CN111816209A - Voice switching method and device based on voice quality evaluation - Google Patents
Voice switching method and device based on voice quality evaluation Download PDFInfo
- Publication number
- CN111816209A CN111816209A CN202010578252.8A CN202010578252A CN111816209A CN 111816209 A CN111816209 A CN 111816209A CN 202010578252 A CN202010578252 A CN 202010578252A CN 111816209 A CN111816209 A CN 111816209A
- Authority
- CN
- China
- Prior art keywords
- voice
- voice quality
- quality evaluation
- average
- switching
- 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
- 238000013441 quality evaluation Methods 0.000 title claims abstract description 77
- 238000000034 method Methods 0.000 title claims abstract description 26
- 230000005540 biological transmission Effects 0.000 claims description 29
- 238000001303 quality assessment method Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 abstract description 10
- 238000010586 diagram Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
Classifications
-
- 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
- G10L25/60—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 measuring the quality of voice signals
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention belongs to the field of communication, and relates to a voice switching method and a voice switching device based on voice quality evaluation, wherein the method comprises the following steps: 1) establishing a voice quality evaluation model by using a voice quality evaluation module; 2) embedding the voice switching module in the intelligent terminal; 3) and (2) evaluating the current voice quality in real time based on the voice quality evaluation model obtained in the step 1), and implementing voice switching operation according to the current voice quality. The invention provides a voice switching method and a voice switching device based on voice quality evaluation, which can effectively and accurately evaluate voice quality, intelligently switch voice in time and ensure normal communication.
Description
Technical Field
The invention belongs to the field of communication, relates to a voice switching method based on network transmission, and particularly relates to a voice switching method and device based on voice quality evaluation.
Background
With the development of communication technology, the existing network transmission can be realized through wired broadband, mobile communication and satellite communication, the actual transmission efficiency is different under different application scenes, and particularly, the normal operation of communication needs to be ensured in large-scale emergency activities. The voice service is one of the most basic and important services in network transmission, and can be used for performing voice switching in time under different conditions, so that normal voice communication is guaranteed, the working efficiency can be effectively improved, and the method has important significance for maintaining public safety.
The voice switching time also needs a certain basis, and it is a common practice to decide whether to switch voice according to the change of voice quality, and when the voice quality is reduced, the voice transmission channel is replaced to ensure the normal operation of voice. There are also many methods for evaluating speech quality, subjective evaluation is given, such as ACR measurement method, objective computer-aided analysis of audio waveform is available, such as PSQM and PESQ method, and speech quality evaluation based on signal strength is also one of the mainstream methods because the most influential factor of speech quality is network transmission.
Disclosure of Invention
In order to solve the technical problems in the background art, the invention provides a voice switching method and device based on voice quality evaluation, which can effectively and accurately evaluate voice quality, perform voice switching intelligently in time and ensure normal communication.
In order to achieve the purpose, the invention adopts the following technical scheme:
a voice switching method based on voice quality assessment is characterized in that: the voice switching method based on voice quality evaluation comprises the following steps:
1) establishing a voice quality evaluation model by using a voice quality evaluation module;
2) embedding the voice switching module in the intelligent terminal;
3) and (2) evaluating the current voice quality in real time based on the voice quality evaluation model obtained in the step 1), and implementing voice switching operation according to the current voice quality.
The specific implementation manner of the step 1) is as follows:
1.1) a voice quality evaluation module establishes a voice quality evaluation test model;
1.2) training and testing a voice sample, performing voice quality evaluation and obtaining a voice quality evaluation score;
1.3) optimizing parameters of the voice quality evaluation test model according to the scores obtained by the voice quality evaluation, and giving a final voice quality evaluation model.
The specific implementation manner of the step 1.1) is as follows: building a voice transmission environment model, wherein the voice transmission environment model comprises a collection end for collecting original voice, a simulator for simulating a real network transmission environment and a playing device for receiving and playing the transmitted voice; and the acquisition end transmits the acquired original voice to the playing equipment through the simulator.
The specific implementation manner of the voice quality evaluation in step 1.2) is to evaluate the uniformity and continuity degree of the actually received audio data through the data amount and standard deviation of the audio packets per second.
The specific formula of the voice quality evaluation in the step 1.2) is as follows:
QL＝(E(Average)-Average)+alpha*SD
wherein:
QL is the voice quality loss;
average is the audio packet sent by the acquisition end every second;
e (average) is the average mathematical expectation of audio packets received by the receiver per second;
average is the Average number of actually received audio packets per second;
alpha is a parameter to be determined;
SD is a standard deviation describing the degree of uniformity of audio packet reception.
The specific implementation manner of implementing the voice switching operation according to the voice quality in the step 3) is as follows: after receiving the voice, the receiving end obtains a corresponding evaluation score through a voice quality evaluation system, and if the score is a high score, the current voice transmission channel is kept unchanged; if the score is low, switching the voice transmission channel, and continuing to evaluate the voice quality of the new transmission channel.
