CN110062117B - Sound wave detection and early warning method - Google Patents
Sound wave detection and early warning method Download PDFInfo
- Publication number
- CN110062117B CN110062117B CN201910277346.9A CN201910277346A CN110062117B CN 110062117 B CN110062117 B CN 110062117B CN 201910277346 A CN201910277346 A CN 201910277346A CN 110062117 B CN110062117 B CN 110062117B
- Authority
- CN
- China
- Prior art keywords
- sound wave
- data
- decibel
- information
- conversation
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 80
- 238000001514 detection method Methods 0.000 title claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 47
- 230000005236 sound signal Effects 0.000 claims abstract description 34
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 230000004044 response Effects 0.000 claims description 43
- 238000001228 spectrum Methods 0.000 claims description 30
- 238000010801 machine learning Methods 0.000 claims description 15
- 230000003321 amplification Effects 0.000 claims description 10
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 10
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 abstract description 39
- 230000008451 emotion Effects 0.000 abstract description 5
- 230000008859 change Effects 0.000 abstract description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 238000011835 investigation Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
-
- 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
- H04M3/5175—Call or contact centers supervision arrangements
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Psychiatry (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Hospice & Palliative Care (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Child & Adolescent Psychology (AREA)
- Telephone Function (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention provides a sound wave detection and early warning method, which can acquire and record sound signals from a conversation party in real time in the conversation process, extract corresponding sound wave frequency data and/or sound wave decibel data from the sound signals, and perform corresponding analysis calculation processing on the sound wave frequency data and/or the sound wave decibel data to obtain corresponding analysis results; when the emotion or the satisfaction degree of the conversation quality of the conversation party in the conversation process changes, correspondingly, the sound wave frequency and the sound wave decibel of the conversation party speaking also change, so that the conversation quality of the current conversation process can be obtained by acquiring the analysis result of the sound wave frequency data and/or the sound wave decibel data in real time.
Description
Technical Field
The invention relates to the technical field of sound detection and recognition, in particular to a sound wave detection and early warning method.
Background
With the development of electronic commerce and online service industries, a considerable number of service industries open corresponding online service platforms at present, and users can perform adaptive operation through the corresponding online service platforms to obtain peer-to-peer services. Since such online service enterprises generally have a large scale, in order to ensure the service quality, telephone consultation and complaint hot lines are also provided for other online services, and the user can make a manual call by dialing the corresponding telephone consultation and complaint hot lines, so as to reflect the related questions and obtain corresponding responses. In order to ensure the service quality of the manual call, in general, an enterprise needs to record the manual call process of each user, and the corresponding recording data is stored and then analyzed. The recorded data is not only used for monitoring and evaluating the call service quality of the service personnel who have a manual call, but also can know the requirements of different users in time so as to make corresponding feedback. However, in the conventional monitoring of the manual call process, after all call records of the manual call are completely recorded, corresponding analysis results can be obtained only by performing manual analysis on corresponding recorded data one by one, and thus, the conventional monitoring operation for the manual call process has a certain time lag and can be completed only by allocating a large amount of manpower and material resources.
Although the existing online service telephone hotline can record the whole manual call process, the quality and the user appeal of the manual call process are analyzed after the manual call process is finished, and the follow-up analysis operation on manual call record data can be realized only by additionally equipping manpower and material resources. It is thus clear that this analysis mode to manual call recording data needs different staff to carry out analysis data to recording data, and inevitable can make recording data revealed and extension at the in-process of analysis data to possibly involve user's individual privacy information at manual call in-process, this kind of analysis operation to recording data is very likely to damage user's recording privacy promptly. In addition, the analysis operation of the recording data cannot obtain a corresponding analysis result in real time in the manual call process, and a large amount of manpower is required for subsequent analysis operation, which cannot realize timely improvement of the quality of the manual call service and obtain the related satisfaction degree of the user in the manual call process in real time, and cannot effectively and timely improve the quality of the manual call service.
