• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    A voice quality measurement system and method for testing a voice-oriented network has dialer devices that initiate telephone sample calls through the voice-oriented network underpinning and monitoring sample call system servers. The speech samples are recorded in the dialers, and used for evaluating speech quality. 18
    公开号:SE1000555A1
    申请号:SE1000555
    申请日:2010-05-24
    公开日:2011-11-25
    发明作者:Peter Olov Lager
    申请人:Peter Olov Lager；
    IPC主号:
    专利说明:

    1015202530other region, and a public mobile telephony network (PLMN) ina third region.
    Many factors contribute to the deterioration ofvoice quality through the voice-oriented network, e.g.deficiencies in codec and eco - cancellation, packet loss anddelay, and radio interference. Depending on the specificsvoice-oriented networks used to carrythe telephone call and other factors, then comes the speech qualitylikely to vary.
    From time to time, the speech quality will not satisfy customersexpectations. Therefore, network operators regularly test theirsnetworks and the networks of their partners with respect toon speech quality.
    Several standards have been developed for testingspeech quality. They can mainly be divided into two categoriesknown as non-interfering and interfering, respectively. Non-disturbingtesting uses real traffic as stimuli, and appreciatespassive voice quality. Disturbing testing uses on the other handside a specific test stimulus, and is dependent on onetelephone calls set up specifically for testing. Disturbingtesting generally gives more accurate test results.
    A standard for disruptive testing of voice quality is calledperceptual evaluation of speech quality (PESQ). It is suitablewell for automated voice quality testing invoice-oriented networks.
    PESQ is based on comparison of an original reference andthe same reference after it has been degenerated by thevoice-oriented networks. Voice sample of the originalthe reference is sent to the voice-oriented networks, andvoice samples that have degenerated during their journey through thevoice-oriented networks are received when they leave10152025voice-oriented networks. In short, the reference is onecomposition of samples from different voices. The reference mimicsthe average telephone voice.
    An algorithm produces measurement data regarding speech quality such asis based on a comparison of speech samples with itoriginal reference and speech samples sent through thevoice-oriented networks.
    The output is a value between -0.5 and 4.5 that represents oneestimate of speech quality, where 4.5 is the bestthe speech quality. A calibrated value is given on a scale from 1to 5, and is called the Mean Opinion Score (MOS). The bestthe speech quality is indicated as 5 on the MOS scale.
    PESQ is used in the "Test-Call System" (TCS). TCS is a product ofArptel Ltd .. In TCS, callers controlled by TCS are placedservers in strategic locations, between which telephone callsshall be set up through the voice-oriented networks in order toperform voice quality testing.
    The callers forward automated telephone calls such asrecommended by the TCS servers. The callers can initiateand / or receive telephone calls. In addition, the callers have onevoice playback and recording capability.
    Copies of the original reference are placed in the TCSthe servers and to the callers. During testing with PESQa phone call is set up between two callers. One ofthe callers play their copy of the originalthe reference, i.e. sends speech samples related tothe original reference. The other caller recordsspeech samples after being sent through the voice-oriented onesnetworks, i.e. receives them and stores them.10152025Comparison of speech samples of the copied originalthe reference stored in the TCS servers and the degenerate onesthe voice samples recorded are executed in the TCS servers.
    TCS can control routing. Routes for calls from a callercaller to a receiving caller can be selected byuse ring prefix. Thereby, problems with badvoice quality is located on a specific network ornetwork operator by changing routes.
    One problem is that it is costly and cumbersome tohouse, operate and maintain callers located onfaraway places. To achieve good test coverage is requiredcallers in many different places.
    Another problem is that sample calls are received, i.e. onecallers send "off hook" or equivalent before itsends voice samples through the voice-oriented networks, whichcauses expenses in the form of call charges.
    Brief description of the inventionThe mentioned problems and other problems are solved bypresent invention. This is achieved through a systemand a method for measuring speech quality in a voice-orientednetwork.
    The voice quality measurement system for the voice-oriented networkincludes a perceptual processing deviceevaluation of speech quality, aoriginal reference speech sample memory to store originalreference number sample, a dialer connected to itvoice-oriented network to initiate a phone callthrough the voice-oriented network, a voice recorderdevice, and a comparative sample memory_1015202530Original reference speech sample memory for storing originalreference number sample is adapted to store originalreference number sample of an original reference obtained froma voice messaging device in the voice-oriented network.
