WO2008131636A1 - Procédé de mesure de données multimédia - Google Patents

Procédé de mesure de données multimédia Download PDF

Info

Publication number
WO2008131636A1
WO2008131636A1 PCT/CN2008/000421 CN2008000421W WO2008131636A1 WO 2008131636 A1 WO2008131636 A1 WO 2008131636A1 CN 2008000421 W CN2008000421 W CN 2008000421W WO 2008131636 A1 WO2008131636 A1 WO 2008131636A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
multimedia data
packet
error
multimedia
Prior art date
Application number
PCT/CN2008/000421
Other languages
English (en)
French (fr)
Inventor
Rui Chen
Xiujiang Guo
Zhiying Wang
Zhijun Dai
Ning Wang
Original Assignee
Zte Corporation
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Priority claimed from CN2007100977929A external-priority patent/CN101296166B/zh
Priority claimed from CN 200710111645 external-priority patent/CN101316198B/zh
Application filed by Zte Corporation filed Critical Zte Corporation
Publication of WO2008131636A1 publication Critical patent/WO2008131636A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0823Errors, e.g. transmission errors
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5003Managing SLA; Interaction between SLA and QoS
    • H04L41/5009Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]

Definitions

  • the present invention relates to a multimedia information transmission and processing technique, and in particular to a method for measuring multimedia data.
  • Multi-media applications include video applications and audio applications, and traditional applications such as WWW, E-mail, etc., have greater bandwidth requirements.
  • Existing multimedia data transmissions such as video and audio draw on high compression coding techniques that utilize time, space, and statistical correlation of multimedia information such as video and audio for data compression.
  • the compressed multimedia data is very sensitive to transmission errors.
  • existing wired or wireless networks do not provide QoS (Quality of Services) guarantee, and packet loss or delay is inevitable, which will result in degraded user experience quality at the receiving end. . Therefore, monitoring the performance of multimedia services has become a research hotspot in the field of multimedia communications.
  • QoS Quality of Services
  • the server side packages the multimedia data and sends it to the receiving end, and passes through several network nodes.
  • multimedia packets may be delayed or lost, causing the receiving end to not receive all packets correctly.
  • the receiving end finds the error information in the multimedia content through the multimedia decoding process, and feeds it back to the server.
  • the monitoring server uses the multimedia decoder to estimate the monitored multimedia data of the receiving end according to the received feedback error information.
  • the monitored multimedia data is compared with the reference multimedia data to monitor the performance of the multimedia service.
  • multimedia decoding processing is required at the receiving end, so that the error information in the multimedia content caused by the network transmission error can be fed back to the server, which is not conducive to the protection of digital copyright.
  • the present invention provides a method for measuring multimedia data, which is applied to a system including a transmitting end, a receiving end, and an intermediate network transmission node, and includes the following steps:
  • the foregoing method may further have the following feature, before step (1), further comprising setting a reference point and a sampling point in the system, where the reference point is set in the network transmission node A, and the sampling point is located in the network transmission node A or After any network transmission node that needs to perform service monitoring, there is a two-way communication channel between the reference point and the sampling point.
  • the foregoing method may further have the following feature: the packet sequence number of the multimedia data packet is a serial number of the encoded multimedia data, and in the RTP protocol, the packet sequence number of the multimedia data packet is an RTP packet sequence number; or, in TCP/IP In the protocol, the packet sequence number of the multimedia data packet is the IP packet sequence number; or, in the H.324M protocol, the packet sequence number of the multimedia data packet is the data frame sequence number.
  • the index includes a data packet number, a data type, and index data, where the data packet number indicates a packet sequence number of the received or sent multimedia data packet.
  • the data type represents a multimedia data packet type
  • the index data represents multimedia data corresponding to a data packet number.
  • the above method may further have the following feature.
  • the packet sequence number of the lost multimedia data packet is obtained by decoding at the sample point.
  • the decoding at the sampling point is sampling The point sorts the received data packets according to the packet sequence number of the multimedia data packet, and detects the integrity of the data packet. If there is a data packet loss, the packet sequence number is fed back to the reference point.
  • the decoding at the sampling point is to perform multimedia data decoding by using a multimedia decoder at the sampling point, and if there is a data packet loss, the packet sequence number is fed back to the reference point.
  • the foregoing method may further have the following feature, in step (2), distinguishing an error region of the error multimedia data according to a packet sequence number of the lost multimedia data packet returned by the sampling point, and information stored in the index Non-error area.
  • step (2) the error multi-media data is constructed by using the found error area and non-error area and the index data stored in the index file.
  • the above method may further have the following features.
  • the reference multimedia data refers to correctly decoded multimedia data, or multimedia data without erroneous data.
  • the foregoing method may further have the following feature: when the data type is video data, the index data includes a video parameter and a frame number of a video image frame, and the video parameter refers to video sequence coding parameter information, where the video The frame number of the image frame refers to the sequence number of the video image frame corresponding to the current multimedia data packet.
  • the above method may further have the following features, the video parameter includes an image height, an image width, a video frame rate, or a video format; the macroblock information includes a type of an image block, a number of an image block, and motion information of the image block. Or refer to the frame information.
  • the foregoing method may further have the following feature: when the data type is video data, the index data further includes macroblock information and/or residual information, where the macroblock information refers to a frame for reconstructing a video image.
  • the residual information refers to differential data between predicted values and decoded values of pixel points inside the image block.
  • the foregoing method may further have the following feature, when the data type is audio data, the index data includes a frame number including an audio parameter and an audio data frame, and the audio parameter refers to an encoding parameter of the audio data, where The frame number of the audio data frame refers to the sequence number of the audio data frame corresponding to the current multimedia data packet.
  • the audio parameters including a sampling rate and a sampling interval of the speech encoding.
  • the foregoing method may further have the following features.
  • the index includes a packet sequence number, a data length of the index entry, segment information, a macroblock, block information, or residual information.
  • the foregoing method may further have the following feature: the error area and the non-error area are distinguished according to the following method, and the frame number of the data frame where the error multimedia data packet is located is found according to the multimedia data corresponding to the erroneous multimedia data packet.
  • the data frame in which the frame number is located is referred to as an erroneous data frame.
  • the area corresponding to the erroneous multimedia data packet in the erroneous data frame and the area affected by the erroneous multimedia data packet in the subsequent multimedia data frame are error areas, and the remaining areas are non-error areas.
  • the foregoing method may further have the following feature: when reconstructing the multimedia data at the reference point, the reconstruction of the non-error area is directly copied from the corresponding position of the reference multimedia data, and the error area is masked according to the video error concealment method or the packet loss compensation. The method is reconstructed.
  • the foregoing method may further have the following feature: when reconstructing the multimedia data at the reference point, the reconstruction of the non-error area is directly copied from the corresponding position of the reference multimedia data, and when the error region is reconstructed, the error region is directly assigned 0, or copy data from the position corresponding to the previous frame of the error multimedia data in the reference multimedia data to the error region, or according to the motion block compensation method according to the macroblock information and the residual information of the image.
