TW479429B - System and method for distributing telephone audio data via a computer network - Google Patents

Abstract

A method for streaming telephone audio data over a computer network, the method comprising: receiving the audio data via a telephone call from an author of the audio data; directly encoding the audio data into a destination file format; and storing the encoded audio data on a file system accessible by a computer.

Description

^ + / ^ 429 ^ + / ^ 429 V. Description of the invention d) The applicant hereby submits the provisional application for the benefit of the application. The owner of this patent document has no objection to the cause, the same, but the invention is particularly relevant here. The Internet There is a lot of communication media, but they are in print. The issue is in the broadcast, regardless of the text, as a third-party broadcast, and its connection with the telephone network has been achieved. It can allow him to enter the Internet. The other allows the same to be used to distribute his application No. 60/154, 769 filed on September 20, 1999 and No. 60/179, 831 filed on February 2, 2000. Here are two cases Incorporated as a reference in the current cost application. The public part of the copyright notice has issues related to copyright protection. Copyright Copy production of any patent document or patent issuance, as it appears in the Patent and Trademark Office patent file or records, and in all other respects all copyright rights are reserved. Background of the invention: Secondly, it is generally related to the voice data stream system. The system and method of the stream voice data of this road. Kushiro reaches the client through a computer network. It has been reported that it may be the heaviest since the plate printing appeared The Internet provides us all around the world to publish their works and teach them ^ ^ _ _ — a brand new threshold, including editors :: t, unified publishing road rises: the needs of the 2 divisions. After that, the authors can let them settle directly and let them resolve the issue in a way that the publisher cannot accept disputes or similar problems. Similar restrictions are also imposed on the field of publication presented by the audio. Audio: The author of the video ::, uses the _ π _ # and the restriction of the author of the roar. _This deduction / deduction does not need to be received from us; second, the use of public radio / electricity, and at the same time the use of the public Internet 4;

479429 V. Description of the invention (2) After omitting the traditional wide-ranging restrictions, such as the scalability and the ability to communicate through the form of speech to the technology, the computer has many transmissions, but there are unreceivable languages.

Boxπ, data with telephone, Internet distribution. And because of other problems put into the prefetcher and some data companies. Lado State This is a network, and it can't be resolved at this time. When accessing directory information and epilogues. At best, scrap merchants impose restrictions on Corolla's infrastructure and its many restrictions. Applications such as personal newsletters can be eliminated in eliminating these entries. When this ubiquitous network of public and small telephones is combined, individuals and public telephones are used by simple telephone equipment that are used by the world to transmit essentially different computer users. Indeed, computer and voice applications are far-reaching. Receiving voice data in different ways to reconnect over the computer network. The system is unable to provide real-time and required broadcasts or provides audio quality issues. One method that allows computer systems to receive incoming audio data is to use the so-called "Gentner compound speech coupler. This box is used as an interface between computers. One input on the box is sent and received from the computer and one input is connected to the electronic court. Network. Data received from the Gentner Box is transmitted to the computer, coding and solutions involve the disadvantages of having only one connection at a time and coding one operation. The case has recorded multiple data streams from the telephone network at the same time. To the voice data, the system will save it as a file and in general. Generally speaking, it is read and sent to the encoder by using UNIX cr0nj〇b. Although this setting allows multiple users to communicate with each other, it is not real-time. This is transmitted through a computer network, and the system is only suitable for on-demand voice data transmission. There is a solution to this problem. For example, Tel ISoft Inc., located in Corodo Spring, has developed a system, '

479429 V. Description of the invention (3) It is used to encode and disseminate the voice data received from the telephone network. D e 1 1 S 〇 f t provides a "one" solution to this problem, using many services that must cooperate to perform the required function. This lack of insulation will limit the scalability that allows the system to receive and encode multiple voice signals with optimal quality. Therefore, when the quality of the voice is an important consideration when recording, this solution is not suitable. No. 5,675,507 (‘507)’ is entitled “Message storage and

The "Payment System" US patent provides another solution to this problem. The '507 patent discusses a system and method for receiving and diffusing fax, voice, and data messages. The system retrieves and stores the data in νχ or AD / pCM format. In the =, these formats are a compressed pCM data format. Then the compressed data will be converted into the AU * WAV format according to the user's preferences. 507 special, by using VOX or AD / PCM, etc. The intermediary file serves as the basis for conversion and will provide the worst voice quality. A system and method is needed to encode and disseminate telephone voice data in an efficient, instant and timely manner, and the voice quality is compared with that provided by the system It looks like a big difference. The purpose of this invention is to provide a system and method for disseminating phonetic materials through a computer network.

—Another purpose of the present invention is to provide a system and method for processing voice data according to a set of predetermined preferences. Another objective of Passing January is to provide a system and method for network clients to disseminate telephone voice data on the Internet. Receive private live sound data and directly encode the data into a purpose photo

479429 V. Description of the invention (4) A system that does not require intermediate voice data conversion (that is, does not require a file format) can achieve the above and other objectives. : Management produces considerable improved speech quality. There is also a system and method at the star connection code that can achieve some of the objects of the present invention. The system and method accepts telephone calls from users, and other computer networks (such as the Internet). , Or both. Dangtong Road Le Tong 3 = Get the user's personal profile information in the Tong Tong Tribune. The user profile contains information that the system uses to decide what to do with the voice data.

Such as whether it should be transmitted by a specific stream voice server in real time and / or compressed on the server, the encoder used to compress the data, the compressed file name, etc. The retrieved data will be formatted according to the xml (extendable standard ^ σ) rule ', which defines the format that the data must be attached for the system to continue to execute. In addition, the present invention can also use other proprietary message formatting rules.

e XML description slot and voice data will lead to voice routing software, which will send the incoming data stream to the appropriate encoder as defined by the 3 user's X ML description file. After the data is routed to the appropriate encoder, it is compressed according to one of many available code algorithms. Sample compression / decompression programs used by the system include QuickTime and RealAudio. These data will be compressed or placed in the file system completely according to the data in the user's XML description file, compressed or transmitted by the data stream voice server. The 5H system also has a workflow and publisher software. These two subsystems can obtain voice from the system database according to the user id included in the workflow.