An apparatus for implementing the voice handover method based on voice quality assessment as described above is characterized in that: the device comprises a voice quality evaluation module and a voice switching module connected with the voice quality evaluation module; the voice quality evaluation module is used for constructing a voice quality evaluation model; the voice quality evaluation model evaluates the current voice quality in real time, and the voice switching module implements voice switching operation according to the current voice quality.
The invention has the advantages that:
the invention provides a voice switching method and device based on voice quality evaluation. The voice quality evaluation algorithm can effectively and accurately evaluate the voice quality, and intelligently switches voices in time to guarantee normal communication.
Drawings
FIG. 1 is a schematic structural diagram of a voice switching apparatus based on voice quality evaluation according to the present invention;
FIG. 2 is a schematic diagram of a model of a speech transmission environment constructed by the present invention;
fig. 3 is a flow chart of a speech quality assessment method.
Detailed Description
The invention provides a voice switching method based on voice quality evaluation, which has the following theoretical basis: the key point of the evaluation of the voice quality evaluation algorithm is whether the voice transmitted by the network is degraded or not and to what extent, and then how to fit the voice quality evaluation algorithm with the algorithm model to obtain the voice quality loss, and then voice switching is carried out according to the loss extent.
The voice switching method based on voice quality evaluation provided by the invention firstly simulates the real network transmission condition through the voice quality evaluation test module to obtain a plurality of groups of sample data, analyzes the corresponding voice loss and determines the final voice quality evaluation parameter. And then applying the model to a real application scene, acquiring the corresponding voice loss in real time, evaluating the voice quality of the model and determining whether to perform voice switching.
The voice switching device based on voice quality evaluation provided by the invention is composed of a voice quality evaluation module and a voice switching module, as shown in fig. 1. The voice quality evaluation module comprises a voice quality evaluation model establishment module, a training test voice sample, a voice quality evaluation model parameter optimization module and a final voice quality evaluation model, wherein the voice quality evaluation model parameter optimization module is used for optimizing the voice quality evaluation model parameter according to the score obtained by voice quality evaluation. And the voice switching module is embedded in the intelligent terminal, and the current voice quality is evaluated in real time based on the final voice quality evaluation model to implement voice switching operation.
1) Voice quality evaluation module
In order to obtain the speech quality evaluation model, a speech transmission environment needs to be established, as shown in fig. 2, including a device for collecting original speech, a simulator for simulating a real network transmission environment, and a device for receiving and playing the transmitted speech. According to the sampling theorem, after the analog audio signal is converted into the digital signal, the analog audio signal can be a uniform and continuous audio data packet, when the transmission network condition is very good, the uniform and continuous data packet generated by the acquisition end still can be uniform and continuous when the acquisition end reaches the playing end, the voice quality at this time can also be as good as that of the acquisition end, but because the voice generally has the conditions of packet loss, time delay and jitter after passing through the transmission network, the model is used for simulating the real network transmission condition by setting 3 parameters of the packet loss, the time delay and the jitter in advance (wherein the packet loss, the time delay and the jitter all belong to the prior art).
The flow chart of the speech quality assessment method is shown in fig. 3. And after the audio sample data reaches a receiving end through the network simulator, the operation is carried out in two ways, firstly, the audio sample data is processed according to the received data information, corresponding characteristics are extracted to establish a voice quality evaluation model, and secondly, a result scoring system gives corresponding voice quality scores according to the voice played after the audio sample data is received. And then, optimizing the parameters of the evaluation model according to the scores to obtain a final voice quality evaluation model.
The voice quality evaluation model evaluates the uniformity and continuity of actually received audio data through the data volume and standard deviation of each second of audio packets, thereby achieving the purpose of evaluating voice quality. The specific formula is as follows:
QL＝(E(Average)-Average)+alpha*SD
QL: loss of speech quality;
average: collecting audio packets sent by a terminal every second;
e (average): the average value mathematical expectation of the audio packets received by the receiving end every second;
average: average number of actually received audio packets per second;
alpha: a parameter to be determined;
SD: a standard deviation describing how evenly the audio packets are received;
the acquisition end knows the number of audio packets sent per second, namely Average, and inputs the audio packets into the model; the receiving end statistically analyzes the obtained audio data to obtain E (average) and SD; substituting the data into the above formula can obtain the corresponding voice quality loss QL, and the QL is an expression about the alpha, and the confirmation of the alpha needs to be optimized by a scoring system.