    The dialer connected to the voice-orientedthe network to initiate telephone calls through itthe voice-oriented network is adapted to initiate acalls to the voice messaging device in the voice-orientednetwork, by dialing an unassigned subscriber number.
    The voice recorder is adapted to receive and recordvoice samples sent by the voice message device through itvoice-oriented network.
    The comparative sample memory is adapted to store the saidspeech sample recorded by the speech player ascomparative sample in said comparative sample memory.
    The method of performing speech quality measurements in voice-orientednetworking involves several steps. Before testing, select oneobtainable copy of speech sample deemed best, andstored in an origin reference sample in a system forto perform speech quality measurements in the voice-orientednetworks.
    During testing, a phone call from one is initiateddialer through the voice-oriented networks tothe voice messaging device in the voice-oriented networks,by dialing an unassigned subscriber number.
    By using said initiated telephone calls are takenvoice sample received through the voice-oriented networks inthe dialer from the voice message device. Thestored in a comparative sample memory.1015202530Through the system and method of the invention it isprocessing device for perceptual evaluation ofspeech quality adapted to produce speech quality measurement databased on said original reference sample and saidcomparative number sample without the need for remote callersplaces.
    Brief description of drawingsThe invention is illustrated in the accompanying drawings in which:Fig. 1 is a schematic drawing of a speech quality measuring systemin accordance with the invention, connected to a voice-orientednetwork,Fig. 2 is a schematic drawing of a speech quality measuring systemin accordance with the invention, which has asample call system server and dialer, and which isconnected to a voice-oriented network,Fig. 3 is a flow chart of a method according tothe invention.
    Detailed description of the inventionThe invention will now be described in more detail. In Fig. 1shows a speech quality measurement system 10 for a voice orientednetwork 11 in accordance with the invention.
    The voice quality measurement system 10 for the voice-oriented network11 includes an original reference speech sample memory 12, acomparative sample memory 13, a processing means forperceptual evaluation of speech quality 14, adialer 15 and a voice recorder 16. They areinterconnected as shown in the drawing.
    The term voice-oriented network can denote a single onenetworks, or several interconnected networks, e.g.voice-oriented networks managed by each network operator.
    This practice is well known in the art of engineering.1015202530In the voice-oriented network 11 there isvoice messaging devices that notify callers of anumbers would have been dialed that have no subscriberassociated with it. The message is a voice message,i.e. speech sample of a pre-recorded voice sent tothe caller through the voice-oriented network 11. Asuch telephone number is called an unassigned telephone numberhere in.
    In the voice-oriented network 11 isthe voice messaging devices each related to onerespective place. By place is meant here e.g. one geographicallyarea or a voice-oriented network controlled by aspecific network operator.
    Figure 1 shows a voice message device 17.
    The voice message device 17 is generally in the same location as,or near a location indicated by a country code and aarea code of a type of unassigned telephone number.
    Thus, if an unassigned phone number has been dialed soa telephone call is routed through the voice-oriented network11, to a specific voice message device 17 as determinedof the country code and area code dialed. In the absence ofcountry code or area code, the phone call is routed through itvoice-oriented network 11 to a localvoice message device, e.g. the voice message device 17.
    In accordance with the invention, the speech quality measurement system 10 performsfor the voice-oriented network 11 perceptual evaluation(PESQ)of speech quality by using the saidvoice messaging devices.
    The speech samples of the specific voice messages fromthe voice messaging devices of the various voice-oriented onesthe networks are used as original reference numbers and10l5202530comparative number sample. The latter are speech samples used forcomparison with the original reference samples. Thecorresponds to those traditionally sent from callers throughthe voice-oriented network during testing with PESQ.
    This will now be further explained. The originalthe reference is traditionally a specific composition ofspeech samples from different voices, selected to produce itmost accurate speech quality measurement result when using PESQ.
    Since speech samples that can be obtained fromthe messaging devices are only a voice that is not selected forto be representative of any mean they are anywayadequate to perform testing with PESQ.
    Also, because it would be beneficial if perfectcopies of the speech samples of the messaging devices couldobtained, it is adequate to use speech samples asdegenerated through the voice-oriented networks thatoriginal reference, provided that the degeneration is reasonable.
    In general, the speech samples of the pre-recordedvoicemails difficult to obtain in original, but throughManual estimation performed by a person can be an adequate copyobtained by trying different routes through the network andchoose the one that is judged to be the best. Different routes are selected bybeat different routing prefixes.
    Voice quality can vary over time, so it pays tocall several times with one and the same routing prefix. OneLocal calls made in a remote location can provide the bestcopy available. The best obtainable copyselected as the original reference.