479429

V. Description of the invention (5) The text address of the material and pass it to any destination by any means: such as: write the location into the email and send it to the user. another. ^ "Transfer Λ ΓΛ text into a building case and transfer it to a specific directory of Kushiro server via FTP (slot case transmission communication protocol =). The feeder was read from the coffin case. : Pull and embed it as a web page link according to the user's request. The link will be activated and the voice will be played to the requested place.

The media file format between receiving and encoding points is no longer required. It can enhance system performance. After receiving the original PCM data from the telephone network, you can directly encode it into a destination file format (such as ReaUudi 0, = 2 ^ or% 11 ^ 0 ...), so you can get excellent = ^^ quality. These directly encoded voice files will then be sent to the streaming voice client over the computer network in real time or as required.

: 2 The system load is distributed among software running on different computers. The workload on any computer can also improve performance. In addition, because it is privately assigned to the computer running the backup software, it can be easily brought into the processor, so the group design is also extensible. For example, if a large number of users want to encode in QuickTime mode, more QuickTime encodings can be brought online. Furthermore, at the same time, the incoming voice data stream is routed to a plurality of j-coders at the same time, which can quickly complete the processing. By directing the voice files to a specific server and cluster, the system can encode them in situ, where a large amount of data can be generated. Second, the encoded and compressed files are distributed to the server 叩 around the network. Combining these benefits will provide enhanced performance and the best utilization of hardware resources.

479429 V. Description of the invention (6) Brief description of the drawings The present invention will be described with reference to the drawings. Its purpose is only for illustration and not limitation. The same reference numbers are used to indicate the same or corresponding parts. Among them: FIG. 1 is a diagram showing settings of many software and hardware components of the system according to a specific embodiment of the present invention; FIG. 2 shows specific embodiments of the software and hardware of the present invention performed by setting as an application service provider (ASP) Fig. 3 shows a specific embodiment of the software and hardware embodiment of the present invention that is set to be executed as an internal service of the company; Fig. 4 shows a step involved in encoding and presenting phone voice data according to a specific embodiment of the present invention High-level browsing flowchart; Figure 5 is a flowchart showing the process performed by the intermediary voice system to accept and process incoming calls according to a specific embodiment of the present invention; Figure 6 is a diagram showing the intermediary voice according to a specific embodiment of the present invention The flow chart of the continuous process of the system for accepting and processing incoming calls is shown in FIG. 7. FIG. 7 is an XML description file according to a specific embodiment of the present invention. FIG. 8 is According to a specific embodiment of the present invention, a flowchart of a process for receiving speech data and transmitting it to an appropriate encoder for compression is described. FIG. 9 is a description of encoding speech data into QuickT according to a specific embodiment of the present invention. Flow chart of the process of i meTM Rea 1TM or Windows Med i aTM format; FIG. 10 is a flowchart describing a subroutine for initializing a <311 1 ^^ 6 ^ encoder;

Page 9 479429 V. Description of the invention (7) Fig. 11 is a description diagram; Fig. 12 is a description flowchart; Fig. 13 is a coded · wav format; Fig. 14 is a description; Fig. 15 is a high-level browser according to the process; Figure 16 is a special chart published according to the data. Figure 17 is a file based on the voice data. Figure 18 is a flowchart of adding and editing according to the device. The implementation of the RealAudi〇TM encoder ^ Windows Medidia ™ The process of the present invention will present the present invention at a glance The present invention is scheduled to work The present invention is placed in the text invention series and a concrete flow chart; a specific implementation of broadcast scheduling Process diagram; a specific actual process flow a specific real word extraction and division a specific overview stored in the system examples, examples, examples, diagrams; examples, formalized examples, sub-normalized data The subroutine description of the initialization of the process will be the flow chart of the voice data; $ now the general workflow I now used to approve the voice # The description will indicate the flowchart of the encoding process; &amp; The comparison of the process of using the data in the network front end $ The details of the preferred embodiments are as follows-the present invention is a system that is based on demand or timely and spreads through a computer network. The invention is composed of a response server (DIVR) 108, which is equipped with an interface card 4 1 06 'that can access the telephone network and receive calls made using standard telephone equipment 104. D I V R 1 08 operates according to the user description and other classified data contained in the system database 1 12 and can execute the workflow and publish software on the DI VR, or it can be a computer. DIVR 108 also contains