The voice quality scoring system scores received voice according to an Absolute Category scoring (ACR) which is a currently widely applied voice quality subjective evaluation method (see table 1), and specifically defines three scores: high score, low score, middle score. The high score corresponds to a MOS value of 3.5 or more, and is considered to be a relatively good voice quality. The low score corresponds to a MOS value of 3.0 or less, which is considered to be poor voice quality; the middle part corresponds to an MOS value of 3.0 to 3.5 minutes; the intermediate score is introduced to avoid system switching oscillation possibly caused by frequent change of quality evaluation results.
TABLE 1 Voice scoring Table
| Audio level | MOS value | Evaluation criteria |
| Superior food | 4.0～5.0 | Very well, clearly audible; small delay and smooth communication |
| Good wine | 3.5～4.0 | Slightly worse, it is clearly heard; small delay, under-smooth AC, somewhat noisy |
| In | 3.0～3.5 | Also, it is not too clear to hear; with a certain delay, can exchange |
| Difference (D) | 1.5～3.0 | It is not very clear to hear; the delay is large, and the exchange needs to be repeated for a plurality of times |
| Bad quality | 0～1.5 | Extremely bad and incomprehensible; large delay and unsmooth AC |
For the same audio sample, different network simulation parameters are used to simulate various different network conditions to obtain corresponding voice quality loss and voice quality scores, the two parameters are combined to select an alpha parameter, and QL critical values of high score, middle score and low score are determined to obtain a final model. Therefore, the input of the voice quality evaluation model is the voice of the receiving end, the output is the corresponding evaluation score, and only the high score and the low score exist.
2) Voice switching module
After receiving the voice, the receiving end obtains a corresponding evaluation score through a voice quality evaluation system, and if the score is a high score, the current voice transmission channel is kept unchanged; if the score is low, the voice transmission channel is switched, and the voice quality evaluation is continuously carried out on the new transmission channel, so that the purpose of guaranteeing the voice transmission quality is achieved.
Claims (7)
1. A voice switching method based on voice quality assessment is characterized in that: the voice switching method based on voice quality evaluation comprises the following steps:
1) establishing a voice quality evaluation model by using a voice quality evaluation module;
2) embedding the voice switching module in the intelligent terminal;
3) and (2) evaluating the current voice quality in real time based on the voice quality evaluation model obtained in the step 1), and implementing voice switching operation according to the current voice quality.
2. The voice switching method based on voice quality assessment according to claim 1, characterized in that: the specific implementation manner of the step 1) is as follows:
1.1) a voice quality evaluation module establishes a voice quality evaluation test model;
1.2) training and testing a voice sample, performing voice quality evaluation and obtaining a voice quality evaluation score;
1.3) optimizing parameters of the voice quality evaluation test model according to the scores obtained by the voice quality evaluation, and giving a final voice quality evaluation model.
3. The voice switching method based on voice quality assessment according to claim 2, characterized in that: the specific implementation manner of the step 1.1) is as follows: building a voice transmission environment model, wherein the voice transmission environment model comprises a collection end for collecting original voice, a simulator for simulating a real network transmission environment and a playing device for receiving and playing the transmitted voice; and the acquisition end transmits the acquired original voice to the playing equipment through the simulator.
4. The voice switching method based on voice quality assessment according to claim 3, characterized in that: the specific implementation manner of the voice quality evaluation in the step 1.2) is to evaluate the uniformity and continuity degree of actually received audio data through the data volume and standard deviation of each second of audio packets.
5. The voice switching method based on voice quality assessment according to claim 4, characterized in that: the specific formula of the voice quality evaluation in the step 1.2) is as follows:
QL＝(E(Average)-Average)+alpha*SD
wherein:
QL is the voice quality loss;
average is the audio packet sent by the acquisition end every second;
e (average) is the average mathematical expectation of audio packets received by the receiver per second;
average is the Average number of actually received audio packets per second;
alpha is a parameter to be determined;
SD is a standard deviation describing the degree of uniformity of audio packet reception.
6. The voice switching method based on voice quality evaluation according to claim 1 or 2 or 3 or 4 or 5, characterized in that: the specific implementation manner of implementing voice switching operation by the voice quality in the step 3) is as follows: after receiving the voice, the receiving end obtains a corresponding evaluation score through a voice quality evaluation system, and if the score is a high score, the current voice transmission channel is kept unchanged; if the score is low, switching the voice transmission channel, and continuing to evaluate the voice quality of the new transmission channel.
