Hiroyuki Kimiyama

Real-time processing method for an ultra high-speed streaming server running PC Linux

Integration of a high-resolution video (such as HDTV) delivery system for IP networks with legacy DBMSs and WWW systems, would make it easy to achieve system for a remote video editing, remote education and digital cinema. For this reason, we have developed a streaming server for IP networks is based on the MAPOS link-layer protocol and has a 2.48-Gbps line speed. The MAPOS protocol enables IP networks over SONET/SDH. The streaming server is constructed on a typical PC with Linux. Choosing a PC and Linux provides high performance at low cost to users. Real-time processing is important if we are to send video data at a specified rate. To implement a streaming server that can deliver Gbps video streams, we developed a real-time processing method specifically for this platform. Consequently, we obtained real-time performance for a streaming server in which the method was implemented. To evaluate its performance, we measured the receiving rate and jitter on receiving terminal while the server was delivering video streams at rates from 1.5 to 2.4 Gbps. The jitter indicates the receive-buffer size that is required and the initial delay which should be inserted before playback. For video delivery at 1.5 Gbps, the maximum buffer size was shown to be 936 KBytes (5.1 msec) and initial delay was 3.0 msec. These results show the practicability of constructing a system for ultra high-speed video (including 1.5-Gbps uncompressed HDTV) delivery with a very short response time within one video frame.

A Basic Study on High Bandwidth Streaming in Realtime over Multipath Using LDPC-IRA Codes

This paper describes a distributed video streaming system using widely disparsed storages, in which each storage hosts send chunked video packets to single receiver through multipath network. By adding parity packets by forward error correction (FEC) along with source video data in each storage hosts, this system enables realtime video streamings even if there are unbalance between those hosts. In this paper, we introduce a model of this unbalance and its effect to the amount of needed packet sending, then discuss how to design redundancy rate in FEC. The result are shown to have a trade off among the range of balancing and the additional amount of sending packet needed for stable video streaming.

An UHD video handling system using a scalable server over an IP network

Ultra-high definition (UHD) videos, such as 4K and 8K, have four times higher resolution than current HDTV videos. Broadcasting stations want to deliver these videos with as little loss as possible. However, uncompressed UHD video handling systems, such as storage and transmission systems, on IP networks are not available. We researched and developed a scalable high-speed video server system to store and deliver HD or 4K videos without compression by adapting a PC cluster technology. To handle videos with higher resolution than 4K, we also developed a method that synchronises with multiple the servers and 4K video transmission systems. We achieved 54 Gbit/s delivery performance with single cluster server comprising 24 PCs. We also implemented the synchronisation method on the server and QG-70 video transmission system. As a result, we were able to demonstrate the worlds first bi-directional, real-time, uncompressed 8K-video transmission system and the 8K video-on-demand system.

Search-Based Content Analysis System on Online Collaborative Platform for Film Production

This paper uses retrieval technology to create a video content analysis system for collaborative film production. Present film production handles all contents as digital files and demands the collaboration of remote locations. This process requires contents transfer among locations and also collaboration tools like comment-based communication. For more creative collaboration, automatic content analysis that supports editing is very effective. We propose an online collaborative platform based on retrieval technologies for film production. Our system bases automatic content analysis on content retrieval techniques for both video and audio clips. A GUI-based interface connects this function to the online collaborative system, and the total system enables efficient and secure collaboration among remote locations. In particular, our search function yields more precise detection through accurate timing, system scalability, and memory savings. Experiments on real film materials show that our search function has enough retrieval accuracy to ease the editing tasks in film production.