Page 10 479429 V. Description of the invention (8) Schedule program and loading software, saved voice data #. [Entry information]: At the beginning of broadcasting, the broadcasting station 118 ^! € i, JDTR0 ^ (inclusive) or more encoder 12. … Once twisted: send the voice to a data stream motif server 124, that is, I, it will store the data from the above dream department or store it in a (inclusive) or more ten series storage device 丨 22 According to the needs of the people * This skill is very clear to those who are proficient in Lee ’s system, those who can be π # 田 μ, 3 those who can be located on a single computer through the use of computer networks (like several physical computers Dissemination system. 疋] Ji, Zhoulu) Connected to the reply please see Figure 1, this figure shows the original 1G2 phone using a telephone connected to the telephone network: live system ^ usually adjusted according to the pulse code The PC data is connected to the telephone network through the interface device 106, and the system can receive PCM data from the telephone network. Hum ^^^ tongue to decode the phone key tone signal, the signal can also be used (DTMF) jealousy. Right-hand recording home office 丨 eight products) 疋 '疋 The known two-tone multi-frequency UMMH Λ has a kind of dry interface is Dialic ^, made &amp; input &amp; card, which can be connected to τ -1 voice network and P0TS (Old ^ Road. This interface is supplied by card, Annong Private Shito first ,,) ,,, within indiscriminately. The text is displayed on the computer that accesses the telephone network. This interface is installed in D0TELL's interactive voice response server (divr) 108, and the software in DI VR 108 will receive and confirm the voice writer to be saved. DIVR 108 also accesses the user profile stored in the system database ^ 2 to confirm the identity of the sound author. Department 112 is an ODBC (Open Database Link) compatible database, which can be accessed through the network by DIVR dither library 479429, invention description (9) 1 08, and is made by Sybase or heart "丨 # 出品 的" # Knowledge-based database management system (such as the one produced by Sybaw I uracieu) DIVR 108 will work the workflow and the king of thorns. Body 1 1 0 is responsible for executing a series of steps or regulations, the software 11 0 initializes the voice of the soft line Records. Each user describes the specific steps that M responds to the author of the voice. The example performed here is to execute a message after the recording is complete, including e-mailing the recording f ^ to respond to the latest recording for approval. Or, send it via FTP: The poor link is sent to the management 126 so that it is included in the web page. The text of the web is presented to the user by the web browser of the web server / through the client PC1 28 上 执 and ϊϊ :: 1 Two t through the network to connect to the management network feeder &quot; 4, = Medium: ΪΓΛ is set to issue queries and update to the system, system database, and the main page 116 Of this network-type front end 11 4 Use gding batch v 卞 ^ 八 化 化 一 ^ ^ 乐 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,-,-,-,- The library 112 stores all the data describing the system status and the preferences of individual users. The centralized user data contains the users; the name and location of the encoder 1 2 〇, which encodes the user's voice data, the encoded audio data format, Whether to authorize the user to broadcast in real time and whether the user needs to be approved before connecting to the recorded voice data sent to the client on the computer network. After DI VR 108 confirms the author and starts the recording operation, di VR accepts the voice data and sends the data Send to DTRouter 118 ° DTRouter 118 receives voice data from DIVR 1 0 8 scattered blocks. According to specific voice

Page 12 479429 V. Description of the invention (10) 2 User description slot, you can specify one or more encoder 120 tone materials. For each designated speech encoder 120, DTRc ^ ter} 丨 8 will create a copy of the received voice data in the buffer. DTRouter 18 wTCP / IP (Transmission Control Protocol / Internet) The communication protocol $) socket connection is opened to each encoder 120, and data is sent out every time a data block is received or the buffer is full. DTR〇uter 1 1 8 makes a decision based on user preference setting data, and passes it as DIVR 1 08 to generate a description and file containing χΜί data from the data contained in the system's database 1 12. The process of generating this XML description file will be explained more clearly below. The system can be set to make the best use of resources by routing the voice data to multiple user-based encoders;] Each encoder performs load balancing. In fact, the system can also be set up to consist of multiple DTR 0uters 1 1 8 containing user preference settings' or depending on the load balancing needs to determine which D Duter 118 ° voice data will be passed to one ( Including) The above encoders 1 2 0 are compacted. Compression encoder includes RealAudioTM compression / decompression program (compression program / decompression program) developed by Rea 1 N etw okrTM I nc, Q uick T developed by A pp 1 e C 〇mpute rTM I nc im eTM compression / decompression program, Windows MediaTM Format 'developed by Microsoft Corporation, or any other compression encoder known to those skilled in the art. However, it lacks the compression advantages of other encoders, so the system can also store data in WAV format. In fact, the system can be configured to use any number or kind of encoders' so that system resources can also be optimally used.

Page 13 479429 Description of the invention (11) After encoding the voice data, the backup device 124, the data storage device m, or both of these * 1 to 1 料 material stream media temple service data stream media corrupt servicer's tribute storage device 122 Both of the media servers 1 and 24 are located on the network whether and by which device the voice data is available. Or the data storage media device 124's ^ storage device 122 publishing system 11 can access these data. This workflow and workflow and publishing system 1 will ^^^^^ λ ^^ t, ν ^ νΛ θ # ^^ 立 次 J τ The publisher will obtain and format the text containing the location of the voice feed. Work industry trial collection _ 1 3 works, the claw right system will obtain the formatted link Asia Via FTP to the web server 126, the web server 126 will dynamically link the link to the web page sent to the client browser 丨 28. ^ The user clicks the voice file through the client browser 1 2 8 When the link access network 2 is 126, the network server will obtain the data from the location, or send an instruction to the appropriate data streaming server 1 2 4 to pass the voice data to the client. Figure 2 presents another specific implementation of the present invention For example, in this configuration, all system components are operated by a single entity 202. In other configurations, system components can be distributed everywhere so as to achieve the best performance. This configuration is a known application service Vendor or ASP. After controlling all parts of the system, The control company can provide a complete voice data stream service to any number of individuals or organizations. Figure 3 presents another specific embodiment of the present invention 302. This creates a situation where the company wants to control the voice capture of the present invention internally Function 3 〇 2

Page 14 479429, Inventive Note G2) One advantage of this configuration is that it is easy to maintain and has the ability to customize all systems to the needs of a particular company, such as unique hidden considerations and so on. Another advantage of this configuration is that when the core functions of the system operate within the company and are directly controlled by the company, the company itself does not need to transmit the voice content to the end users. 0 The company's internal main control system can be accessed through D IV R, The system database, encoder and management web server 302 are controlled. The system connected to the storage device and the data streaming server is controlled by the system license device, or a trusted third-party device capable of guiding and compressing the voice to meet the company's needs. These devices can be on any network that the system can access. Please refer to Figure 4, which shows a high-level overview of the system program flow. The voice writer calls the system using the telephone device 402, and the system is answered by D IVR 404. The identification of the caller will authenticate its record in the system database 4 06. If the information provided by the author is invalid, DI VR will ignore the call and hang up the connection 41 8. If the author's identification is authenticated 40 6, DIVR It will open a connection to DTRouter and send recorded metadata 40 8. The DIVR starts to receive voice data from the telephone network through the telephone interface panel 41 0, and transmits the received data to DTRou ter to reach 41 2. The author issues a predetermined DTMF tone 414 to stop the system recording, thus completing the voice recording. Once the recording is complete, the D I V R will update the database with information related to the recording and its arrangement ❶ 4 1 6. Then the system disconnects the call 4 1 8 ′ and starts the work flow 4 2 0 at the same time. The flow will be executed according to a set of user-defined rules and parameters 422. Return to step 412, when the DIVR receives the data, it will direct the data to