  • step (3) the video quality monitoring value is calculated for the reconstructed error multimedia data and the reference multimedia data by using an effective peak signal to noise ratio method, and the video shield monitoring value is used. Compare, perform multimedia service performance monitoring.
  • the method of the invention establishes an index between the packet serial number and the multimedia data of the transmitted multimedia data packet, returns the packet serial number of the multimedia data packet when the multimedia data transmission error, searches for the multimedia data in the index according to the packet serial number, and performs multimedia data.
  • Reconstruction compare the reconstructed multimedia data with the reference multimedia data, obtain the evaluation index of multimedia service performance, conveniently realize the monitoring and evaluation of multimedia service performance, thereby simplifying the process of reconstructing multimedia data during multimedia service quality evaluation, which is beneficial to Multi-service performance monitoring of multiple sampling points simultaneously at the reference point.
  • Figure 1 is a diagram of a multimedia service quality evaluation system based on index and feedback information
  • FIG. 2 is a schematic structural diagram of an index file according to an embodiment of the present invention.
  • FIG. 3 is a flowchart of a method for reconstructing an error video sequence according to an embodiment of the present invention
  • FIG. 4 is a schematic structural diagram of an index of a second embodiment of the present invention.
  • Figure 5 is a schematic diagram of block division in H.264;
  • FIG. 6 is a schematic diagram of a reconstruction process of an error video sequence according to Embodiment 2 of the present invention.
  • the invention simplifies the reconstruction process of the multimedia data in the multimedia service quality evaluation by establishing the index between the packet serial number and the multimedia data of the transmitted multimedia data packet, obtaining the location of the erroneous multimedia data through the index, and simplifying the process of reconstructing the multimedia data during the multimedia service quality evaluation, which is beneficial to
  • the reference point performs multimedia service performance monitoring on multiple sampling points.
  • the transmitting end packs the multimedia data and sends it to the receiving end, and passes through several network transmission nodes.
  • the reference point is set at the network transmission node A.
  • the sampling point is located at the network transmission node A or any network transmission node that needs service monitoring. There is two-way communication between the reference point and the sampling point. aisle.
  • Reference Point In the process of packaging, an index between the transmitted multimedia data packet number and the multimedia data is established, that is, an index between the packet serial number of the transmitted multimedia data packet and the corresponding multimedia data is established.
  • multimedia packets may be delayed or lost, causing the sampling points to not receive all multimedia packets correctly.
  • the sampling point detects the packet number of the lost packet and feeds it back to the reference point.
  • the reference point searches for an area of the multimedia data included in the packet in the index file according to the received packet number of the lost packet, determines a non-error area and an error area, and performs different reconstruction on the non-error area and the error area to obtain the reconstructed multimedia.
  • the data is compared with the reference multimedia data (correctly decoded multimedia data) to monitor the performance of the multimedia service.
  • Performance monitoring of a typical multimedia video service based on the H.264 video coding standard For example, the performance monitoring principle and implementation method of the present invention are further described, but are not limited to the application.
  • the video data is encoded into an RTP packet after being encoded by an encoder conforming to the H.264 video coding standard, and passes through the network node 1 and the network node 2 to reach the receiving end. Since the data transmission network cannot guarantee the quality of service, the multimedia data received by the receiving end may be inconsistent with the data of the transmitting end. There is a possibility of packet loss between the sender to the network node 1, between the network node 1 to the network node 2, or between the network node 2 and the receiving end.
  • Step 110 Construct an index-based multimedia data packet, build an index, and write an index file.
  • the reference point performs a normal decoding operation when transmitting the multimedia code stream, and the reference point establishes an index between the transmitted multimedia data packet number and the multimedia data in the process of packaging, that is, establishes a packet serial number and corresponding multimedia for transmitting the multimedia data packet.
  • the index between the data.
  • the packet sequence number of the transmitted multimedia data packet is the serial number of the encoded multimedia data transmission.
  • the packet sequence number of the transmission multimedia data packet is the packet sequence number and other information required for reconstructing the multimedia data.
  • the packet sequence number of the transmitted multimedia data packet is the RTP packet sequence number.
  • the multimedia information used for indexing includes multimedia data information corresponding to the data packet.
  • the reference point indexes and writes the index data to the index file.
  • the index file contains the index data written.
  • the structure of the finally created index file is shown in Figure 2, but it is also possible that the names and order of the fields in the index file are different.
  • the index structure shall contain (without limitation) the following information:
  • Segment information indicates different meanings in different audio and video coding standards, such as macroblock group (GOB, Group of Blocks) in h.263, fragmentation in other video standards (Slice ) and strips.
  • GOB macroblock group
  • Slice fragmentation in other video standards
  • IDCT inverse discrete cosine transform
  • Step 120 At the reference point, if the multimedia service quality of a certain receiving end (ie, the sampling point) is to be monitored and evaluated, according to the packet number of the lost packet returned by the sampling point, the index file is searched. According to the information stored in the index file, the multimedia data included in the packet is found in the area where the error frame is located, and the error area and the non-error area of the frame where the error is located are distinguished.
  • sampling point 1 sampling point 2 and sampling point 3 are established at the network node 1, network node 2, and receiving end positions, respectively. In practical applications, the sampling point can be at any network node location.
  • Each sampling point judges the serial number of the received video data packet (for example, the RTP packet serial number), and then sends the sequence number of the error video data packet to the reference point through the feedback channel, and the reference point can be known according to the serial number of the error video data packet. What video information has been lost when the sample is decoded.
  • step 130 the multimedia data with errors is reconstructed, and the error region and the non-error region are reconstructed using different methods.
  • the multimedia data with error is quickly reconstructed at the reference point analog sampling point.
  • the missing point letter can be known.
  • the so-called partial decoding refers to the use of the information stored in the index file for multimedia data reconstruction, without reading the code stream from the code stream file for complete decoding operation. Therefore, there is no need to perform additional multimedia decoding process processing at the reference point, thereby reducing the resource requirement of the reference point, and facilitating monitoring of the multimedia service performance of the plurality of sampling points at the reference point.
  • the process of error video reconstruction is simulated by using an index file, a correctly decoded file, and error information fed back from the sample points.
  • the video frame between two I frames (including the previous I frame, excluding the next I frame) is used as a picture group (GOP, Group Of Picture), as shown in the figure, II, 2 ⁇ , 3 ⁇ , 4 ⁇ ,
  • There are no errors in the frame such as II and 2B in Figure 3
  • the corresponding video frame can be copied from the correctly decoded file to the corresponding position of the reconstructed video file.
  • the general error processing is performed on the frame where the packet is lost.
  • the simplest method is to copy the corresponding block from the previous frame to the error position (such as the black part in the 3P frame in FIG. 3), and for other positions without errors, Copy the corresponding image block from the correct decoded file.
  • the error part in the subsequent frame affected by the error (such as the black part in 4B and 5P in FIG. 3) is predicted according to the motion vector of the corresponding block and the reference frame, and the predicted value of the corresponding image block is obtained.