Disclosure of Invention
The existing analysis of the service quality and the user requirement in the manual call process is realized by manually analyzing the recording data one by one after recording the whole manual call process, although the analysis mode can completely and comprehensively identify and analyze all the recording data so as to improve the effectiveness of the analysis result, the analysis mode can be realized on the premise of completely obtaining the recording data, that is to say, the analysis mode can not obtain the analysis result in real time in the manual call process, and has a certain time lag. Since many unexpected emergencies may occur during the manual calling process, the user may also have emotional fluctuation with the progress of the manual calling, and the above-mentioned situations may cause the degradation of the quality of the manual calling if the situations cannot be processed in time during the manual calling process, it is important to analyze the emotion and satisfaction of the calling about the user in real time during the manual calling process. In addition, the existing analysis and processing of the manual call recording data are realized by means of extra human hands, the manual call recording data often relate to personal privacy data of users, the possibility that the personal privacy data are leaked exists, and human hand analysis of the manual call recording data requires that an analyst analyzes the content of all the recording data one by one, so that a large amount of manpower and material resources are inevitably wasted.
Aiming at the defects in the prior art, the invention provides a sound wave detection and early warning method which can acquire and record sound signals from a conversation party in real time in the conversation process, extract corresponding sound wave frequency data and/or sound wave decibel data from the sound signals, and perform corresponding analysis calculation processing on the sound wave frequency data and/or the sound wave decibel data to obtain corresponding analysis results; when the emotion of the conversation party in the conversation process or the satisfaction degree of the conversation quality changes, correspondingly, the sound wave frequency and the sound wave decibel of the speaking of the conversation party also change, so that the conversation quality in the current conversation process can be obtained by acquiring the analysis result of the sound wave frequency data and/or the sound wave decibel data in real time; in addition, the sound wave detection and early warning method can also determine the call response level corresponding to the current call process according to the analysis result of the sound wave frequency data and/or the sound wave decibel data, and can switch the call mode adaptively according to the call response level in real time, so as to realize the timely investigation and real-time manual intervention on the call process. Therefore, the sound wave detection and early warning method can analyze in real time to obtain a corresponding call analysis result in the call process, and performs preset call mode switching according to the call analysis result to realize instant manual intervention.
The invention provides a sound wave detection and early warning method which is characterized by comprising the following steps:
the method comprises the following steps that (1) sound signals from a conversation party are obtained, and sound wave frequency data and/or sound wave decibel data corresponding to the sound signals are recorded within a preset time length;
step (2), acquiring corresponding first sound wave frequency information and second sound wave frequency information and/or first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data;
a step (3) of determining a current call response level of the conversation partner based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information;
step (4), based on the current call response level of the conversation party, the current call mode with the conversation party is subjected to adaptive mode switching operation;
further, in step (1), after acquiring the voice signal from the conversation party, the method further includes performing validity judgment processing on the voice signal from the conversation party, specifically, acquiring voiceprint data of the voice signal, generating conversation party feature information about the voice signal according to the voiceprint data, matching the conversation party feature information with a preset voiceprint database, if the conversation party feature information is matched with the preset voiceprint database, determining that the corresponding voice signal is a valid voice signal, and if the conversation party feature information is not matched with the preset voiceprint database, determining that the corresponding voice signal is an invalid voice signal;
further, in step (1), generating dialogue-side feature information about the voice signal according to the voiceprint data and matching the dialogue-side feature information with the preset voiceprint database specifically includes extracting a corresponding voiceprint distribution spectrum from the voiceprint data, converting the voiceprint distribution spectrum into identity information about the dialogue side through a first machine learning algorithm model to serve as the dialogue-side feature information, calculating a correlation value between the identity information and the preset voiceprint database through a second machine learning algorithm model, and performing the matching according to the correlation value;