Page 15 479429 V. Description of the invention (13) DTR0uter. DTRouter receives relays and voice data 424 from a DIVR or scheduler and loader system. After receiving the data, j) TRouter will open the socket to the appropriate encoder 426 defined by the voice relay data. The data is passed to the encoder 428 via an open connection. The data is encoded in accordance with the complex 2 compression / decompression algorithm 43. The encoded voice is sent to a streaming server for live broadcast 432, or to a compressed storage device for broadcast 434 at a later time as needed. The media server then makes the encoded data accessible from the network 43 6. The result = the final result is the voice data created by the author over the phone 438 ^ namely ^ broadcast or broadcast on demand. Mingmai reads Figure 5, which is presented as a summary by the answering desk. A graded music player went to Jinshan and Shishi. The customer's front end of the burial wooden bag 4 "7 = traditionally connected to the telephone network while the phone is called 502," Chang's (CPE ') such as a telephone), also through DIVRi to the telephone network via if Waiting to answer the call 504 ^ When you receive a gas call, you will get the calling number 5506 and delivery code 51 0. If the shame 1. The purpose of ^ &amp; an incoming call 504, the system will terminate the call and wait for a new call Writing ^ ^ ^ put 'will mention that the author enters the company identification code 512. After the author receives it, he will enter the identification information | Shima. DIVR uses round-robin funding: No provide-a valid company request, the company code of the author, the program flow will return to the cattle] 9 1 Enter an invalid opportunity to provide a valid code.]; = 2 In -in, the author will have another-the right author cannot enter a valid 479429 within the second time. 5. The company identification code of invention description (14), the system will display ns 〇rry '' message indicates that the author can not provide a valid company code 5 2 0 and terminate the call, then return to step 504 and wait for other incoming calls. If provided by the author A valid company identification code, the system will prompt the user to provide a user identification code 522, which is also input through DTMF tone 524. The supplied identification code will be checked with the system database 5 2 6 to determine whether the author is valid Of users 5 3 0. As before, the user has three opportunities to enter a valid user ID. If a valid identification cannot be entered

Code ’system will display a message stating that the author cannot provide a valid user code 5 2 G and terminate the call, then return to step 504 to wait for another incoming call. A valid user code is provided on the right. The system checks the database to determine if the user has exceeded the maximum allowed number of calls 532. If the maximum number of calls is exceeded, the system will display a "sorry" message indicating the fact 52 ° and terminate the call, and then return to step 5 0 4 to wait to receive other calls. If the maximum number of calls is not exceeded, the system will update the system database The author's record is that the number of new calls that have been made should be 5 3 4. The system will also determine whether the user wants to assign voice messages on-site 53 6 and issue appropriate prompts 53 8 and 540. Then the general recording instructions will be played 542 and the system will wait for a liver tone 544 indicating that it is time to start recording.

When did the author provide 0? Tone 5 44 indicates that the recording has begun, 1) 1 ”??: Connect with one heart 546. D ϊ VR will also query the system database 2 ^ Establish an XML · description file, which includes encoding, storage, and guidance electronics 咭All the parameters required by the ancestors are 548. An eta data will be transmitted from i °, i Bekoku back to the author's preferences, and an XML description file will be created, which will be transmitted to DTR〇uter. 479429 5 Description of the invention (15) The system checks to determine whether the voice data has been generated, that is, to determine whether the author started to speak 5 50. If the author is talking, DIVR records the PCM data packet received from the telephone network 5 52 and records it 574 . The system also listens to the D T M tone 5 7 4 to indicate the end of the message. A copy of the original PC M data is stored in the local cache 5 5 4 so the author can recheck this message without having to access the storage device via the network. Get. The data packet will be sent to DTR〇uter556 and control will return to step 5 50, where the system checks the next voice data packet at a time. When the system fails to receive the voice data 5 5 0, the recording will end 5 5 8. System decision Whether the author paused the message 5 0. If the author did pause the message, the system will wait for 562 DTMF tones 5 64 indicating the recording to continue, or force the recording program to continue overtime at step 5 50. 566. If the recording is not paused, The system will decide whether the recording is completed 5 6 8. If the recording is not completed, it means that the “End Recording” DTMF tone is not received in step 9 552. The most likely issue is 1 Hang up or the line is disconnected 57 2 And then the system will go back to waiting for $ ^ corpse until another call is received. If you receive the ending recording tone :: ^ The system will play a prompt to the author to select the recording configuration 5 7 0. Figure 6 presents the process of determining the configuration of the completed record. The system presents the configuration option 6 0 2 selected by the user. # 者 by selecting "DTMF tones" to choose 604. If the author chooses to return to Shun to 606, the system will store the PCM data in the local cache two, and then replay the configuration option 6 08. If The author chooses to ignore: "Play and release the new recording message 61. This will cache the local player and re-start it from fi 1 9 Μ Ping Gong and re-record the heart 612, and return control to step 542. If The author decides to save the recording

479429, invention description (16), the number of recorded materials in the greed database will be increased by 6 1 6 and the name and location of the recording will be stored in the database 61 8 and the author's workflow will execute 6 2 G. The system will hang up the connection 6 2 2 and return control to step 5 0 ^. Finally ’will instruct the author to ignore the recording and exit 6 2 4. If this system ignores the local cache and hangs up the connection 626, return control to step 5 0 4 〇

