Research And Realization Of Streaming Media Transmission Quality Based On FFmpeg Hardware Encoding

In recent years, with the continuous development of mobile terminal business and Internet technology matures, the market there have been a lot of streaming media business-related product applications, Such as video conferencing, video on demand and high-definition television and other services. Most of streaming media applications using Wireless network transmission are based on the design of mobile terminals.But with people’s increasingly strict on video quality and subjective feelings,the computation of mobile terminal on Audio and video codecs will enhance, Resulting in the lower overall processing performance and the decline of user experience. Therefore, more and more mobile phone manufacturers are considering taking in the design of the hardware encoding module. But how to allow developers to efficiently use hardware encoding function will become a common problem faced by hardware vendors.The way of transmission for media data currently is generally used with TCP or UDP protocol,However, the TCP module with its own congestion control mechanisms and network retransmission mechanism cannot meet the transmission requirements of real-time media streaming, Meanwhile the UDP protocol itself does not have any congestion control mechanisms, if we do not take in some related network congestion control algorithm, To use the UDP for network transport will inevitably results in unfriendly problem, and also aggravates its own transmission packet loss rate, delays and other issues.For the above two types of problems raised, we proposed the following solutions,in this thesis, which is based on the existing study of streaming media transmission quality control. Firstly, under the Android platform, we designed a hardware encoding scheme based on the extension of StageFright plugin into FFmpeg. Then we put forward a comprehensive assessment algorithm of the network state, combining with the network condition evaluation algorithm, we also designed the streaming adaptive transmission control algorithm, the frame rate shaping filter control algorithm. In addition, an adaptive switching algorithm of the hardware and software encoder based on network status was brought up. Finally, we designed a set of network audio and video monitoring systems, to assess the status of the transmission control system under different network conditions through a lot of tests, and verify the correctness of algorithms in this thesis.