further, in the step (1), the matching processing according to the correlation value specifically includes comparing the calculated correlation value with a preset correlation threshold, determining that the dialogue party feature information matches with the preset voiceprint database and determines that the corresponding voice signal is a valid voice signal if the correlation value exceeds the preset correlation threshold, and determining that the dialogue party feature information does not match with the preset voiceprint database and determines that the corresponding voice signal is an invalid voice signal if the correlation value does not exceed the preset correlation threshold; wherein, for the valid sound signal, the recording processing of the sound wave frequency data and/or sound wave decibel data is executed, and for the invalid sound signal, the recording processing of the sound wave frequency data and/or sound wave decibel data is not executed;
further, in the step (1), recording the sound wave frequency data and/or the sound wave decibel data corresponding to the sound signal within a preset time length specifically includes recording a plurality of corresponding different sound wave frequency data and/or a plurality of different sound wave decibel data within a plurality of different preset time lengths, calculating a data recording integrity degree value corresponding to each of the plurality of different sound wave frequency data and/or each of the plurality of different sound wave decibel data through a third machine learning algorithm model, and taking the sound wave frequency data and/or the sound wave decibel data corresponding to the data recording integrity degree value with the highest data recording integrity degree value as the calculation processing object data in the step (2);
further, in the step (2), based on the sound wave frequency data and/or the sound wave decibel data, acquiring corresponding first sound wave frequency information and second sound wave frequency information, and/or first sound wave decibel information and second sound wave decibel information specifically includes acquiring an average value of sound wave spectrum peak-trough difference values corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, acquiring a sound wave spectrum peak-trough difference value corresponding to the sound wave frequency data immediately after another preset time length within the preset time length as the second sound wave frequency information, and/or acquiring a sound wave decibel average value corresponding to the sound wave decibel data within the preset time length as the first sound wave decibel information, and acquiring a sound wave decibel value corresponding to the sound wave decibel data immediately after another preset time length within the preset time length as the second sound wave decibel information Sound wave decibel information;
further, in the step (3), the determining of the current talk response level with respect to the dialer based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information specifically includes, calculating a first accumulated total time length that the difference value between the wave peaks and the wave troughs of the sound wave spectrum exceeds the average value of the difference values between the wave peaks and the wave troughs of the sound wave spectrum in the total call time length corresponding to the call of the conversation party, and determining the call response level according to the ratio of the first accumulated total duration to the total call duration, and/or, calculating a second accumulated total time length of the sound wave decibel value exceeding the sound wave decibel average value in the total call time length corresponding to the call of the other party, determining the call response level according to the ratio of the second accumulated total duration to the total call duration;
further, in the step (2), acquiring corresponding first sound wave frequency information and second sound wave frequency information, and/or first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data specifically includes acquiring an amplification value of a sound wave spectrum peak-trough difference value corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, and determining an amplification threshold value of the sound wave spectrum peak-trough difference value corresponding to a current conversation with a conversation party as the second sound wave frequency information, and/or acquiring a sound wave decibel amplification value corresponding to the sound wave decibel data within a preset time length to serve as the first sound wave decibel information, and determining a sound wave decibel amplification threshold value corresponding to the conversation of the conversation party at present to serve as the second sound wave decibel information;
further, in the step (3), the determining of the current call response level of the party to be conversed based on the first sound wave frequency information and the second sound wave frequency information, and/or based on the first sound wave decibel information and the second sound wave decibel information specifically includes determining that an amplitude value of the sound wave spectrum peak-trough difference exceeds an amplitude threshold value of the sound wave spectrum peak-trough difference during the conversation with the party to be conversed, and determining the call response level accordingly, and/or determining that the sound wave decibel amplitude value exceeds the sound wave decibel amplitude threshold value during the conversation with the opposite party to be conversed, and determining the call response level accordingly;
further, in the step (4), the performing, based on the current call response level of the party, an adaptive mode switching operation on the current call mode with the party specifically includes presetting a plurality of call response levels with different event occurrence probabilities according to a fourth machine learning algorithm model, and setting a plurality of call modes corresponding to the plurality of call response levels one to one according to the fourth machine learning algorithm model, so as to perform a corresponding call mode switching operation based on the call response level determined in the current call process with the other party.