When the DI VR queries the system database and obtains a set of predetermined speech author parameters in step 548 of FIG. 5, an XML description file will be generated. This data is responsible for driving all system decisions regarding the recording and storage of the author's voice data. Figure 7 shows an example XM L profile, as expected by a specific embodiment of the present invention. All the data is contained in a set of &lt; ENCODELIST &gt; tags 702, and just before the &lt; ENCODELISTM $ signature is a tag for identifying the company and user identification code of the author of the voice data 704. There are also tags that set wait and priority values, which will be discussed later, and are used when scheduling live broadcasts. <SOURCE> Block 70 6 describes the format of the voice data received by the system. What the source format or raw audio data looks like will be described in terms of its generation, sample rate, and size (how many bytes). Most telephone systems use the compacted // L aw and a L aw formats for broadcast data. The system will confirm receipt of the data and decompress it to produce linear PCM data. The remainder of the XML description file is composed of a plurality of <ENCODE> blocks 708. The example XML description building presented in Figure 7 consists of three coding blocks, one is Sun Audio, one is RealAudio, and the other is Windows.

Med i a. Each block starts with the media name that the encoder wants to generate, followed by the file name used to perform live broadcast. There is also a

479429 V. Description of the invention (17) Compression coding and <c o D E c L [$ τ>] block 710 containing the name of the compression / decompression program. In the case of Real Audio, a single data stream can be encoded at multiple ratios using multiple compression / decompression programs. The data stream server and client then use the protocol at the fastest rate (such as the highest quality) that can successfully pass the user connection. Data section server addresses for performing live broadcasts, and storage devices for storing compressed copies of voice data are also provided here. The flexibility provided by the two XMLs allows voice data generated by a particular author to be routed to a particular server. 'It may be the closest server, located on a less congested network segment, or more based on the author's identification code. of

Please refer to FIG. 8, which illustrates a DTRouter process that frequently receives voice data to the encoder. Once the server has initialized 802, it will wait for 804 until it receives the connection. When a connection 806 from d I v r is received, the XML description file will describe the speech data and encoding parameters received first. DTRouter parses the XML description file 8 07 and decides whether to attach it to the standard specification 80 8. If the XML description file contains abnormal data, DTRuter will delete the job and return to step 804 and wait for other connections. Show

If the XML description file is valid, the data will contain the AXMUnf0 object 81 0, which is a data structure that passes data to the encoder. Here, an XMLInfo object will be created for each encoder specified in the XML description file, and a double check will be performed to determine whether the received XML description file matches the existing job. The job is currently handled by DTR0Uter. Any voice call within 8 & 2 will be processed. If the X M L profile does not match the existing job, control will be passed to the newly created consumer thread 8 1 4 who is responsible for receiving incoming

479429 V. Description of the invention (18) Voice materials. DTRouter is also set up to send the received data to the encoder specified in the author's XML description file 8 16. After the appropriate threads have been established and control is passed to them, the main thread will return to step 804 and wait for the next connection. The consumer attempted to read the data from the open STCP / IP socket starting with the XML description file 8 1 8 and will check to see if the data 8 2 0 is available. For example, D IVR will send a capture from the voice author Pcm information. If the data exists, the data will be read from socks 8 22. The data is read and placed in the buffer 8 2 4 and the signal is sent to the encoder 826 where the data is available. Control then returns to step 820 to see if other data blocks are available. If there is no more data, the recording will be terminated and consumer thread 838 will be deleted. A group of encoder threads operates in parallel. One thread is used for the encoder specified in each XML description file to encode the speech data. The encoder opens &amp; TCp socket to the encoder 82 8. Once the connection has been opened, it will check whether the consumer has generated a signal 8 30 that the data is available. If information is available, 832 will be read from the buffer and written to the encoder 834 via the socket. Once the data has been written, the encoder thread will send a signal to the consumer who has access to the data buffer area 8 3 6. Control will then return to step 8300, where it will be checked to see if there is the next data block. If there is no more data, the recording will be terminated and the encoder thread will be deleted. Each executing consumer thread is identified by its processed XML description file, which can be viewed back to the peak-fingerprint of step 812. The existing XML description file can be compared with the XML description processed by multiple consumer threads in DTRouter. Compare

Page 21 4ff9429 V. Description of Invention (19). If the system defines the equivalent XML description file (for example, it contains consistent company and user identification codes), it will be processed in the current thread, and it will be checked in 840 to see if the newly received XML description file has its <PRIORITY> Higher value identified by label 704. If the received XML description file has no higher priority, DTRouter will ignore the data and return to step 804 to wait for a new connection. If the XML description file has higher priority, the pre-existing consumer thread will be suspended 8 4 2 and a new consumer thread 8 4 4 will be created. Because existing consumer threads work with encoder threads, the newly created consumer thread 844 will continue to pass data to it. A check is made on 846 'to determine if the newly received XML description file specifies a different compression location from the pre-stored XML description file (which was previously paused in step 842). If a new compression location has been specified, an additional encoder thread will be created to receive data from the consumer thread for encoding into the specified compression file 8 4 8. If no new compression file 8 4 6 is specified, only the existing encoder can encode the incoming speech and continue to use the newly created consumer thread. As explained later, this routine is used for scheduling and loading of real-time broadcast voice. FIG. 9 is a process performed by a server configured to encode data into a format such as DingTuo, Reai, or Windows Media. When the server starts processing, it will initialize 902 the listening socket, which is used to wait for the connection 90 4 from the DTRouter. After the connection is established, the connection specific thread (produced by the main thread 908) will be used to read data from sock e t 9 0 6 and then the host will return to step 9 0 4 and wait for the next connection. The connection specific thread will receive an XML description file, which is attached to the voice data to become "? € 9 1 0. The X M L description file is parsed 9 1 2 and the encoder decides this