    Speech sample of the best obtainable copy is used as oneoriginal reference of the PESQ algorithm, and is thus namedoriginal reference sample and stored in1015202530the original reference speech sample memory. Once an adequate copyhas been obtained from a voicemail device is available ingenerally no need to regularly obtain a new copyfrom the voice message device for use asoriginal reference.
    When the speech quality measurement system 10 performs perceptual evaluationof voice quality (PESQ) it uses the voice messaging device17, by typing a routing prefix, a country code, aarea code and an unassigned phone number.
    The voice message device 17 responds by sending the speech samplesof the voice message through the voice-oriented network 11,to the speech quality measurement system 10.
    The speech recorder 16 of the speech quality measurement system 10 for thevoice-oriented networks 11 record speech samples that are heardto the voicemail. They are then called comparative numbers samplesand stored in the comparison number sample memory 13. The number samples inthe original reference speech sample memory 12 and the speech samples inthe comparative sample memory 13 is processed by the PESQprocessing means 14, and the result is presented as oneMOS value.
    In addition to poor speech quality, a low MOS value may indicateother problems with the voice-oriented network 11, e.g.false answers. False answers are a condition when itvoice-oriented network 11 signals "connect" despitea phone call has not been answered. False answers resultin incorrect invoicing.
    The voice-oriented network 11 has manyvoice messaging devices, many of which do not share one andsame voicemail. Voice mode, frequency spectrum and otherscharacteristics differ in the voice samples of the respectivethe voice messaging devices. It is important that onespeech quality measurement using a specific9lO15202530voice messaging device 17 uses a related originalreference.
    Therefore, the original reference numeral sample memory 12 is ithe speech quality measurement system 10 adapted to include aseveral original references, each attributable to onerespective voice message device 17. Each input, i.e.memory position, in the original reference numeral sample memory 12 isrelated to a respective unassigned telephone numbereach related to a specificvoice message device 17.
    When speech quality measurements are performed by the speech quality measurement system10, the PESQ processing means 14 uses the voice message assent through the voice-oriented network 11 and whichstored in the comparative sample memory 13, and selects itsrelated original reference in a related input oforigin reference number sample memory 12.
    Fig. 2 shows an embodiment of the speech quality measuring system10. A TCS server 21 and a dialer 15 are shown.
    They are interconnected to enable data communicationthrough a data connection 23. The connection is a physical link,a data-oriented network or other meansprovide TCP / IP or other types of data communication.
    The dialer 15 is in turn connected to onevoice-oriented network 11. The dialer 15 is aPOTS dialer, an ISDN dialer, a voice-over-IPcaller or similar device capable ofi.e.initiate calls, make calls.
    The TCS server 21 orders the initiation of a test call toperform a speech quality test using PESQ.
    The dialer responds by initiating the sample call. TheThe exact details of initiating a test call vary101015202530depending on the type of dialer, voice-orientednetwork interfaces and other factors. As an examplesignals a POTS caller “off hook.” It voice-orientedthe network 11, in this example a PSTN network, responds with adial tone indicating that it is ready forthe dialer to send Dual-Tone-Multi-Frequency(DTMF) signals representing digits of aphone number. The telephone number is e.g. a number in an E-l63-or E-164 numbering plan.
    When the POTS caller receives the dial tone, it startssend a DTMF-coded telephone number. Phone number ashas no subscriber associated with it. It issubscriberless. However, the dialed number has onean area code and a non-network prefix, a country code,assigned subscriber number. The network prefix and country code areoptional.
    The phone number is selected so that it is related to onespecific route and remote voice-oriented network location,thus using the network prefix, country code andthe area code in an appropriate manner.
    Most voice-oriented networks 11 respond by settingup a call to a voice messaging device 17 playingup a pre-recorded voice message announcingthe caller that the call cannot be completed because the numberas it is beaten has no subscriber.
    The location of the voice message device 17 is generally closethe one indicated by the country code and area code.
    The dialer 15 continues by recording itpre-recorded the voice message when played byrust message device 17. It pre-recordedthe voicemail sent is used as degeneratelllOl5202530speech sample of a PESQ algorithm. The PESQ algorithm is located inTCS server 21.
    The speech recorder 16 is thus located in the dialer15, and the recorded speech samples are sent as data overthe data connection 23 to the TCS server 21. The TCS server 21stores the recorded speech samples in the comparative speech sample memory 13located therein.