7. An apparatus for implementing a voice handover method based on voice quality assessment according to any of claims 1-6, characterized in that: the device comprises a voice quality evaluation module and a voice switching module connected with the voice quality evaluation module; the voice quality evaluation module is used for constructing a voice quality evaluation model; the voice quality evaluation model evaluates the current voice quality in real time, and the voice switching module implements voice switching operation according to the current voice quality.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010578252.8A CN111816209A (en) | 2020-06-19 | 2020-06-19 | Voice switching method and device based on voice quality evaluation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010578252.8A CN111816209A (en) | 2020-06-19 | 2020-06-19 | Voice switching method and device based on voice quality evaluation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111816209A true CN111816209A (en) | 2020-10-23 |
Family
ID=72845461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010578252.8A Pending CN111816209A (en) | 2020-06-19 | 2020-06-19 | Voice switching method and device based on voice quality evaluation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111816209A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117577098A (en) * | 2024-01-15 | 2024-02-20 | 云天智能信息(深圳)有限公司 | Voice communication method and system for satellite broadband short message communication |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101500131A (en) * | 2009-03-09 | 2009-08-05 | 深圳市源富创新电子有限公司 | Audio and video radio transmission system and transmission method |
| CN102195806A (en) * | 2010-03-10 | 2011-09-21 | 通用电气公司 | Handoff metric for multiple transmission technologies |
| CN103023630A (en) * | 2013-01-11 | 2013-04-03 | 中国人民解放军国防科学技术大学 | Method for hiding information of speech stream on basis of speech coding by pulse code modulation |
| CN104575521A (en) * | 2014-12-26 | 2015-04-29 | 大连理工大学 | Method for evaluating voice quality of LTE communication system |
| US9876901B1 (en) * | 2016-09-09 | 2018-01-23 | Google Inc. | Conversational call quality evaluator |
| CN109413701A (en) * | 2017-08-15 | 2019-03-01 | 中国移动通信集团公司 | A kind of VoLTE business switch method and system |
-
2020
- 2020-06-19 CN CN202010578252.8A patent/CN111816209A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101500131A (en) * | 2009-03-09 | 2009-08-05 | 深圳市源富创新电子有限公司 | Audio and video radio transmission system and transmission method |
| CN102195806A (en) * | 2010-03-10 | 2011-09-21 | 通用电气公司 | Handoff metric for multiple transmission technologies |
| CN103023630A (en) * | 2013-01-11 | 2013-04-03 | 中国人民解放军国防科学技术大学 | Method for hiding information of speech stream on basis of speech coding by pulse code modulation |
| CN104575521A (en) * | 2014-12-26 | 2015-04-29 | 大连理工大学 | Method for evaluating voice quality of LTE communication system |
| US9876901B1 (en) * | 2016-09-09 | 2018-01-23 | Google Inc. | Conversational call quality evaluator |
| CN109413701A (en) * | 2017-08-15 | 2019-03-01 | 中国移动通信集团公司 | A kind of VoLTE business switch method and system |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117577098A (en) * | 2024-01-15 | 2024-02-20 | 云天智能信息(深圳)有限公司 | Voice communication method and system for satellite broadband short message communication |
| CN117577098B (en) * | 2024-01-15 | 2024-04-19 | 云天智能信息(深圳)有限公司 | Voice communication method and system for satellite broadband short message communication |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108389592B (en) | Voice quality evaluation method and device | |
| CN108346434B (en) | Voice quality assessment method and device | |
| CA2334906C (en) | Method for executing automatic evaluation of transmission quality of audio signals | |
| CN103632680A (en) | Speech quality assessment method, network element and system | |
| CN103249061B (en) | Voice quality method of testing and system | |
| CN102685789B (en) | Method for evaluating QoE (Quality Of Experience) of voice service user perception experience by simulating user behaviors | |
| CN102075988A (en) | System and method for locating end-to-end voice quality fault in mobile communication network | |
| CN104581758A (en) | Voice quality estimation method and device as well as electronic equipment | |
| CN105282347A (en) | Method and device for evaluating voice quality | |
| CN1327407C (en) | Method and device for determing quality of speech signal | |
| CN109451303A (en) | A kind of modeling method for user experience quality QoE in VR video | |
| CN105529037A (en) | Communication equipment voice quality evaluation testing system and testing method | |
| CN111816209A (en) | Voice switching method and device based on voice quality evaluation | |
| CN106803797A (en) | A kind of VoIP speech quality method for objectively evaluating based on network parameter | |
| CN112272393A (en) | Method for intelligently switching networks of mobile Internet of things platform | |
| CN100499694C (en) | Method and device for testing speech quality | |
| US7050924B2 (en) | Test signalling | |
| GB2407952A (en) | Quality assessment tool | |
| EP1443496B1 (en) | Non-intrusive speech signal quality assessment tool | |
| Mohamed et al. | A method for quantitative evaluation of audio quality over packet networks and its comparison with existing techniques | |
| CN117061378A (en) | Voice call quality detection method | |
| CN116597856A (en) | Voice quality enhancement method based on frogman intercom | |
| CN116933620A (en) | Noise quality evaluation and model building method | |
| CN103079218B (en) | Auto-dial testing control method and equipment | |
| CN106714226A (en) | Voice quality evaluation method, device and 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 |