Page 479 429

Whether the data contained here contains valid parameters can encode the incoming voice data 914. Where the XML data contained in the profile is invalid, the thread will be deleted and the connection maintained by it will be closed. If the data is valid, the encoder will initialize, initialize 918 and read the voice data 92 from the socket. Once the transmission of voice data from DTRouter begins, the data will be read from the socket and encoded by the compression / decompression program 9 2 4. If the code of the compressed speech data is a Qu 1 ckT 1 me encoder, once all encoding is completed, it will be combined in the movie building of Shima to add a cue track 92 6. After encoding is completed, ^ will be closed 9 2 8 and the encoded voice data will be stored according to the XML description file 9 3 〇 ^

Data storage device or streaming server. The connection is stopped and its connection to DTRouter will be closed 91 6. Figure 10 presents the procedure for compressing the incoming speech data using a Quick coder. Set the compression / decompression program in 00 2 first, and then the voice magnetic preparation can prepare to write the incoming data to 1 004, and then the code will be set to the standard number. The status of receiving data for encoding is 1 0 0 6. This completes the initialization and ^ bundles subroutine 1008, returning control to the main routine.

Figure 11 shows the procedure for compressing incoming voice data using a Real Audio encoder. The encoder first determines whether the XML description file indicates that the encoder has more than one compression / decompression program to encode speech data. If it does not use a single compression / decompression program, initialization will be performed. Where multiple compression / decompression programs are instructed, SureStream initialization will be performed for each compression / decompression program. SureStream is

A patented technology developed by Real Networks that encodes incoming voice data with multiple compression / decompression programs and stores it in a single file

/ ^ 29 V. Invention Description (21) &quot; 一 ~ 一 '' Kinai. Then the client asks to connect, and Su r e s t r e a m starts to coordinate the servo cry with the client, using the best encoding supported by the client connection to send the data / ¾. 〇. 2. This process continuously determines whether the voice is broadcasted in real time, which is also determined by the XML description file received by the encoder. If instructed to broadcast in real time, the server will initialize the streaming server software to directly receive and broadcast its encoded voice data 1 π 〇, and then the encoder will determine whether the voice should be compressed into the storage device 1 1 1 2 If you want to compress the speech, the output file position will be initialized to 1 Π4, and then the encoding engine will be set to be ready to start encoding, and the initialization will be completed. If it is instructed not to compress, the compression / decompression program is set to a semicolon number to start encoding 1116, and 9, 1111 is initialized. At the beginning of the war, Figure η shows the procedure for compressing the incoming voice data using Windows Media encoder. The compression / decompression program is initialized in step 丨 202, and the L: coder determines whether the voice is to be broadcasted in real time. If instructed to do so immediately, the U server will run the breeding stream media server software 1206 with an additional alias in the file directory path used by the server, and then the encoder will decide to go to if I; factory = Jian: 12 .8, if you want to compress the voice, it will; The first two = Cheng has already broken x and Fancheng initialized 1214. The right indicates that no compression is performed. The compression / decompression program will be set to the reference number to start encoding 1 2 1 2 and the initialization 1 2 U will be completed. Please refer to FIG. 13 ′ This figure ^ presents the process of encoding a standard uncompressed .w. t — At the beginning the voice data is stored as

479429 V. Description of Invention (22) Case ’socket initializes 1302 and waits to accept connection 1 304 from Router. When the connection is established, it will use the connection specific thread (produced by the main thread 3 06) to accept data across s 0 c k e t 1 3 0 8. The thread then returns to step 1304 and waits for the next connection. The connection specific thread will receive an XML description file, which is attached to the voice data as the title 1310. The cXML description file is parsed 1312, and the encoder determines whether the data contained here contains valid parameters to input the incoming voice data = Ma 1314. Invalid XML profile 4, the thread will be deleted and the connection 1316 caused by the thread will be closed. If there is a text in the data, the encoder will prepare v 题: The title is not supplied with the value of the data length 1 3 8 contained in the file. The flat coder reads the data from the open socket 13.0 a, say 41 322, it will write the data · wav data-B p ^ pull all the data / materials, the encoder will produce data long: · To the stall to slaughter 132fi. The name 纴 $ + 丄 bedin ten negative degrees entered the wav title and closed Kodagi U26. After finishing the coding

The description file is stored in the data storage dream department $ 次 u… said that Beco will stop in the two media feeders according to the XML connection specification 1 328. The present invention provides that a connection of $ = 0 Uter will be closed 1 31 6. ability. Figure "Inside and outside: the processing of the real-time voice call broadcast via the computer network is responsible for broadcasting in the program (S / L system) to perform the broadcast. The system's database stores: procedures; immediately save voice data The _ shell material that has been scheduled and ended ^ event, especially started live broadcasting at 2:00 PM schedule, for example, if the user has: until the author starts broadcasting on his own. Cannot perform schedule Guang #, if the author is in fine J; there is a case of 479429

The data accompanying the 0 event will instruct the system to terminate the transmission of the stored voice. The j program will continue to loop and poll the system. It is determined that the broadcast needs to be deducted until the live author's license number starts the scheduled live broadcast = no start . The loop starts when the S / L system queries the system database and decides whether it has been: Scheduled and events starting in the next minute 1402. If the current user opens: A scheduled live broadcast 1404, the S / L system will generate a copy of the XML description file 1 40 6 and send it to DTRouter 14 0. The system database stores voice data attached to scheduled events, such as a Southern "waiting to start transmission" message or a message to a specific user. After sending the XML description file at 1 408, the S / L system will access the local or network file system 1 412 to open the voice data attached to the scheduled event. Once turned on, the voice data will be transmitted to DTRouter 1414. The system will then query the database to determine if there are currently other events scheduled to start -141 6. If there are other events scheduled to start 141 8 then Re-execute the process of broadcasting music from step 1406. If there is no event scheduled to start 1 4 1 8 and 1 4 0 4, the system will once again obtain the event data 1 that has been scheduled to end at the current time. 422. At the moment, there is no event to end at 1 424, the system will wait for a minute 1420 and begin to end processing 1 2 0 2. If the database returns a message 1 424 indicating that the saved voice is at the end, the system will generate a The event description of the negative priority XML description coffin is used to indicate that the broadcast is to be terminated, and then the XML description file will be transmitted to DTR 〇uter 1 4 2 8. A query is generated here to obtain additional events that have been scheduled to end at this time. Data 1 4 30 If the database returns instructions to save the voice

