基于FFmpeg硬件编码的流媒体传输控制算法研究与实现

发布时间：2018-05-23 23:50

本文选题：流媒体 + 硬件编码　；参考：《华南理工大学》2016年硕士论文

【摘要】：近年来,随着移动终端业务以及互联网技术的不断发展成熟,市场上流媒体相关的产品应用如视频会议、网络视频点播服务以及高清电视等也层出不穷。然而大部分产品业务都是基于手机终端设计的,随着人们对视频质量及主观感受的要求越来越严格,将会导致手机承载音视频编解码的运算加重造成整体处理计算性能降低。因此,越来越多的手机厂商在设计中都考虑加入了硬件编码模块,然而如何让开发者高效地使用硬件编码功能将是硬件厂商共同面临的问题。目前的媒体数据传输方式基本上是基于TCP或者UDP来实现的,然而TCP本身模块自带的网络重传和拥塞控制机制无法满足实时流媒体的传输要求,同时UDP协议本身并未有任何拥塞控制机制,倘若不加入相关的网络拥塞控制算法,使用UDP传输必然会造成网络不友好现象,同时加重自身传输的丢包率、延迟等问题。针对以上提出的两类问题,本论文在已有的流媒体传输质量控制的研究基础之上,提出了以下解决方案:首先,在Android系统平台下,设计了一种基于FFmpeg框架扩展Stagefright插件实现硬件编码的方案,然后提出了一套完善的网络拥塞状况评估算法,同时结合网络状况评估算法,分别设计了流媒体自适应传输控制算法、帧率整形过滤控制算法,另外提出一种自适应网络状况的软硬编码器切换算法。最后设计实现了一套网络音视频监控测试系统,并在不同的网络状况下通过大量实验测试来评估系统的传输控制状况以验证本论文算法方案的正确性。
[Abstract]:In recent years, with the continuous development of mobile terminal services and Internet technology, the application of media related products in the market, such as video conferencing, network video on demand service and high-definition TV, has emerged in endlessly. However, most of the product services are based on mobile terminals, with the quality of video and subjective feelings. More and more stringent requirements will lead to the increase of the computing performance of the handset and video coding and decoding. Therefore, more and more mobile phone manufacturers consider adding hardware coding modules in the design. However, how to make the developers use the hardware coding function efficiently will be a common problem for the hardware vendors. The former media data transmission mode is basically based on TCP or UDP. However, the network retransmission and congestion control mechanism of TCP itself can not meet the transmission requirements of real-time streaming media. At the same time, the UDP protocol itself does not have any congestion control mechanism. If it does not join the related network congestion control algorithm, the UDP transmission is used. It will inevitably cause the unfriendliness of the network, and aggravate the packet loss rate and delay of its own transmission. In this paper, based on the two kinds of problems mentioned above, the following solutions are proposed on the basis of the existing quality control of streaming media transmission. First, a FFmpeg framework is designed to extend Stage under the Android system platform. The fright plug-in implements the hardware coding scheme, and then puts forward a set of perfect network congestion evaluation algorithm. At the same time, combined with the network status evaluation algorithm, the adaptive transmission control algorithm of streaming media, the frame rate shaping filtering control algorithm are designed respectively. In addition, a soft and hard coder handoff algorithm of the adaptive network condition is proposed. Finally, the algorithm of the adaptive network condition is proposed. The design and implementation of a network audio video monitoring and testing system, and in different network conditions through a large number of experimental tests to evaluate the transmission control status of the system to verify the correctness of the algorithm in this paper.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN919.8

【相似文献】