  • the upper residual is the value of the reconstructed pixel block, which is the method of partial decoding.
  • the portion of the frame that is not affected by the erroneous reference frame is copied and filled from the correctly decoded video, thereby simulating the decoding process with the error of the sample point.
  • Step 140 service shield quantity evaluation.
  • the video quality measurement can be calculated for the reconstructed error video and the correctly decoded video by using an effective peak signal noise ratio (EPSNR) method. (VQM) value, This VQM value was taken as the final evaluation result.
  • EPSNR effective peak signal noise ratio
  • the process described in the present invention can reduce the following calculations compared to the conventional process of completely decoding one error video: 1) no need to process chrominance information; 2) no need to read the code stream and VLC decoding. 3) No prediction of motion vectors is required. 4) No IDCT transformation is required. 5) The computational complexity is related to the location and number of lost packets. The farther the packet is dropped from the I frame for the first time, the lower the calculation.
  • the index file may also use other structures different from the first embodiment, and will be specifically described below by another embodiment.
  • Embodiment 2 of the present invention specifically includes the following steps:
  • an index between the packet sequence number of the transmitted multimedia data packet and the corresponding multimedia information is established.
  • the packet sequence number of the transmitted multimedia data packet is the serial number of the encoded multimedia data.
  • the packet sequence number of the transmitted multimedia data packet is the RTP packet sequence number; in the TCP/IP protocol, the packet sequence number of the transmitted multimedia data packet It is the IP packet sequence number; in the H.324M protocol, the packet sequence number of the transmitted multimedia data packet is the data frame sequence number.
  • the multimedia information can be compressed video or audio data, or other data types.
  • the established index should contain (without limitation) the following information:
  • Packet No. indicates the packet number of the received or transmitted multimedia data packet.
  • the data packet number can be used as an index search key to retrieve the required multimedia data.
  • Packet Type Indicates the multimedia packet type of the current index.
  • the data type can be video or audio (Video Type/Audio Type).
  • Index Data Indicates the feature data of the multimedia data corresponding to the packet number.
  • the content of the index data is different for different multimedia data types (video or audio).
  • the index data should include video parameters, frame numbers of video image frames, macroblock information, and residual information. Macroblock information and residual information can be used as optional index data content.
  • Video parameters refer to video sequence coding parameter information, such as image height, image width, video frame rate, video format, and the like.
  • the frame number of the video image frame refers to the sequence number of the video image frame corresponding to the current multimedia data packet.
  • the index data should contain the frame number of multiple image frames.
  • Macro block information refers to image block feature data for reconstructing a video image frame, such as the type of image block, the number of the image block, the motion information of the image block, the reference frame information, and the like.
  • the macroblock information should contain feature data of a plurality of macroblocks.
  • a frame of image consists of multiple macroblocks.
  • Residual information refers to feature data for reconstructing a video image frame, including differential data between a predicted value and a decoded value of each pixel point inside the image block.
  • a picture contains a plurality of slices, each of which contains a plurality of macroblocks (Macroblock).
  • Each of the macroblocks has a size of 16 ⁇ 16 pixels, which is further divided into 16 4 x 4 pixel sub-blocks (Blocks).
  • Blocks When storing macroblock information, it is necessary to record the position of each macroblock in the image, as well as the motion vector, subblock type and reference frame letter of each subblock.
  • the index data should include the audio parameter and the frame number of the audio data frame.
  • Audio Parameters refers to the encoding parameters of audio data. For example, the sampling rate and sampling interval of speech coding.
  • the frame number of the audio data frame refers to the sequence number of the audio data frame corresponding to the current multimedia data packet.
  • the index data should contain the frame number of multiple audio frames.
  • the second step is the reconstruction of the multimedia data based on the index data.
  • the correspondence between the packet sequence number of the multimedia data packet and the index data is established, and the index data can be obtained according to the packet sequence number of the multimedia data packet, and then the multimedia data is reconstructed according to the index data.
  • the data status information includes the packet serial number of the error multimedia data packet, and after the sequence number of the error data packet is fed back to the reference point, the reference point can follow the packet serial number and index of the multimedia data packet in the index.
  • the relationship of the data retrieves the corresponding multimedia information, thereby reconstructing the multimedia data.
  • the reconstructed multimedia data is compared with the reference multimedia data to obtain a multimedia service performance evaluation index, and the reference multimedia data is the correctly decoded multimedia data.
  • the specific multimedia data reconstruction methods are as follows:
  • the reference point acquires the data status information of the multimedia data packet.
  • the data status information contains the sequence number of the lost packet.
  • the status information of the multimedia data packet can be obtained through a feedback channel.
  • the reference point searches for the data packet number in the index according to the serial number of the lost data packet, and retrieves the corresponding index data according to the data packet number of the index.
  • video image reconstruction is performed.
  • the specific reconstruction method is as follows:
  • the reference point searches for the index data according to the sequence number of the lost data packet, and obtains the frame number of the data frame where the lost data packet is located.
  • the data frame where the frame number is located is referred to as an erroneous data frame, and the area corresponding to the lost data packet in the erroneous data frame and The area of the subsequent multimedia data packet affected by the lost data packet is an error area, and the remaining area is a non-error area.
  • a video frame between two I frames is used as a picture group GOP (Group Of Picture), II, P2, P3, B4, P5, ... is a GOP before the next I frame.
  • GOP Group Of Picture
  • Error (such as II and P2 in Figure 6), is not bad Wrong area.
  • the error area is black, corresponding to the location where the lost packet is located, and the blank portion in P3 is the non-error area.
  • the error region is judged according to the motion vector and the reference frame information of the corresponding block, as shown by the stripe portion in B4, P5.
  • the other blank portions are non-error areas.
  • the reference video sequence is the reference multimedia data saved by the reference point, is the correctly decoded multimedia data, or has no erroneous data.
  • the audio data reconstruction is performed according to the audio reconstruction method.
  • the reconstructed multimedia data is compared with the reference multimedia data to obtain an evaluation index of the multimedia service performance.
  • the reconstructed multimedia data After the reconstructed multimedia data is obtained, it can be compared with the reference multimedia data to obtain an evaluation index of the performance of the multimedia service. For example, when evaluating the performance of a video service, the Peak Signal to Noise Ratio of the reconstructed image and the reference image can be used as an evaluation index.
  • the present invention does not require additional multimedia decoding process processing at the reference point, thereby reducing resource requirements of the reference point, and facilitating multimedia serviceability at multiple reference points at the reference point. It can be monitored, which greatly reduces the resource consumption of the reference point.
  • the method of the invention establishes an index between the packet sequence number of the transmitted multimedia data packet and the multimedia data, and returns the packet sequence number of the multimedia data packet when the multimedia data transmission error is compared, and compares the reconstructed multimedia data with the reference multimedia data to obtain a multimedia service.
  • the evaluation index of performance can easily realize the monitoring and evaluation of multimedia service performance through evaluation indicators. This simplifies the process of reconstructing multimedia data in the evaluation of multimedia service quality, and facilitates multimedia service performance monitoring of multiple sampling points at the same time.