Page 26 479429 V. Description of the invention (24) The information to be ended at this time 1 4 3 2 will save the processing of the voice broadcast from step 1 4 2 6. At present there are no events that require 1 432. The system will wait for 1 4 20 and start and end. As explained earlier, DIVR will initially execute a workflow software for a series of user-defined actions or events. This series is a combination of a series of executable executable programs written in C / C ++ language, such as PERL or PYTHON. The workflow is executed here in response to the voice action (see step 6 2 0 in FIG. 6). Figure 15 presents a workflow according to a specific embodiment of the invention. The system that executes the workflow will run a series of parameters of the software 1504, and the parameters received by the workflow software will execute the query 1506 in the system database, and the incident 1502 will be attached. In 1508, the working state will also be set to avoid executing multiple copies of the same workflow at the same time. The organization will also obtain other metadata 1 5 10 and user information related to the records required to properly perform the workflow. Events with workflows 1 5 1 4 and 1 5 1 6 until everything is over. After all events are completed, the workflow is as shown in Figure 16. According to a specific embodiment of the present invention, a specific embodiment of the process is presented. When an author goes through the system (he must be approved by the administrator), the process is performed. After the recording is completed, the workflow will be executed 1 6 0 2 e-mails to the administrator, explaining the re-execution termination in the hyperlink of the web page, the end of the processing 1 2 0 2. The column is suitable for the script code that makes the workflow included, or a call that is completed by both of them. After passing a common number to the job, it will use this to obtain the execution status of the workflow process. The workflow material 1 5 1 2 These software execution packages have been executed and will terminate 1 520. Create a workflow, record a sample workflow during a call, and send a

Page 27 479429 V. Description of the Invention (25) Record 1 6 0 4 The workflow generates information from data taken from system data. The administrator clicks the hyperlink 1 600 in the email message to navigate to the management web page 16 0 5. The management webpage allows the administrator to set the properties of the recorded voice data, such as the title naming and set its "approved" status. The administrator names the title of the record 16 06 and checks its validity 16 08. If the title is invalid, the administrator will return to the management page presented in step 16 0 5 and provide the opportunity to name the record again. If the title is valid, the record recorded in the system database will be updated 16 1 0, and the approver will set the recorded “approval” attribute 1 6 1 4. If the recording is not approved 16 16, the approver will present a screen summarizing the changes made to the recording records in the database 1 6 1 8 and terminate the workflow 1 62 0. When the administrator approves the recording 1 6 1 6, the database will be updated to reflect the change in approval status 1 6 2 2 and then three actions will be performed simultaneously. Send a copy of the recording as an e-mail attachment to a specific archive custodian 1 6 2 6 After the administrator approves the recording, a day-to-day summary of the changes made to the recording in the database will appear 1 6 1 8 and then Perform the publisher 1 6 2 4 described later. The results of the publisher will be transmitted to a specific web server via FTP, and the web page will be part of it. After these three steps have been performed, the workflow will end 1620. The publisher is also a software subsystem initialized by DIVR to generate text data indicating that the location of the voice data is on a network storage device, a streaming server performing live broadcasting, or both. Just like workflow software, publishers include a series of directives like PERL or PYTHON

Page 28 V. Description of the invention (26) Instruction code written by σ, a series of 丨 ν r / r 丄 丄 μ ^ program, Chi Qu 4: Tuilei translation in c / c + + language executable Coarse or k. After the application of the workflow, the jitter can be transmitted via email and then embedded in the web page. To the user or to the web server, and execute: = 本 = a specific embodiment 'present-general publisher's publishers a series of parameters to perform the publishing re-invention-a structure based on the publisher's internal data Set to 1704. ;:! System Database 17.8. Whether or not to get its :: number? = Establish a s〇l 杳% in the greed received by shame P. Perform 1714 query in the system database = format im, and this parameter can be changed according to = different ... and then return the formatted result to the call = As a result, users with advanced access permissions can view and store the data contained in the database, and the administrator may have access to the &gt; side J ’s access auditing, which may allow users of the company, Add or edit work methods and scheduled events. And there are more: two, out; ^ ^ Μ (,. | Superuser &quot;) / ^ ^ external workflow, publisher and job events to add new addition In addition, superuser can also add and edit temple server, servo Device storage flat series. Company 0 The Internet-based consultation platform t 4 Most collections and publishers, iik —, 枓 Kuli ^ t, Ying and other features allow the administrator 3 is now involved in the addition, editing and general programming of Liu Lan ; J :. Figure 18 shows

