Research On Panoramic Video Live Broadcast Technology With Adaptive Rate

With the integration of Virtual Reality in the live broadcasting industry,the VR live broadcast industry has developed rapidly.As one of the main contents of VR live broadcast,panoramic video provides users with a 360-degree viewing angle,enabling users to obtain a good immersive experience.However,the huge amount of data of panoramic video has brought huge challenges to its network transmission.Tile-based dynamic adaptive transmission scheme of panoramic video can effectively solve the above-mentioned problems by dividing tiles to transmit high-quality video within the user’s field of view and low-quality video outside the field of view.This paper focuses on the research of panoramic video live broadcast technology,designs and implements a tile-based panoramic video live broadcast system,and further studies the perspective switching strategy and adaptive scheduling algorithm.The research results of this paper can effectively reduce the live broadcast delay and improve the user’s viewing experience.This article first designs and implements a panoramic video live broadcast system based on Tile.In this system,the server is responsible for the generation of Media Presentation Description files,and the client is responsible for the update of MPD files and the decoding and rendering of panoramic videos.On this basis,this article combines the MPEG-DASH standard protocol and proposes an MPD update scheme suitable for live broadcast scenes.This solution reduces the size of a single MPD file by dividing a standard MPD file into multiple small MPD files connected in multiple stages,thereby increasing the proportion of effective information and reducing the live broadcast delay.The experimental results show that the MPD update scheme proposed in this paper effectively shortens the live broadcast delay.In a tile-based panoramic video live broadcast system,viewing angle switching will cause low-quality images in some areas of the field of view,making it difficult to switch to highquality video images in a timely manner,which greatly reduces user experience.In order to solve the above-mentioned problems,this paper proposes a rapid viewing angle switching technology based on RAPF.In the process of viewing angle switching,this technology provides random access images for subsequent video streams,enabling fast display and playback of high-quality video content in the field of view.At the same time,the technology uses the coding method of inter-frame prediction to reduce the code rate of the RAPF stream,thereby reducing bandwidth consumption.Experimental results show that the fast viewing angle switching technology proposed in this paper can effectively shorten the viewing angle switching delay and improve user experience while reducing bandwidth consumption.In addition,due to the fluctuation of network bandwidth and the existence of viewing angle prediction errors,the experience of the panoramic video live broadcast system has been affected to a certain extent.In order to solve the above problems,this paper combines the Model Predictive Control algorithm,and proposes a bit rate adaptive algorithm with a fixed viewing angle and a bit rate adaptive algorithm for switching viewing angles for different scenarios.The two algorithms use the user experience quality as a measure to select an appropriate coding quality level for the tile in the prediction field of view to ensure the best video quality in the field of view.Experimental results show that both bit rate adaptive algorithms can effectively improve the video quality in the user’s field of view and optimize the user experience.