energy and export to the energy computing module less than the signal of quiet threshold value.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007100981197A CN101287029A (en) | 2007-04-13 | 2007-04-13 | Method and apparatus for automatically respond to detection |
| PCT/CN2008/070644 WO2008125047A1 (en) | 2007-04-13 | 2008-03-31 | A method for detecting the automatic answering and an apparatus thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007100981197A CN101287029A (en) | 2007-04-13 | 2007-04-13 | Method and apparatus for automatically respond to detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101287029A true CN101287029A (en) | 2008-10-15 |
Family
ID=39863268
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2007100981197A Pending CN101287029A (en) | 2007-04-13 | 2007-04-13 | Method and apparatus for automatically respond to detection |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN101287029A (en) |
| WO (1) | WO2008125047A1 (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102523356A (en) * | 2011-12-21 | 2012-06-27 | 上海会畅通讯科技发展有限公司 | Method and device for detecting sound of meeting passages |
| CN103856642A (en) * | 2014-01-22 | 2014-06-11 | 中国移动（深圳）有限公司 | Detection method and system |
| CN107392757A (en) * | 2017-07-24 | 2017-11-24 | 重庆小雨点小额贷款有限公司 | Signal auditing method and device |
| CN109949813A (en) * | 2017-12-20 | 2019-06-28 | 北京君林科技股份有限公司 | A kind of method, apparatus and system converting speech into text |
| CN111405131A (en) * | 2020-03-30 | 2020-07-10 | 深圳震有科技股份有限公司 | Method, system and storage medium for detecting far-end off-hook signal |
| CN114070935A (en) * | 2022-01-12 | 2022-02-18 | 百融至信(北京)征信有限公司 | Intelligent outbound interruption method and system |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103973872B (en) * | 2013-01-31 | 2016-12-28 | 联想(北京)有限公司 | A kind of method of information processing and electronic equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100428757C (en) * | 2003-09-28 | 2008-10-22 | 维音数码（上海）有限公司 | Telephone auto externally dialing method for client service and its intelligent analysis and management |
| US7184521B2 (en) * | 2004-06-10 | 2007-02-27 | Par3 Communications, Inc. | Method and system for identifying a party answering a telephone call based on simultaneous activity |
| CN100496070C (en) * | 2005-03-08 | 2009-06-03 | 华为技术有限公司 | Identification system of automatic answering equipment and its identification method |
-
2007
- 2007-04-13 CN CNA2007100981197A patent/CN101287029A/en active Pending
-
2008
- 2008-03-31 WO PCT/CN2008/070644 patent/WO2008125047A1/en active Application Filing
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102523356A (en) * | 2011-12-21 | 2012-06-27 | 上海会畅通讯科技发展有限公司 | Method and device for detecting sound of meeting passages |
| CN102523356B (en) * | 2011-12-21 | 2015-04-15 | 上海会畅通讯股份有限公司 | Method and device for detecting sound of meeting passages |
| CN103856642A (en) * | 2014-01-22 | 2014-06-11 | 中国移动（深圳）有限公司 | Detection method and system |
| CN103856642B (en) * | 2014-01-22 | 2016-08-17 | 中国移动（深圳）有限公司 | A kind of detection method and system |
| CN107392757A (en) * | 2017-07-24 | 2017-11-24 | 重庆小雨点小额贷款有限公司 | Signal auditing method and device |
| CN109949813A (en) * | 2017-12-20 | 2019-06-28 | 北京君林科技股份有限公司 | A kind of method, apparatus and system converting speech into text |
| CN111405131A (en) * | 2020-03-30 | 2020-07-10 | 深圳震有科技股份有限公司 | Method, system and storage medium for detecting far-end off-hook signal |
| CN114070935A (en) * | 2022-01-12 | 2022-02-18 | 百融至信(北京)征信有限公司 | Intelligent outbound interruption method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2008125047A1 (en) | 2008-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101287029A (en) | Method and apparatus for automatically respond to detection | |
| CN110557451B (en) | Dialogue interaction processing method and device, electronic equipment and storage medium | |
| CN101490741B (en) | Detecting an answering machine using speech recognition | |
| US4914692A (en) | Automatic speech recognition using echo cancellation | |
| CN102404462B (en) | Call progress analyzing method for phone dialing system and device | |
| CN102137194B (en) | Call detection method and device | |
| CA2569660A1 (en) | Method and system for identifying a party answering a telephone call based on simultaneous activity | |
| CN106851839A (en) | Frame structure determines method and base station | |
| CN103747511B (en) | information broadcasting method and system | |
| CN105700660A (en) | Electronic Device Comprising a Wake Up Module Distinct From a Core Domain | |
| CN102104473A (en) | Method and system for conversation between simplex terminal and duplex terminal | |
| CN103295585A (en) | Processing system and method identifying whether telephone automatic dialing is responded by real human | |
| WO2009111957A1 (en) | A method and corresponding device for detecting line status, and a predictive outbound dialing system | |
| CN103617801A (en) | Voice detection method and device and electronic equipment | |
| CN101540797A (en) | Incoming number filter method, system and mobile terminal | |
| CN110705926A (en) | Method, device and system for acquiring logistics object distribution information | |
| CN102651857A (en) | Busy telephone line processing system and method | |
| CN109151148A (en) | Recording method, device, terminal and the computer readable storage medium of dialog context | |
| CN1175152A (en) | Intensified delayed information business | |
| CN102523356B (en) | Method and device for detecting sound of meeting passages | |
| EP3937474A1 (en) | Computer-generated speech detection | |
| ATE327617T1 (en) | COLLISION PROCESSING SCHEME FOR A DISCRETE MULTI-TONE DATA COMMUNICATION NETWORK | |
| CN100407732C (en) | Mobile communication terminal and method of uleave-mode auto-responsion using the same | |
| CN102083231B (en) | A kind of device and method for eliminating polyphonic ringtone playing noise | |
| CN102790958B (en) | A kind of method of command transmitting 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 | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20081015 |