Page 29 479429 V. Description of the invention (27) When adding data, the administrator first selects the appropriate form or template for adding the required data 1 8 02, for example: if the administrator wants to add a new user to the system Inside, you should use `` add user '' form. After completing the information field on the form, 1 0 0 4 will check the consistency of the poor material. 8 0 6 If the data fish 'fixed template does not match 1 808, for example, more than one encoder address is entered, ^ The error message 1810 is displayed. Here is an opportunity for the user to submit a positive 2 order 1 8 1 2 is to check the correctness again. When the information entered is correct 1 8 0 8, the record will be added to the database 1 8 1 4. &quot; When editing the data, the system will get the "item" form from the database. The offending item is user, company or server. Then the administrator selects item 1818 from the list of returned items, and then the user selects component 1819 from the selected '&lt;. For example, ‘If the administrator wants to edit a user,, first select ,, users, from the list of items returned from the database, and then manage = and then select a specific user from the list of all users. The system obtains the information 182 () about 70 items from the database T, and presents it to the administrator 1 822 in an editable morning. The data were processed through the form, and then checked to confirm that the new data was consistent with the predetermined template 1 826. If the data and the pre-samples are inconsistent 1 8 28, an error message 丨 83 〇 will be displayed and the administrator will be given an opportunity to correct the error 1 832, and then the consistency of the corrected data will be checked again 1 82 6. When the format of the data is correct 1 828, &amp; the data is written back to the database ′ to update the record 1 834. ^ At the time of Liu Lan, the item list 1 836 was obtained from the database. The administrator selected an item 1 83 8 and the component 1 839 under the item. Information about the selected component is obtained from the database 184 and actions 842 are performed. Produce

479429 V. Description of the invention (28) The results produced will be displayed to the administrator 1 8 4 4 and the administrator can modify the components through the editing interface. Although the present invention is described and illustrated by the preferred embodiments, those skilled in the art can make many changes and modifications without departing from the spirit and field of the present invention, so the present invention is not limited to precise methods The details of the study or the structure of the above publication, so many changes and modifications are included in the field of the present invention.

Page 31

Claims (1)

^ / ^ 429 Method of data: Speech method • ~ Voice of phone call through computer network includes:, Receive voice data through incoming call from the author of the voice data; 2 Encode the voice data into the target file format: and =, store the pasted voice data In the file system of the executive weeping access. … The computer that catches soft peptides every day "2 · For example, the data in the first item of patent application range to one of a plurality of encoders. Go, the Hai method includes routing voices. 3 The method of the second scope of the patent application is as follows. The author of the audio-visual materials is Ding Mo Rengu, Bajiayi ... and the accompanying words. The steps of-are to import the voice data to a plurality of editors. 4. If the method of the scope of patent application is the first, the method includes: assigning a position to the encoded voice data; and the value is in! ! Before the voice data is stored in the file system, transfer the voice data to the official for approval. 5. Rushen, the method of the first item of the patent scope, which includes a set of rules that accompany the author of Ik δ, so that the voice client of the data stream can obtain the position of the voice data that has been coded. .6. The method of claim 1 in which the encoding step is performed by a Real Audi ™ compression / decompression program. 7. The method according to item 1 of the patent, wherein the encoding step is performed by a QuickTimeTM compression / decompression program. 8. The method according to item 1 of the scope of patent application, wherein the encoding step is performed by a Windows MediaTM compression / decompression program. Page 32 6. Scope of patent application Among them, according to the w av format guide 9. If the method of the scope of the patent application item No. 〖enter the encoding step. 〆1 0 · —continued via a computer network. The system includes: ㈣ 电话 Phone voice data system-a first computer, of which the voice data resides on the right; σ to receive a younger brother via a phone network-computer, Among them, the ancient Pt 4 ·, the data is encoded into the destination file format, and the software can directly receive the received voice into a data storage device, using a third computer, of which the right and left are stored, and the coded voice data; and software. ° The coded voice data stored in circulation is used. 11 · If a patent application is filed] Λ / Contains a plurality of computers composed of a plurality of different codes, where the second computer package is also a computer composed of software And the second computer: "brain" is used to execute software to route voice data to a selected source containing the π ii author to route voice data to one of a plurality of '3 different computers (including) or more . Brain, the computer according to the real-time / j1 item, the system 'the system contains a power to a plurality of authors including different papers to set the voice data routing. )the above. System with Rea 1 A u d i oTM start / item ’where the coded software package 15. If applying for a special = ^ contraction program. The system of item 10, wherein the encoding software package Page 33 479429 6. Scope of patent application ^ Contains Qu i c kT i meTM compression / decompression program. 16. The system according to item 10 of the scope of patent application, wherein the encoding software includes a Winn d 0 w s M d e d i aTM compression / decompression program. 1 7. The system according to item 10 of the patent application scope, wherein the encoding software includes a · wav format. 1 8. —A system for distributing telephone voice data through a computer network, the system includes: a computer including: the first software for receiving voice data transmitted from the telephone network; directly encoding the received voice data into a purpose The second software in the file format; the third software that circulates the encoded voice data; and a data storage device for storing the encoded voice data; and a computer connected to the network by two circuits, which allows the client to run The software client computer accesses to receive streaming voice data. 1 9. A method for distributing telephone voice data via a computer network 'The method includes: receiving a call from a computer, the call is made by a voice data author; obtaining personal description data based on the author; Describe the data contained in the file, directly encode the voice data into the destination file format; and store the encoded voice data in a file system that can be accessed by a computer running voice streaming software. , Page 34 479429 6. Application for Patent Scope 2 〇. The method for applying for Patent Scope Item 19, the method includes: setting the data contained in the description file according to user preferences, and routing the voice data to a plurality of encoders; And the speech data is encoded by a plurality of encoders. 2 1. The method according to item 20 of the scope of patent application, wherein the encoding steps of the voice data occur almost simultaneously. 2 2. The method according to item 19 of the scope of patent application, wherein the personal description file is data formatted in accordance with XML rules. 2 3. The method according to item 19 of the scope of patent application, wherein the encoding step of the voice data includes setting the encoding characteristics according to the data contained in the author's personal profile. 2 4 · The method of item 19 in the scope of patent application, wherein the step of storing the encoded voice data includes storing the encoded voice data on a storage device designated in the author's personal profile. 2 5 · The method of item 19 in the scope of patent application, including the execution of the workflow according to the author's personal profile. Page 35