Compared with the prior art, the sound wave detection and early warning method can acquire and record the sound signals from a conversation party in real time in the conversation process, extract corresponding sound wave frequency data and/or sound wave decibel data from the sound signals, and perform corresponding analysis calculation processing on the sound wave frequency data and/or the sound wave decibel data to obtain corresponding analysis results; when the emotion of the conversation party in the conversation process or the satisfaction degree of the conversation quality changes, correspondingly, the sound wave frequency and the sound wave decibel of the speaking of the conversation party also change, so that the conversation quality in the current conversation process can be obtained by acquiring the analysis result of the sound wave frequency data and/or the sound wave decibel data in real time; in addition, the sound wave detection and early warning method can also determine the call response level corresponding to the current call process according to the analysis result of the sound wave frequency data and/or the sound wave decibel data, and can switch the call mode adaptively according to the call response level in real time, so as to realize the timely investigation and real-time manual intervention on the call process. Therefore, the sound wave detection and early warning method can analyze in real time to obtain a corresponding call analysis result in the call process, and performs preset call mode switching according to the call analysis result to realize instant manual intervention.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a sound wave detection and early warning method provided by the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a sound wave detection and early warning method according to an embodiment of the present invention. The sound wave detection and early warning method can comprise the following steps:
step (1), acquiring a sound signal from a conversation party, and recording sound wave frequency data and/or sound wave decibel data corresponding to the sound signal within a preset time length.
Preferably, in step (1), after obtaining the sound signal from the conversation party, the method may further include performing validity judgment processing on the sound signal from the conversation party, specifically, acquiring voiceprint data of the sound signal, generating conversation party feature information about the sound signal according to the voiceprint data, matching the conversation party feature information with a preset voiceprint database, determining that the corresponding sound signal is a valid sound signal if the conversation party feature information matches the preset voiceprint database, and determining that the corresponding sound signal is an invalid sound signal if the conversation party feature information does not match the preset voiceprint database.
Preferably, in the step (1), the generating of the feature information of the speech party with respect to the sound signal according to the voiceprint data and the matching of the feature information of the speech party with the preset voiceprint database may specifically include extracting a corresponding voiceprint distribution spectrum from the voiceprint data, converting the voiceprint distribution spectrum into the identification information of the speech party through a first machine learning algorithm model to serve as the feature information of the speech party, calculating a correlation value between the identification information and the preset voiceprint database through a second machine learning algorithm model, and performing the matching according to the correlation value.
Preferably, in the step (1), the matching process according to the correlation value may specifically include comparing the calculated correlation value with a preset correlation threshold, if the correlation value exceeds the preset correlation threshold, determining that the feature information of the dialog party is matched with the preset voiceprint database while determining that the corresponding voice signal is a valid voice signal, and if the correlation value does not exceed the preset correlation threshold, determining that the feature information of the dialog party is not matched with the preset voiceprint database while determining that the corresponding voice signal is an invalid voice signal; wherein the recording process of the sound wave frequency data and/or sound wave decibel data is performed for the valid sound signal, and the recording process of the sound wave frequency data and/or sound wave decibel data is not performed for the invalid sound signal.
Preferably, in the step (1), the recording of the sound wave frequency data and/or the sound wave decibel data corresponding to the sound signal within the preset time length may specifically include recording a plurality of corresponding different sound wave frequency data and/or a plurality of different sound wave decibel data within a plurality of different preset time lengths, calculating a data recording integrity degree value corresponding to each of the plurality of different sound wave frequency data and/or each of the plurality of different sound wave decibel data through a third machine learning algorithm model, and using the sound wave frequency data and/or the sound wave decibel data corresponding to the data recording integrity degree value having the highest data recording integrity degree value as the calculation processing object data in the step (2).
And (2) acquiring corresponding first sound wave frequency information and second sound wave frequency information and/or first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data.
Preferably, in the step (2), the obtaining of the corresponding first sound wave frequency information and second sound wave frequency information, and/or the first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data may specifically include obtaining an average value of sound wave spectrum peak-valley difference values corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, and obtaining a sound wave spectrum peak-valley difference value corresponding to the sound wave frequency data immediately after another preset time length in the preset time length as the second sound wave frequency information, and/or acquiring a sound wave decibel average value corresponding to the sound wave decibel data within a preset time length as the first sound wave decibel information, and acquiring a sound wave decibel value corresponding to another preset time length following the preset time length of the sound wave decibel data as the second sound wave decibel information.