Landscapes

  • Engineering & Computer Science (AREA)
  • Environmental & Geological Engineering (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Description

一种多媒体数据的测量方法
技术领域
本发明涉及一种多媒体信息传输和处理技术, 具体说, 涉及一种多媒体 数据的测量方法。
背景技术
随着有线网络和无线网络的发展和用户需求的驱动,多媒体技术和相关 的应用得到了越来越多的关注,被认为是未来高速网络的主流应用之一。 多 媒体应用包括视频应用和音频应用, 相对于 Internet的传统应用如 WWW、 E-mail等, 对带宽的需求更大。 现有的视频和音频等多媒体数据传输釆取了 高压缩率的编码技术, 这些编码技术利用了视频和音频等多媒体信息在时 间、 空间和统计上的相关性来进行数据压缩。压缩后的多媒体数据对传输错 误非常敏感,特别是现有的有线或者无线网络不提供传输的 QoS ( Quality of Services )保证, 数据包的丟失或者延迟不可避免, 这都会造成接收端的用 户体验质量下降。因此对多媒体服务性能进行监测成为多媒体通信领域的研 究热点。
在传统的多媒体服务性能监测方法中, 服务器端对多媒体数据进行打 包, 并将其发送到接收端, 期间要经过若干个网络节点。 在多媒体数据的传 送过程中, 多媒体数据包可能会发生延迟或者丟失,造成接收端不能正确接 收所有的数据包。 此时, 接收端通过多媒体解码处理, 找到多媒体内容中的 错误信息, 并将其反馈至服务器端。监测服务器再根据接收到的反馈错误信 息, 利用多媒体解码器估计出被监测的接收端多媒体数据。将被监测的多媒 体数据与参考多媒体数据进行比较, 实现多媒体服务性能的监测。
上述方法存在的缺点包括:
一、需要在接收端进行多媒体解码处理, 才能将由于网络传输错误引起 的多媒体内容中的错误信息反馈给服务器端 , 不利于数字版权的保护。
二、在服务器端进行多媒体服务性能评价时, 需要对每个被监测的接收 端进行完整的多媒体解码处理。因此当一个服务器端对多个接收端同时进行 监测时, 会占用大量的计算资源。
发明内容
为了解决上述技术问题, 本发明提供了一种多媒体数据的测量方法, 应 用于包括发送端、 接收端和中间网络传输节点的系统, 包括如下步骤:
( 1 )构建多媒体数据包, 在参考点建立所述多媒体数据包的包序号与 对应的多媒体数据之间的索引;
( 2 )根据采样点返回来的丟失多媒体数据包的包序号, 以及所述索引 中保存的索引数据构建差错多媒体数据;
( 3 )在参考点利用构建的差错多媒体数据和参考多媒体数据进行比较, 得出多媒体数据的质量评价结果。
进一步地, 上述方法还可具有以下特点, 在步骤( 1 )之前还包括在系 统中设置参考点和采样点, 所述参考点设置在网络传输节点 A, 所述采样点 位于网络传输节点 A或之后的任意需要进行服务监测的网络传输节点, 参 考点和采样点之间存在双向通信通道。
进一步地, 上述方法还可具有以下特点, 所述多媒体数据包的包序号是 编码后多媒体数据发送的序号, 在 RTP协议中, 多媒体数据包的包序号是 RTP包序号;或者,在 TCP/IP协议中,多媒体数据包的包序号是 IP包序号; 或者, 在 H.324M协议中, 多媒体数据包的包序号是数据帧序号。
进一步地, 上述方法还可具有以下特点, 步骤(1 ) 中, 所述索引中包 含数据包号、数据类型和索引数据, 所述数据包号表示接收或者发送的多媒 体数据包的包序号, 所述数据类型表示多媒体数据包类型, 所述索引数据表 示与数据包号对应的多媒体数据。
进一步地, 上述方法还可具有以下特点, 步骤(2 ) 中, 所述丟失多媒 体数据包的包序号通过在釆样点的解码得到。
进一步地, 上述方法还可具有以下特点, 所述在采样点的解码是在采样 点根据多媒体数据包的包序号对接收的数据包进行排序 ,检测数据包的完整 性, 如有数据包丟失, 将其包序号反馈至参考点得到。
进一步地, 上述方法还可具有以下特点, 所述在采样点的解码是在采样 点用多媒体解码器进行多媒体数据解码,如有数据包丟失,将其包序号反馈 至参考点得到。
进一步地, 上述方法还可具有以下特点, 步骤(2 ) 中, 根据所述采样 点返回来的丟失多媒体数据包的包序号, 以及所述索引中保存的信息, 区分 差错多媒体数据的差错区域和非差错区域。
进一步地, 上述方法还可具有以下特点, 步骤(2 ) 中, 利用找出的所 述差错区域和非差错区域,以及所述索引文件中保存的索引数据构建差错多 媒体数据。
进一步地, 上述方法还可具有以下特点, 步骤(3 ) 中, 所述参考多媒 体数据是指正确解码的多媒体数据, 或者是没有错误数据的多媒体数据。
进一步地, 上述方法还可具有以下特点, 所述数据类型为视频数据时, 所述索引数据包括视频参数和视频图像帧的帧号,所述视频参数是指视频序 列编码参数信息,所述视频图像帧的帧号是指当前多媒体数据包对应的视频 图像帧的序号。
进一步地, 上述方法还可具有以下特点, 所述视频参数包括图像高、 图 像宽、视频帧速率或者视频格式; 所述宏块信息包括图像块的类型、 图像块 的编号、 图像块的运动信息或者参考帧信息。
进一步地, 上述方法还可具有以下特点, 所述数据类型为视频数据时, 所述索引数据还包含宏块信息和 /或残差信息, 所述宏块信息是指用于重建 视频图像帧的图像块特征数据,所述残差信息是指图像块内部象素点的预测 值和解码值之间的差分数据。
进一步地, 上述方法还可具有以下特点, 所述数据类型为音频数据时, 所述索引数据包含包括音频参数和音频数据帧的帧号,所述音频参数是指音 频数据的编码参数,所述音频数据帧的帧号是指当前多媒体数据包对应的音 频数据帧的序号。 进一步地, 上述方法还可具有以下特点, 所述音频参数包括语音编码的 采样率和采样间隔。 进一步地, 上述方法还可具有以下特点, 步骤(1 ) 中, 所述索引中包 括包序号、 索引项的数据长度、 段信息、 宏块、 块信息或者残差信息。