    An original reference is stored in advance, i.e. before testingin the TCS server 21. It is stored inthe original reference speech sample memory 12. The originalthe reference is a speech sample of the same voicemail atthe voice message device 17. The difference is thatthe reference sample has been evaluated by one person and assessedvoice with no onerepresent high quality voice, i.e.or slight degeneration.
    Voice samples of the pre-recorded message inthe voice message device 17 is not an equivalent compensationfor the reference traditionally used. Nor is itdegenerate copy used as original reference equalas the original pre-recorded voice message inthe voice message device 17 in the voice-oriented network11.Still, the PESQ algorithm performs quite well, but not quiteas accurately as with the previously mentioned speech samples asrepresents the average voice.
    There are several benefits. It performs better than non-interfering test of speech quality, e.g. NIQA. In addition, cansavings are made in operation and maintenance andaccommodation costs as there is no need fordialers in remote locations. In addition, there are nonel210l52025call charges, as the calls are not consideredanswered.
    Test coverage is greatly improved because it is notpractically possible to have callers in as many places asthere are voicemail devices of the aforementioned type.
    The previously mentioned functions can be distributed in several wayswithout departing from the scope and spirit of the invention.
    Figure 3 shows a method in accordance with the invention. Beforespeech quality measurement, a obtainable copy of is storedvoice sample from a voice message device 17 ithe original reference speech sample memory 12. The stored copy is onewhich is judged to be the best. It is seen in the first box infigure 3.
    The dialer 15 dials an unassignedphone number. This is seen in the second box in Figure 3.
    The voice message device 17 of the voice orientedthe network responds by sending a voice message through itvoice-oriented network 11. This is a comparative sample.
    The voice samples of the voice message are received inthe speech quality measurement system 10. This is shown in the third box infigure 3.
    Comparative number samples received in the speech quality measurement system10 is stored in the original reference speech sample memory 12. This is seen inthe fourth box in Figure 3.
    The original reference number samples in the original number sample memory12 and the comparative sample in the comparative sample memory 13compared using perceptual evaluation ofspeech quality. It is seen in the fifth box in Figure 3.l3
    权利要求:
    Claims (7)
    [1]
    A speech quality measuring system (10) (11), for a voice-oriented network comprising - a processing means for perceptual evaluation of speech quality (14), - an original reference speech sample memory (15) (12), - a dialer connected to said voice-oriented network for initiating a telephone calls through the voice-oriented network (II), - a speech recorder (16), (13), - said original reference sample memory characterized by: (12) - a comparative speech sample memory adapted to store original reference speech samples of an original reference obtained from a voice message device ( 17) in the voice-oriented network (11), - that said dialing device (15) connected to said voice-oriented network (11) for initiating a telephone call through the voice-oriented network (11) is adapted to initiate a call to said voice-message device (11). 17) (11), through the voice-oriented network by dialing an unallocated subscriber number, - that said speech recorder (16) is adapted to receive and record speech samples sent by said voice message device (17) (11), through the voice-oriented network - that said comparison speech sample memory (13) is adapted to store said speech sample recorded by said speech recorder (16). ) as a comparative sample in said comparative sample memory, - that said processing means for perceptual evaluation of speech quality is adapted to produce speech quality measurement data based on said original reference sample without the need for callers in remote locations.
    [2]
    A speech quality measuring system (10) (II) for a voice-oriented network according to claim 1, characterized by: - said original reference speech sample memory (12) for storing original reference speech samples is further adapted to include speech samples of a plurality of original references, each related to a respective voice message device, - that said speech quality measuring system (10) is further adapted to associate said included original speech samples from each of said plurality of original references related to a respective voice message device with an unassigned telephone number, whereby said means for perceptual evaluation of voice quality selects the appropriate original reference based on the unallocated phone number dialed.
    [3]
    The voice quality measuring system (10) for voice-oriented networks (II), characterized by: (15) according to claim 2, - said dialer being adapted to record voice samples without answering calls, thereby avoiding call charging.
    [4]
    The voice quality measurement system for voice-oriented networks (II) (10) according to claim 3, characterized by: - said dial-up device (15) being adapted to initiate a call to a specific voice-message device in the voice-oriented networks (II), by dialing an area code. related to voice-oriented networks at a specific location, whereby speech quality measurement data can be obtained relating to voice-oriented networks (II) at a specific location. 15 10 15 20 25 30
    [5]
    The voice quality measurement system (10) for voice-oriented networks (11) characterized by: (15) according to claim 4, - said dialer being adapted to initiate a call through specific voice-oriented networks (11), by dialing a routing prefix related to the specific voice-oriented networks, whereby speech quality measurement data can be obtained relating to voice-oriented networks in a route to a specific location.