Preferably, in the step (2), the obtaining of the corresponding first sound wave frequency information and second sound wave frequency information, and/or the first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data may specifically include obtaining an amplitude value of a sound wave spectrum peak-trough difference corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, and determining an amplitude threshold value of a sound wave spectrum peak-trough difference corresponding to a current conversation with a conversation party as the second sound wave frequency information, and/or acquiring a sound wave decibel amplification value corresponding to the sound wave decibel data within a preset time length as the first sound wave decibel information, and determining a sound wave decibel amplification threshold value corresponding to the conversation of the conversation party at present as the second sound wave decibel information.
Preferably, the obtaining of the corresponding first sound wave frequency information and second sound wave frequency information, and/or the corresponding first sound wave decibel information and second sound wave decibel information is realized based on obtaining a normal reference line corresponding to the sound wave frequency data and/or the sound wave decibel data, where the obtaining of the sound wave frequency data or the normal reference line corresponding to the sound wave decibel data may specifically be assuming that a sound wave trough-peak difference value or a sound wave decibel value of a sound signal from a conversation party is B at any time in the process of monitoring and obtaining the sound signal from the conversation partynWherein n represents the duration of a callIn seconds, the average growth rate G of the wave trough-peak difference or the average growth rate G of the sound decibel value at any time n on the dialogue sidenComprises the following steps:
for any time n +1, the theoretical sound wave trough-peak difference or the theoretical sound decibel value BlComprises the following steps:
Bl＝α+βGn
in the above expression, α and β are constants, GnThe average increasing speed of the wave trough-wave crest difference value or the average increasing speed of the sound decibel value of the conversation party at any time n. The constants α and β are calculated as follows:
performing minimum multiplicative regression fitting on constants alpha and beta according to all historical normal dialogue data, namely making deltai＝Bi-Bl＝Bi-α-βGnIn which B isiIs the actual sound wave trough-peak difference or actual sound decibel value, BlThe sum of the squares of the errors is minimized, i.e. obtained, for the theoretical sound wave trough-peak difference or the theoretical sound decibel value according to the principle of least square method
Minimum, then can find
And (3) determining the current call response level of the conversation party based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information.
Preferably, in the step (3), the determining of the current talk response level with respect to the dialer based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information may specifically include, calculating a first accumulated total time length that the difference value between the wave crest and the wave trough of the sound wave spectrum exceeds the average value of the difference value between the wave crest and the wave trough of the sound wave spectrum in the total call time length corresponding to the call of the conversation party, and determining the call response level according to the ratio of the first accumulated total duration to the total call duration, and/or, calculating a second accumulated total time length that the sound wave decibel value exceeds the sound wave decibel average value in the total time length corresponding to the opposite party call, and determining the call response level according to the ratio of the second accumulated total duration to the total call duration.
Preferably, in the step (3), the determining of the current call response level regarding the party on the basis of the first sound wave frequency information and the second sound wave frequency information, and/or the first sound wave decibel information and the second sound wave decibel information may specifically include determining that an amplitude of the sound wave spectrum peak-trough difference exceeds an amplitude threshold of the sound wave spectrum peak-trough difference during the call with the party, and determining the call response level accordingly, and/or determining that the sound wave decibel amplitude exceeds the sound wave decibel amplitude threshold during the call with the other party, and determining the call response level accordingly.
And (4) carrying out adaptive mode switching operation on the current call mode with the conversation party on the basis of the current call response level of the conversation party.
Preferably, in the step (4), the performing, based on the current call response level of the party, an adaptive mode switching operation on the current call mode with the party may specifically include presetting a plurality of call response levels with different event occurrence probabilities according to a fourth machine learning algorithm model, and setting a plurality of call modes one-to-one corresponding to the plurality of call response levels according to the fourth machine learning algorithm model, so as to perform a corresponding call mode switching operation based on the call response level determined in the current call process with the other party.