进一步地, 上述方法还可具有以下特点, 在参考点按如下方法区分差错 区域和非差错区域,依据所述错误多媒体数据包对应的多媒体数据, 查找错 误多媒体数据包所在数据帧的帧号,该帧号所在数据帧称为错误数据帧,错 误数据帧中对应于错误多媒体数据包的区域及后续多媒体数据帧中受该错 误多媒体数据包影响的区域为差错区域, 其余区域为非差错区域。
进一步地, 上述方法还可具有以下特点,在参考点进行多媒体数据的重 建时, 非差错区域的重建直接从参考多媒体数据的对应位置复制得到, 对差 错区域, 按照视频差错掩盖方法或丟包补偿方法进行重建。
进一步地, 上述方法还可具有以下特点,在参考点进行多媒体数据的重 建时, 非差错区域的重建直接从参考多媒体数据的对应位置复制得到,对差 错区域进行重建时, 对差错区域直接赋值为 0, 或从参考多媒体数据中与 错多媒体数据的前一帧对应的位置的数据复制到差错区域,或根据图像的宏 块信息和残差信息按照运动补偿方法得到。
进一步地, 上述方法还可具有以下特点, 步骤(3 ) 中, 用有效峰值信 噪比方法, 对重建的差错多媒体数据和参考多媒体数据计算视频质量监测 值, 利用所述视频盾量监测值进行比较, 进行多媒体服务性能监测。
本发明方法通过建立发送的多媒体数据包的包序号和多媒体数据之间 的索引, 在多媒体数据传输错误时, 返回多媒体数据包的包序号, 根据该包 序号在索引中查找多媒体数据,进行多媒体数据重建,将重建后多媒体数据 与参考多媒体数据进行比较,得到多媒体服务性能的评价指标, 方便地实现 多媒体服务性能的监测和评价,从而简化了多媒体服务质量评价时的多媒体 数据的重建过程,有利于在参考点同时对多个采样点进行多媒体服务性能监 测。 附图概述
图 1是基于索引和反馈信息的多媒体服务质量评测系统图;
图 2是本发明实施例一索引文件的结构示意图;
图 3是本发明实施例一差错视频序列的重建方法流程图;
图 4是本发明实施例二索引的结构示意图;
图 5是 H.264中的块划分示意图;
图 6是本发明实施例二差错视频序列的重建过程示意图。
本发明的较佳实施方式
本发明通过建立发送的多媒体数据包的包序号和多媒体数据之间的索 引, 通过索引获得错误多媒体数据的位置, 重建视频, 从而简化了多媒体服 务质量评价时的多媒体数据的重建过程,有利于在参考点对多个采样点进行 多媒体服务性能监测。
下面参照附图, 对本发明的优选实施例作详细描述。
参照图 1所示, 发送端对多媒体数据进行打包, 并将其发送到接收端, 期间要经过若干个网络传输节点。在系统中设置参考点和釆样点, 参考点设 置在网络传输节点 A, 采样点位于网络传输节点 A或之后的任意需要进行 服务监测的网络传输节点, 参考点和采样点之间存在双向通信通道。 参考点 在进行打包过程中要建立发送的多媒体数据包号和多媒体数据之间的索引, 即建立发送多媒体数据包的包序号与对应的多媒体数据之间的索引。 多媒 体数据的传送过程中, 多媒体数据包可能会发生延迟或者丟失,造成采样点 不能正确接收所有的多媒体数据包。 采样点检测得到丢包的包序号, 并将其 反馈至参考点。参考点根据接收到的丢包的包序号在索引文件中查找该包所 包含的多媒体数据的区域,确定非差错区域和差错区域,对非差错区域和差 错区域进行不同的重建, 得到重建的多媒体数据, 与参考多媒体数据(正确 解码的多媒体数据 )进行比较, 实现多媒体服务性能的监测。
下面以典型的基于 H.264视频编码标准的多媒体视频服务的性能监测 为例, 进一步描述本发明的性能监测原理和实施方法,但是并不局限于该应 用。
在图 1所示的实施例中,视频数据在发送端经符合 H.264视频编码标准 的编码器编码后被打成 RTP包, 经过网络节点 1和网络节点 2后, 到达接 收端。 由于数据传送网络无法保证服务质量,接收端得到的多媒体数据可能 与发送端的数据不一致。发送端到网络节点 1之间、 网络节点 1到网络节点 2之间或者网络节点 2到接收端之间都存在数据包丟失的可能。
本发明实施例一的多媒体数据的测量方法如下:
步骤 110,构建基于索引的多媒体数据包,建立索引, 并写入索引文件。 参考点在发送多媒体码流时进行正常的解码操作,参考点在进行打包过 程中要建立发送的多媒体数据包号和多媒体数据之间的索引,即建立发送多 媒体数据包的包序号与对应的多媒体数据之间的索引。发送多媒体数据包的 包序号是编码后多媒体数据发送的序号,在基于包传输协议的流媒体传送过 程中,发送多媒体数据包的包序号是包序号和其他一些重建多媒体数据所需 要的信息。 在 RTP协议中, 发送多媒体数据包的包序号是 RTP包序号。 用 于建立索引的多媒体信息包括数据包对应的多媒体数据信息。
参考点建立索引并向索引文件写入索引数据,索引文件包含写入的索引 数据, 最终建立的索引文件结构如图 2所示,但也有可能索引文件中的字段 名字和顺序不同。
参照图 2所示, 对索引结构作详细描述。
在现有基于块的混合视频编码标准中, 索引结构应包含(不限于)以下 信息:
( 1 ) 包序号。
为索引关键字, 用于标示该包内所包含的多媒体数据。
( 2 ) 索引项的数据长度(可选) 。
标示该索引项内包含多少个字节的索引项数据,它可以不包含包序号和 其本身的数据长度, 其主要用于包的快速查找和索引。 ( 3 )段信息(Segment信息), Segment在不同的音视频编码标准中表 示不同的意思, 如 h.263中的宏块组(GOB, Group of Blocks ) , 其它视频 标准中的分片 ( Slice )和条带。
它主要包括(但不限于):每个包内包括 Segment的总数、 当前 Segment 的号码(Index ) 、 当前 Segment所在的帧号、 当前 Segment的类型、 当前 Segment所包含的宏块的开始位置、当前 Segment所包含的宏块的结束位置。