    [6]
    The speech quality measuring system (10) for measuring speech quality in (11) characterized by: (12), voice-oriented networks according to claim 1, - said original reference speech sample memory (13) comparative speech sample memory and processing means for perceptual evaluation of speech quality (14) are included in a sample call server (21), - that said speech recorder (16) is included in said calling device (15), - that said sample call system server (21) and said calling device (15) are interconnected for data communication by TCP / IP, whereby said speech quality measuring system (10) for a voice-oriented network (10) is divided into practical structural units.
    [7]
    A method of performing speech quality measurements in a voice-oriented network (11), characterized by the following method steps: - before testing, selecting from a voice message device (17) in the voice-oriented network (11) a obtainable copy of speech sample which is judged to be best, and storing it in a (12) original reference speech sample memory in a voice quality measuring system (10) for the voice-oriented network (II), - during testing, initiate a telephone call from a dialer (15) through the voice-oriented network (11) to the voice message device (17) in the ( ll), voice-oriented network by dialing an unallocated subscriber number, - receiving voice samples from the voice message device (17) sent through the voice-oriented network (ll) in the dialer (15) using said initiated telephone calls, - storing said voice samples sent through the voice-oriented network (II) (13), in a comparative sample memory whereby the processing means for perceptual In the evaluation of speech quality (14) are adapted to produce speech quality measurement data based on said original reference number sample and said comparative number sample without the need for callers in remote locations. 17
    类似技术:
    公开号 | 公开日 | 专利标题
    US9942406B2|2018-04-10|Method and apparatus for managing telephone calls
    US8379818B2|2013-02-19|Method and apparatus for conveying a calling party identifier
    US5933475A|1999-08-03|System and method for testing a telecommunications apparatus
    US8712029B2|2014-04-29|Method and system for routing telephony communications together with modified calling party identifier information
    US8078155B2|2011-12-13|Call processing for group conferencing
    US8693671B2|2014-04-08|Network predictive customer service queue management
    US20060188081A1|2006-08-24|Systems and methods for call measurement
    US5479473A|1995-12-26|Method and apparatus for testing call paths over long distance carrier networks
    CN101662544A|2010-03-03|Method, device and system for external-call processing
    US5875230A|1999-02-23|Interactive measurement system and method for telecommunication networks
    CN101827176A|2010-09-08|Method and equipment for a user to call back
    US8942670B1|2015-01-27|Carrier network determination for providing real-time customer feedback of call information
    SE1000555A1|2011-11-25|Voice quality measurement system and method for a voice-oriented network
    US8477909B2|2013-07-02|Testing call routing
    KR101051723B1|2011-07-25|Internet extension telephone apparatus and method for transmitting an existing extension telephone, and computer-readable recording medium accordingly
    WO2016202144A1|2016-12-22|Media channel testing method and device
    SE0900802A1|2010-12-13|Systems for measuring telecommunication quality with operator-common test equipment
    US9106755B2|2015-08-11|Method and system for a gateway transfer
    CN105611082A|2016-05-25|Network voice testing system
    WO2007148037A1|2007-12-27|Communications terminal with preference information relating to a messaging system
    US20210067626A1|2021-03-04|Methods and systems for placing calls
    US20110051905A1|2011-03-03|Method of and a device for quantifying a tone signal received in a telecommunications network
    JP3118111B2|2000-12-18|Telephone exchange
    JP2008011579A|2008-01-17|Method and program for public network connection
    RU2005140545A|2006-07-27|SYSTEM AND METHOD FOR PROVIDING TONAL SIGNALS FOR RETURNING A CALL TO A COMMUNICATION NETWORK
    同族专利:
    公开号 | 公开日
    SE534860C2|2012-01-24|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    CN107221339A|2017-05-22|2017-09-29|华北电力大学|Based on gain compensation audiphone voice quality W PESQ method for objectively evaluating|
    CN109767786A|2019-01-29|2019-05-17|广州势必可赢网络科技有限公司|A kind of online voice real-time detection method and device|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    SE1000555A|SE534860C2|2010-05-24|2010-05-24|Voice quality measurement system and method for a voice-oriented network|SE1000555A| SE534860C2|2010-05-24|2010-05-24|Voice quality measurement system and method for a voice-oriented network|
    [返回顶部]