It can be seen from the above embodiments that the sound wave detection and early warning method can acquire and record the sound signal from the conversation party in real time during the conversation process, extract the corresponding sound wave frequency data and/or sound wave decibel data from the sound signal, and perform corresponding analysis, calculation and processing on the sound wave frequency data and/or the sound wave decibel data to obtain the corresponding analysis result; when the emotion of the conversation party in the conversation process or the satisfaction degree of the conversation quality changes, correspondingly, the sound wave frequency and the sound wave decibel of the speaking of the conversation party also change, so that the conversation quality in the current conversation process can be obtained by acquiring the analysis result of the sound wave frequency data and/or the sound wave decibel data in real time; in addition, the sound wave detection and early warning method can also determine the call response level corresponding to the current call process according to the analysis result of the sound wave frequency data and/or the sound wave decibel data, and can switch the call mode adaptively according to the call response level in real time, so as to realize the timely investigation and real-time manual intervention on the call process. Therefore, the sound wave detection and early warning method can analyze in real time to obtain a corresponding call analysis result in the call process, and performs preset call mode switching according to the call analysis result to realize instant manual intervention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (8)
1. A sound wave detection and early warning method is characterized by comprising the following steps:
the method comprises the following steps that (1) sound signals from a conversation party are obtained, and sound wave frequency data and/or sound wave decibel data corresponding to the sound signals are recorded within a preset time length;
step (2), acquiring corresponding first sound wave frequency information and second sound wave frequency information and/or first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data;
a step (3) of determining a current call response level of the conversation partner based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information;
step (4), based on the current call response level of the conversation party, the current call mode with the conversation party is subjected to adaptive mode switching operation;
in step (2), the obtaining of the corresponding first sound wave frequency information and second sound wave frequency information, and/or the first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data specifically includes obtaining an average value of sound wave spectrum peak-trough difference values corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, obtaining a sound wave spectrum peak-trough difference value corresponding to the sound wave frequency data immediately after another preset time length within the preset time length as the second sound wave frequency information, and/or obtaining a sound wave decibel average value corresponding to the sound wave decibel data within the preset time length as the first sound wave decibel information, and obtaining a sound wave decibel value corresponding to the sound wave decibel data immediately after another preset time length within the preset time length as the second sound wave decibel information Information;
the determining of the current talk response level with respect to the dialer based on the first sound wave frequency information and the second sound wave frequency information and/or based on the first sound wave decibel information and the second sound wave decibel information in step (3) may specifically include, calculating a first accumulated total time length that the difference value between the wave peaks and the wave troughs of the sound wave spectrum exceeds the average value of the difference values between the wave peaks and the wave troughs of the sound wave spectrum in the total call time length corresponding to the call of the conversation party, and determining the call response level according to the ratio of the first accumulated total duration to the total call duration, and/or, calculating a second accumulated total time length of the sound wave decibel value exceeding the sound wave decibel average value in the total call time length corresponding to the call of the other party, and determining the call response level according to the ratio of the second accumulated total duration to the total call duration.
2. The acoustic detection and warning method of claim 1, wherein: in the step (1), after the voice signal from the conversation party is obtained, the judgment processing on the validity of the voice signal from the conversation party is further included, specifically, voiceprint data is collected on the voice signal, conversation party characteristic information about the voice signal is generated according to the voiceprint data, then the conversation party characteristic information is matched with a preset voiceprint database, if the conversation party characteristic information is matched with the preset voiceprint database, the corresponding voice signal is determined to be a valid voice signal, and if the conversation party characteristic information is not matched with the preset voiceprint database, the corresponding voice signal is determined to be an invalid voice signal.
3. The acoustic detection and warning method of claim 2, wherein: in step (1), generating dialogue-side feature information about the voice signal according to the voiceprint data and matching the dialogue-side feature information with the preset voiceprint database specifically includes extracting a corresponding voiceprint distribution spectrum from the voiceprint data, converting the voiceprint distribution spectrum into identity information about the dialogue side through a first machine learning algorithm model to serve as the dialogue-side feature information, calculating a correlation value between the identity information and the preset voiceprint database through a second machine learning algorithm model, and performing the matching according to the correlation value.
4. The acoustic detection and warning method of claim 3, wherein: in step (1), the matching process according to the correlation value specifically includes comparing the calculated correlation value with a preset correlation threshold, determining that the feature information of the dialog party is matched with the preset voiceprint database and the corresponding voice signal is an effective voice signal at the same time if the correlation value exceeds the preset correlation threshold, and determining that the feature information of the dialog party is not matched with the preset voiceprint database and the corresponding voice signal is an ineffective voice signal at the same time if the correlation value does not exceed the preset correlation threshold; wherein the recording process of the sound wave frequency data and/or sound wave decibel data is performed for the valid sound signal, and the recording process of the sound wave frequency data and/or sound wave decibel data is not performed for the invalid sound signal.