( 4 )宏块和块信息。
它主要包括(但不限于): 宏块的地址, 宏块的类型, 运动矢量, 参考 帧号。
( 5 ) 残差信息。
表示每个块经反离散余弦变换(IDCT ) 变换后的残差。
步骤 120, 在参考点, 如果想对某一接收端(即采样点)进行多媒体服 务质量进行监测和评价的时候,根据该采样点返回来的丟包的包号, 在索引 文件中进行查找,根据索引文件中保存的信息, 查找该包所包含的多媒体数 据在差错帧所在区域, 区分差错所在帧的差错区域和非差错区域。
在多媒体视频数据传输过程中建立若干个采样点。在图 1的示例中, 分 别在网络节点 1、 网络节点 2和接收端位置建立采样点 1、 采样点 2和采样 点 3。 在实际应用中, 采样点可以在任意的网络节点位置。 每个采样点对接 收的视频数据包的序号 (例如 RTP 包序号)进行判断, 然后将错误视频数 据包的序号通过反馈信道发送至参考点,参考点根据错误视频数据包的序号 可以知道在釆样点进行解码的时候已经丢失了哪些视频信息。
步骤 130, 重建带有差错的多媒体数据, 对差错区域和非差错区域使用 不同的方法进行重建。
根据釆样点返回来的丢包的包号、索引文件和参考点保存的正确的解码 视频文件, 在参考点模拟采样点快速重建带有差错的多媒体数据。
根据釆样点反馈过来的丢包的包号和索引文件就能知道釆样点丟失信 息在视频中所在的位置, 重建差错视频。 在重建差错多媒体数据的过程中, 只对差错多媒体数据中受差错影响的区域进行部分解码操作,不受差错影响 的区域直接从已经正确解码的多媒体数据中拷贝过来填充。 所谓部分解码, 是指利用索引文件中保存的信息进行多媒体数据重建,不需要从码流文件中 读取码流进行完整的解码操作。因此不需要在参考点进行额外的多媒体解码 过程处理, 从而降低了参考点的资源要求,有利于在参考点对多个采样点的 多媒体服务性能进行监测。
参照图 3所示,是用索引文件、正确解码文件和采样点反馈过来的差错 信息模拟差错视频重建的过程。
假设两个 I帧之间的视频帧 (包含前一个 I帧, 不包含后一个 I帧)作 为一个图片组(GOP, Group Of Picture ) , 如图所示, 将 II、 2Ρ、 3Β、 4Β、
5Ρ 到下一个 I帧之前作为一个 GOP。根据索引文件和丟包的包号在当 前 GOP中找到该 GOP中的第一个丢包所在的帧 (如图 3中的 3P ) , 从该 GOP的 I帧到该丟包所在帧之前的所有帧没有错误(如图 3中的 II和 2B ), 可以从正确解码文件拷贝相应的视频帧到重建视频文件相应的位置。
然后, 在丟包所在帧进行一般的错误处理, 例如最简单的方法是从前一 帧拷贝相应的块到差错位置(如图 3中 3P帧中的黑色部分) , 对于其它没 有错误的位置, 则从正确解码文件拷贝相应的图像块填补。
最后, 对受该错误影响的后续帧中的错误部分(如图 3中 4B和 5P中 的黑色部分),根据其中相应块的运动矢量和和参考帧进行预测得到相应图 像块的预测值,加上残差后即为重建的像素块的值,此即为部分解码的方法。 对该帧内其它不受错误参考帧影响的部分,则从正确解码的视频中拷贝过来 填充, 从而模拟采样点的带有差错的解码过程。
步骤 140, 服务盾量评价。
对重建的差错多媒体数据和正确解码的多媒体数据进行比较,得到评价 结果, 可以用有效峰值信噪比(EPSNR, Effective Peak Signal Noise Ratio ) 方法, 对重建的差错视频和正确解码视频计算视频质量测量(VQM )值, 将该 VQM值作为最终的评价结果。
从上面的叙述可以看出,本发明所描述的过程与传统的基于完整解码一 次差错视频的过程相比, 可以减少以下计算: 1 ) 不需要处理色度信息; 2 ) 不需要读码流和 VLC解码。 3 )不需要运动矢量的预测。 4 )不需要做 IDCT 变换。 5 )计算复杂度与丟包位置和数量有关, 第一次丢包离 I帧越远, 计 算量越低。
其中, 索引文件也可以釆用不同于实施例一的其他结构, 下面通过另一 实施例进行具体说明。
本发明实施例二具体包含如下步骤:
第一步,建立发送的多媒体数据包的包序号与对应的多媒体信息之间的 索引。
发送的多媒体数据包的包序号是编码后多媒体数据发送的序号,在 RTP 协议中, 发送的多媒体数据包的包序号是 RTP包序号; 在 TCP/IP协议中, 发送的多媒体数据包的包序号是 IP包序号; 在 H.324M协议中, 发送的多 媒体数据包的包序号是数据帧序号。 多媒体信息可以是压缩后的视频或者音频数据, 或者其它数据类型。 如图 4所示, 建立的索引应包含(不限于) 以下信息:
1、 数据包号(Packet No. ): 表示接收或者发送的多媒体数据包的包序 号, 该数据包号可以作为索引的检索关键字, 检索出需要的多媒体数据。
2、 数据类型 (Packet Type ) : 表示当前索引的多媒体数据包类型, 该 数据类型可以是视频或者音频( Video Type/Audio Type ) 。
3、 索引数据(Index Data ): 表示与数据包号对应的多媒体数据的特征 数据。 对于不同的多媒体数据类型(视频或者音频), 索引数据的内容也不 同。
A、 对于视频类型的多媒体数据, 其索引数据应包括视频参数、 视频图 像帧的帧号、宏块信息以及残差信息等。其中宏块信息和残差信息可以作为 可选的索引数据内容。 ( 1 )视频参数 ( Video Parameters )是指视频序列编码参数信息, 例如 图像高、 图像宽、 视频帧速率、 视频格式等内容。
( 2 )视频图像帧的帧号( Video Frame No. )是指当前多媒体数据包对 应的视频图像帧的序号。
当一个视频数据包中包含有多个图像帧时,索引数据应包含有多个图像 帧的帧号。
( 3 )宏块信息(Macro_block Information )是指用于重建视频图像帧的 图像块特征数据, 例如图像块的类型、 图像块的编号、 图像块的运动信息、 参考帧信息等。
当一个视频数据包中包含有多个宏块时,宏块信息应包含有多个宏块的 特征数据。 一帧图像由多个宏块组成。
( 4 )残差信息( Residual Information )是指用于重建视频图像帧的特征 数据, 包括图像块内部的每个象素点的预测值和解码值之间的差分数据。
如图 5所示,以 H.264为例 ,一帧图像 ( Picture )包含多个条带 ( Slice ) , 每个条带包含多个宏块(Macroblock )。 其中每个宏块大小为 16 χ 16象素, 它又分为 16个 4 x 4象素的子块(Block ) 。 在存储宏块信息时需要记录每 个宏块在图像中的位置, 以及每个子块的运动矢量、子块类型以及参考帧信
以 Η.264为例, 当一个宏块包含有 16个 4 x 4象素子块时, 需要记录每 个 4 x 4子块中的 16个象素的残差信息。
B、 对于音频类型的多媒体数据, 索引数据应包括音频参数和音频数据 帧的帧号。
( 1 )音频参数 ( Audio Parameters )是指音频数据的编码参数。 例如语 音编码的采样率和采样间隔等。
( 2 )音频数据帧的帧号( Audio Frame No. )是指当前多媒体数据包对 应的音频数据帧的序号。 当一个音频数据包中包含有多个音频帧时, 索引数 据应包含有多个音频帧的帧号。 第二步, 基于索引数据的多媒体数据的重建。
在索引中, 建立了多媒体数据包的包序号与索引数据之间的对应关系, 可以根据多媒体数据包的包序号获得索引数据,再根据索引数据重建多媒体 数据。
当采样点根据接收到数据状态信息,该数据状态信息包含错误多媒体数 据包的包序号,将错误数据包的序号反馈给参考点后, 参考点就可以按照索 引中多媒体数据包的包序号和索引数据的关系, 检索出对应的多媒体信息, 从而进行多媒体数据的重建。将重建的多媒体数据与参考多媒体数据进行比 较,得到多媒体服务性能评价指标, 参考多媒体数据为正确解码的多媒体数 据。
具体的多媒体数据重建方法如下:
1、 参考点获取多媒体数据包的数据状态信息。
如果釆样点出现丟包情况, 该数据状态信息包含所丟数据包的序号。 多媒体数据包的状态信息可以通过反馈信道获取。
2、 参考点根据该所丢数据包的序号查找索引中的数据包号, 根据索引 的数据包号检索出对应的索引数据。
3、 根据索引数据, 重新构建出错的多媒体数据。
若当前数据为视频类型, 则进行视频图像重建, 具体重建方法如下:
( 1 )根据视频索引数据确定非差错区域和差错区域。
参考点根据所丢数据包的序号查找索引数据,获得所丢数据包所在数据 帧的帧号, 该帧号所在数据帧称为错误数据帧,错误数据帧中对应于所丟数 据包的区域及后续多媒体数据包中受该丟失数据包影响的区域为差错区域, 其余区域为非差错区域。
如图 6所示, 假设两个 I帧之间的视频帧 (包含前一个 I帧, 不包含 一个 I帧 )作为一个图片组 GOP ( Group Of Picture ) , II , P2, P3, B4, P5,…到下一个 I帧之前为一个 GOP。根据索引数据和丟包信息在当前 GOP 中找到该 GOP中的第一个丢包所在的帧(如图 6中的 P3 ) , 从该 GOP的 I 帧到该丟包所在帧之前的所有帧没有错误(如图 6中的 II和 P2 ) , 是非差 错区域。 在丟包所在帧(P3 )中, 差错区域为黑色部分, 对应于丟失数据包 所在的位置, P3中的空白部分为非差错区域。 对受该错误影响的后续帧中, 如 B4、 P5,直到当前 GOP结束,根据其中相应块的运动矢量和参考帧信息, 判断差错区域, 如 B4、 P5中的条纹部分所示。 其它的空白部分为非差错区 域。
( 2 )对非差错区域的视频图像进行重建, 可以直接从参考视频序列中 的对应位置复制得到, 参考视频序列是参考点保存的参考多媒体数据, 是正 确解码的多媒体数据, 或者是没有错误数据的多媒体数据, 或者是发送端原 始发送的多媒体数据。
( 3 )对差错区域的视频图像进行重建, 可以对差错区域直接赋值为 0; 也可以从参考视频序列中与差错视频图像对应的视频图像前一帧的对应位 置复制得到; 也可以根据图像特征数据(宏块信息和残差信息)按照运动补 偿方法得到; 或者按照其它视频差错掩盖方法得到。
若当前数据为音频类型, 则按照音频重建方法进行音频数据重建。
( 1 )根据音频索引数据确定非差错数据和差错数据。
( 2 )对非差错数据进行重建, 可以直接从参考音频数据中复制得到。
( 3 )对差错数据进行重建, 可以对差错区域直接赋值为 0; 或者按照 其它丟包补偿方法得到。
第三步, 将重建后的多媒体数据与参考多媒体数据进行比较, 得到多媒 体服务性能的评价指标。
在得到重建的多媒体数据后, 可以将其与参考多媒体数据进行比较,得 到多媒体服务性能的评价指标。 例如, 在对视频服务性能进行评价时, 可以 使用重建图像和参考图像的峰值信噪比(Peak Signal to Noise Ratio )作为评 价指标。
当然, 索引的建立不限于本发明所描述的实施方式, 也可以采用其他结 构, 包含其他信息。
综上所述, 本发明不需要在参考点进行额外的多媒体解码过程处理, 从 而降低了参考点的资源要求,有利于在参考点对多个采样点的多媒体服务性 能进行监测, 从而大大减少了参考点的资源消耗。
工业实用性
本发明方法通过建立发送的多媒体数据包的包序号和多媒体数据之间 的索引,在多媒体数据传输错误时,返回多媒体数据包的包序号将重建后多 媒体数据与参考多媒体数据进行比较,得到多媒体服务性能的评价指标, 通 过评价指标能够方便地实现多媒体服务性能的监测和评价。从而简化了多媒 体服务质量评价时的多媒体数据的重建过程,有利于在参考点同时对多个采 样点进行多媒体服务性能监测。

Claims

权 利 要 求 书
1、 一种基于索引的多媒体数据的测量方法, 应用于包括发送端、 接 收端和中间网络传输节点的系统, 包括如下步骤:
( 1 )构建多媒体数据包, 在参考点建立所述多媒体数据包的包序号与 对应的多媒体数据之间的索引;
( 2 )根据采样点返回来的丟失多媒体数据包的包序号, 以及所述索引 中保存的索引数据构建差错多媒体数据;
( 3 )在参考点利用构建的差错多媒体数据和参考多媒体数据进行比较, 得出多媒体数据的质量评价结果。
2、 如权利要求 1所述的方法, 其特征在于, 在步骤( 1 )之前还包括 在系统中设置参考点和采样点, 所述参考点设置在网络传输节点 A, 所述釆 样点位于网络传输节点 A或之后的任意需要进行服务监测的网络传输节点, 参考点和采样点之间存在双向通信通道。
3、 如权利要求 1所述的方法, 其特征在于, 所述多媒体数据包的包 序号是编码后多媒体数据发送的序号, 在 RTP协议中, 多媒体数据包的包 序号是 RTP包序号; 或者, 在 TCP/IP协议中, 多媒体数据包的包序号是 IP 包序号; 或者, 在 H.324M协议中, 多媒体数据包的包序号是数据帧序号。
4、 如权利要求 1所述的方法, 其特征在于, 步骤( 1 )中, 所述索引 中包含数据包号、数据类型和索引数据, 所述数据包号表示接收或者发送的 多媒体数据包的包序号, 所述数据类型表示多媒体数据包类型, 所述索引数 据表示与数据包号对应的多媒体数据。
5、 如权利要求 1所述的方法, 其特征在于, 步骤(2 ) 中, 所述丟失 多媒体数据包的包序号通过在采样点的解码得到。
6、 如权利要求 5所述的方法, 其特征在于, 所述在采样点的解码是 在釆样点根据多媒体数据包的包序号对接收的数据包进行排序,检测数据包 的完整性, 如有数据包丟失, 将其包序号反馈至参考点得到。
7、 如权利要求 5所述的方法, 其特征在于, 所述在采样点的解码是 在釆样点用多媒体解码器进行多媒体数据解码,如有数据包丢失,将其包序 号反馈至参考点得到。
8、 如权利要求 1所述的方法, 其特征在于, 步骤(2 ) 中, 根据所述 采样点返回来的丟失多媒体数据包的包序号, 以及所述索引中保存的信息, 区分差错多媒体数据的差错区域和非差错区域。
9、 如权利要求 8所述的方法, 其特征在于, 步骤(2 )中, 利用找出 的所述差错区域和非差错区域,以及所述索引文件中保存的索引数据构建差 错多媒体数据。
10、 如权利要求 1所述的方法, 其特征在于, 步骤(3 )中, 所述参考 多媒体数据是指正确解码的多媒体数据, 或者是没有错误数据的多媒体数 据。
11、 如权利要求 4所述的方法, 其特征在于, 所述数据类型为视频数 据时, 所述索引数据包括视频参数和视频图像帧的帧号, 所述视频参数是指 视频序列编码参数信息,所述视频图像帧的帧号是指当前多媒体数据包对应 的视频图像帧的序号。
12、 如权利要求 11所述的方法,其特征在于, 所述视频参数包括图像 高、 图像宽、 视频帧速率或者视频格式; 所述宏块信息包括图像块的类型、 图像块的编号、 图像块的运动信息或者参考帧信息。
13、 如权利要求 4或 11所述的方法,其特征在于, 所述数据类型为视 频数据时, 所述索引数据还包含宏块信息和 /或残差信息, 所述宏块信息是 指用于重建视频图像帧的图像块特征数据,所述残差信息是指图像块内部象 素点的预测值和解码值之间的差分数据。
14、 如权利要求 4所述的方法, 其特征在于, 所述数据类型为音频数 据时, 所述索引数据包含包括音频参数和音频数据帧的帧号, 所述音频参数 是指音频数据的编码参数,所述音频数据帧的帧号是指当前多媒体数据包对 应的音频数据帧的序号。
15、 如权利要求 14所述的方法,其特征在于, 所述音频参数包括语音 编码的采样率和采样间隔。
16、 如权利要求 1所述的方法, 其特征在于, 步骤( 1 )中, 所述索引 中包括包序号、 索引项的数据长度、 段信息、 宏块、 块信息或者残差信息。
17、 如权利要求 1或 8或 9所述的方法, 其特征在于, 在参考点按如 下方法区分差错区域和非差错区域,依据所述错误多媒体数据包对应的多媒 体数据, 查找错误多媒体数据包所在数据帧的帧号, 该帧号所在数据帧称为 错误数据帧 ,错误数据帧中对应于错误多媒体数据包的区域及后续多媒体数 据帧中受该错误多媒体数据包影响的区域为差错区域,其余区域为非差错区 域。
18、 如权利要求 1或 8或 9所述的方法, 其特征在于, 在参考点进行 多媒体数据的重建时,非差错区域的重建直接从参考多媒体数据的对应位置 复制得到, 对差错区域, 按照视频差错掩盖方法或丟包补偿方法进行重建。
19、 如权利要求 1或 8或 9所述的方法, 其特征在于, 在参考点进行 多媒体数据的重建时,非差错区域的重建直接从参考多媒体数据的对应位置 复制得到, 对差错区域进行重建时, 对差错区域直接赋值为 0, 或从参考多 媒体数据中与差错多媒体数据的前一帧对应的位置的数据复制到差错区域, 或根据图像的宏块信息和残差信息按照运动补偿方法得到。
20、 如权利要求 1所述的方法, 其特征在于, 步骤(3 ) 中, 用有效峰 值信噪比方法,对重建的差错多媒体数据和参考多媒体数据计算视频质量监 测值, 利用所述视频质量监测值进行比较, 进行多媒体服务性能监测。
PCT/CN2008/000421 2007-04-29 2008-03-03 Procédé de mesure de données multimédia WO2008131636A1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN2007100977929A CN101296166B (zh) 2007-04-29 2007-04-29 基于索引的多媒体数据的测量方法
CN200710097792.9 2007-04-29
CN 200710111645 CN101316198B (zh) 2007-06-01 2007-06-01 基于索引的多媒体数据的监测方法
CN200710111645.2 2007-06-01

Publications (1)

Publication Number Publication Date
WO2008131636A1 true WO2008131636A1 (fr) 2008-11-06

Family

ID=39925192

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2008/000421 WO2008131636A1 (fr) 2007-04-29 2008-03-03 Procédé de mesure de données multimédia

Country Status (1)

Country Link
WO (1) WO2008131636A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0982874A2 (en) * 1998-08-27 2000-03-01 Matsushita Electric Industrial Co., Ltd. Adaptive transmission power control apparatus and method
WO2005006568A2 (en) * 2003-06-26 2005-01-20 Interdigital Technology Corporation Method for generating a channel quality indicator by biasing signal-to-interference ratio
JP2005192073A (ja) * 2003-12-26 2005-07-14 Matsushita Electric Ind Co Ltd マルチメディアデータ通信方法および装置
JP2006086941A (ja) * 2004-09-17 2006-03-30 Matsushita Electric Ind Co Ltd 送信方法と送信装置及び受信方法と受信装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0982874A2 (en) * 1998-08-27 2000-03-01 Matsushita Electric Industrial Co., Ltd. Adaptive transmission power control apparatus and method
WO2005006568A2 (en) * 2003-06-26 2005-01-20 Interdigital Technology Corporation Method for generating a channel quality indicator by biasing signal-to-interference ratio
JP2005192073A (ja) * 2003-12-26 2005-07-14 Matsushita Electric Ind Co Ltd マルチメディアデータ通信方法および装置
JP2006086941A (ja) * 2004-09-17 2006-03-30 Matsushita Electric Ind Co Ltd 送信方法と送信装置及び受信方法と受信装置

Similar Documents

Publication Publication Date Title
US8649278B2 (en) Method and system of multimedia service performance monitoring
US20200029130A1 (en) Method and apparatus for configuring content in a broadcast system
EP1482681B1 (en) Medium streaming distribution system
US9246630B2 (en) Method, device, and system for forward error correction
US20130254611A1 (en) Recovering data in multimedia file segments
MX2007011093A (es) Una arquitectura de descodificador para manejo optimizado de errores en multimedia continua.
RU2486679C2 (ru) Устройство и способ для хранения и чтения файла, имеющего хранилище медиа данных и хранилище метаданных
US20070127488A1 (en) Transmission format, communication control apparatus and method, recording medium, and program
EP2876881B1 (en) Method and system for determining a quality value of a video stream
KR102288815B1 (ko) Dash 표준 및 flute 프로토콜에 기초한 전송 시 패킷 손실을 처리하는 방법
CN102106146A (zh) 可缩放视频编码(svc)解码中由于增强层分组丢失而引起的错误的隐藏方法
US8300705B2 (en) Method for generating and processing hierarchical PES packet for digital satellite broadcasting based on SVC video
WO2015174894A1 (en) Methods, source device, target device and analyser for managing video coding
CN101296166B (zh) 基于索引的多媒体数据的测量方法
CN101179353A (zh) 一种多媒体服务性能监测的方法和系统
CN101056414A (zh) 一种视频数据流信源信道联合译码方法
US20140119445A1 (en) Method of concealing picture header errors in digital video decoding
US8068721B2 (en) Method of transmitting video data
JP4791443B2 (ja) 客観品質評価方法、客観品質評価装置およびプログラム
WO2008131636A1 (fr) Procédé de mesure de données multimédia
CN115604481A (zh) 提升编解码与传输并行的方法、装置及系统
CN101316198B (zh) 基于索引的多媒体数据的监测方法
JP2019029688A (ja) 通信システム及び通信方法
CN101176353B (zh) 用于流式多媒体中优化错误管理的解码器架构
Shaheen et al. Using MAC frame header for efficient multimedia streaming over IEEE 802.11 wireless LAN

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 08714875

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 08714875

Country of ref document: EP

Kind code of ref document: A1