5. The acoustic detection and warning method of claim 1, wherein: in the step (1), recording the sound wave frequency data and/or the sound wave decibel data corresponding to the sound signal within a preset time length specifically includes recording a plurality of corresponding different sound wave frequency data and/or a plurality of different sound wave decibel data within a plurality of different preset time lengths, calculating a data recording integrity degree value corresponding to each of the plurality of different sound wave frequency data and/or each of the plurality of different sound wave decibel data through a third machine learning algorithm model, and using the sound wave frequency data and/or the sound wave decibel data corresponding to the data recording integrity degree value with the highest data recording integrity degree value as the calculation processing object data in the step (2).
6. The acoustic detection and warning method of claim 1, wherein: in the step (2), acquiring corresponding first sound wave frequency information and second sound wave frequency information, and/or first sound wave decibel information and second sound wave decibel information based on the sound wave frequency data and/or the sound wave decibel data specifically includes acquiring an amplification value of a sound wave spectrum peak-trough difference value corresponding to the sound wave frequency data within a preset time length as the first sound wave frequency information, and determining an amplification threshold value of the sound wave spectrum peak-trough difference value corresponding to a current conversation with a conversation party as the second sound wave frequency information, and/or acquiring a sound wave decibel amplification value corresponding to the sound wave decibel data within a preset time length to serve as the first sound wave decibel information, and determining a sound wave decibel amplification threshold value corresponding to the conversation of the conversation party at present to serve as the second sound wave decibel information.
7. The acoustic detection and warning method of claim 1, wherein: in the step (3), the determining of the current call response level of the conversation party based on the first sound wave frequency information and the second sound wave frequency information, and/or based on the first sound wave decibel information and the second sound wave decibel information specifically includes determining that an amplitude value of the sound wave spectrum peak-trough difference value exceeds an amplitude threshold value of the sound wave spectrum peak-trough difference value in a conversation process with the conversation party, and determining the call response level accordingly, and/or determining that the sound wave amplitude value exceeds the sound wave decibel amplitude threshold value in a conversation process with the opposite party, and determining the call response level accordingly.
8. The acoustic detection and warning method of claim 1, wherein: in the step (4), the performing of adaptive mode switching operation on the current call mode with the conversation party based on the current call response level of the conversation party specifically includes presetting a plurality of call response levels with different event occurrence probabilities according to a fourth machine learning algorithm model, and setting a plurality of call modes corresponding to the plurality of call response levels one to one according to the fourth machine learning algorithm model, so as to perform corresponding call mode switching operation based on the call response level determined in the current call process with the other party.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910277346.9A CN110062117B (en) | 2019-04-08 | 2019-04-08 | Sound wave detection and early warning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910277346.9A CN110062117B (en) | 2019-04-08 | 2019-04-08 | Sound wave detection and early warning method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110062117A CN110062117A (en) | 2019-07-26 |
| CN110062117B true CN110062117B (en) | 2021-01-08 |
Family
ID=67318477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910277346.9A Active CN110062117B (en) | 2019-04-08 | 2019-04-08 | Sound wave detection and early warning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110062117B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111508529A (en) * | 2020-04-16 | 2020-08-07 | 深圳航天科创实业有限公司 | Dynamic extensible voice quality inspection scoring method |
| CN111599379B (en) * | 2020-05-09 | 2023-09-29 | 北京南师信息技术有限公司 | Conflict early warning method, device, equipment, readable storage medium and triage system |
Family Cites Families (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7940897B2 (en) * | 2005-06-24 | 2011-05-10 | American Express Travel Related Services Company, Inc. | Word recognition system and method for customer and employee assessment |
| CN103701999B (en) * | 2012-09-27 | 2017-03-01 | 中国电信股份有限公司 | The monitoring method of call center's voice call and system |
| JP6341092B2 (en) * | 2012-10-31 | 2018-06-13 | 日本電気株式会社 | Expression classification device, expression classification method, dissatisfaction detection device, and dissatisfaction detection method |
| JPWO2014069444A1 (en) * | 2012-10-31 | 2016-09-08 | 日本電気株式会社 | Dissatisfied conversation determination device and dissatisfied conversation determination method |
| US9191510B2 (en) * | 2013-03-14 | 2015-11-17 | Mattersight Corporation | Methods and system for analyzing multichannel electronic communication data |
| CN105810205A (en) * | 2014-12-29 | 2016-07-27 | 中国移动通信集团公司 | Speech processing method and device |
| US10194027B1 (en) * | 2015-02-26 | 2019-01-29 | Noble Systems Corporation | Reviewing call checkpoints in agent call recording in a contact center |
| KR101681653B1 (en) * | 2015-03-26 | 2016-12-01 | 김윤희 | Device and System for providing phone number service by providing customer's inquiry contents to phone number owner and method thereof |
| CN107305773B (en) * | 2016-04-15 | 2021-02-09 | 美特科技(苏州)有限公司 | Voice emotion recognition method |
| CN106024015A (en) * | 2016-06-14 | 2016-10-12 | 上海航动科技有限公司 | Call center agent monitoring method and system |
| CN107886951B (en) * | 2016-09-29 | 2021-07-23 | 百度在线网络技术（北京）有限公司 | Voice detection method, device and equipment |
| CN107452385A (en) * | 2017-08-16 | 2017-12-08 | 北京世纪好未来教育科技有限公司 | A kind of voice-based data evaluation method and device |
| CN107452405B (en) * | 2017-08-16 | 2021-04-09 | 北京易真学思教育科技有限公司 | Method and device for evaluating data according to voice content |
| CN107464573A (en) * | 2017-09-06 | 2017-12-12 | 竹间智能科技（上海）有限公司 | A kind of new customer service call quality inspection system and method |
| CN109451188B (en) * | 2018-11-29 | 2022-03-18 | 平安科技（深圳）有限公司 | Method and device for differential self-help response, computer equipment and storage medium |
-
2019
- 2019-04-08 CN CN201910277346.9A patent/CN110062117B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110062117A (en) | 2019-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108922538B (en) | Conference information recording method, conference information recording device, computer equipment and storage medium | |
| CN108833722B (en) | Speech recognition method, speech recognition device, computer equipment and storage medium | |
| US8005675B2 (en) | Apparatus and method for audio analysis | |
| US7822605B2 (en) | Method and apparatus for large population speaker identification in telephone interactions | |
| US10289967B2 (en) | Customer-based interaction outcome prediction methods and system | |
| EP0776567B1 (en) | Analysis of audio quality | |
| US5414755A (en) | System and method for passive voice verification in a telephone network | |
| US9837078B2 (en) | Methods and apparatus for identifying fraudulent callers | |
| US9300790B2 (en) | Multi-party conversation analyzer and logger | |
| US20040170258A1 (en) | Predictive dialing system and method | |
| CN110062117B (en) | Sound wave detection and early warning method | |
| CN113411454B (en) | Intelligent quality inspection method for real-time call voice analysis | |
| CN111128241A (en) | Intelligent quality inspection method and system for voice call | |
| CN112261230B (en) | Express call answering rate assessment method, device, equipment, system and storage medium | |
| CN106847305A (en) | A kind of method and device of the recording data for processing service calls | |
| US11545159B1 (en) | Computerized monitoring of digital audio signals | |
| CN111508527B (en) | Telephone answering state detection method, device and server | |
| CN114449105A (en) | Voice-based electric power customer service telephone traffic quality inspection system | |
| Dubey et al. | Comparison of subjective and objective speech quality assessment for different degradation/noise conditions | |
| Audhkhasi et al. | Two-scale auditory feature based non-intrusive speech quality evaluation | |
| Hegde et al. | Voice activity detection using novel teager energy based band spectral entropy | |
| Blatnik et al. | Influence of the perceptual speech quality on the performance of the text-independent speaker recognition system | |
| CN117354421A (en) | Intelligent voice analysis method and system | |
| CN115565523A (en) | End-to-end channel quality evaluation method and system based on neural network | |
| TR2023018175A2 (en) | CALL CENTER WAITING TIME ESTIMATION SYSTEM |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |