Electrical computers and digital processing systems: multicomput – Computer-to-computer protocol implementing – Computer-to-computer data streaming
Reexamination Certificate
2001-12-28
2004-09-07
Cuchlinski, Jr., William A. (Department: 2144)
Electrical computers and digital processing systems: multicomput
Computer-to-computer protocol implementing
Computer-to-computer data streaming
C709S233000, C709S247000, C370S232000, C370S235000
Reexamination Certificate
active
06789123
ABSTRACT:
BACKGROUND OF INVENTION
1. Technical Field
The invention is related to a codec-independent system for efficiently delivering media content, such as, for example, scalable coded audio and/or video content, over a network, such as the Internet or a wireless network, and in particular, to a system and method for automatically and dynamically delivering streaming media content which is optimally scaled in real time to match current network bandwidth and packet loss ratio.
2. Related Art
Reliable delivery of streaming audio or video media content or some combination thereof over an inherently unreliable packet-based network such as the Internet is a challenging task. During any given connection between a server and one or more clients, the available bandwidth between the server and any given client can vary greatly, and individual data packets representing encoded portions of the streaming media can be lost or delayed. Consequently, it is difficult to guarantee a smooth and consistent playback quality for streaming media.
For example, one common problem frequently observed with a network such as the Internet is that because such networks have very little guarantee of quality of service (QoS), data packets are often lost or delayed during transmission. Consequently, data packets comprising portions of media data files may arrive at a client either late, out of sequence, or may not arrive at all. Further, where data packets representing a media type of data file are lost or overly delayed beyond a predetermined minimum time constraint, the result is typically a degraded or irreparably damaged media file. Such loss or delay tends to produce noticeable artifacts in the media as the encoded packets are decoded and combined for playback on the client.
Another common problem is that the available bandwidth of a network such as the Internet typically fluctuates considerably over time for a variety of reasons, including network traffic, number of users, etc. Consequently, the available bandwidth between any given server and client, or any given source and destination, will typically fluctuate during any given connection session. Such variance in available bandwidth is not typically of great concern with non-media data files, however, with streaming media, the fluctuations can result in drastic changes in the quality of the media playback over time, along with noticeable artifacts in the playback as the playback quality changes.
In view of the aforementioned problems, a number of conventional media delivery schemes have been created in an attempt to deliver streaming media over a network such as the Internet. For example, one of the most basic schemes for streaming audio or video files simply compresses the file into a single bitstream. The packets representing this bitstream are then sent sequentially over the Internet from a server to a client where they are decoded, reassembled, and presented for playback. However, because the bitstream cannot typically be altered after it has been compressed, it is difficult to adapt to fluctuating network bandwidth conditions.
Several conventional schemes for streaming media files have expanded on the aforementioned media delivery scheme by using a multi-rate scheme to generate several compressed media files at different bit rates for each media file. The server then determines the available bandwidth between the server and the client, and sends the compressed media file having the highest bit rate that can be successfully transmitted using the given bandwidth. The server will then automatically change to either a higher or lower bit rate version of the media file, as appropriate, where the bandwidth between the server and client changes during transmission. One of the problems with switching to a file having a different bit rate is that there tends to be noticeable artifacts in file playback where file bit rate is changed during playback. Another problem is that more storage space is required on the server because multiple versions of each media file, compressed at different bit rates, must be stored to account for the available bandwidth.
The playback provided by the aforementioned schemes has been greatly improved by the simple addition of the concept of buffering. With buffering, playback of the media file is delayed on the client for a period of time, typically measured in a number of seconds. Such buffering tends to smooth out bandwidth fluctuations, thereby reducing, but not entirely eliminating the need to sometimes switch between different media file bit rates. As with the previous schemes, data packets are sometimes lost during transmission. However, where a packet is lost during transmission, the use of a buffer typically provides a window of time during which any lost packets can be retransmitted. If the retransmitted packets are received in time, the playback of the media file is not interrupted. However, if any of the retransmitted packets are not received in time, the playback of the media file will have noticeable artifacts corresponding to the lost packets.
Because lost packets can seriously degrade media playback, several schemes have been developed to address occasional packet loss. For example, several conventional schemes use an Automatic Retransmission Request (ARQ) which retransmits lost packets after the server receives a negative acknowledgement (NACK) from the client for any given packet. Such schemes begin to degrade rapidly as the packet loss ratio increases.
Other conventional schemes address the packet loss problem by using Forward Error Correction (FEC). FEC involves the transmission of parity packets along with the data packets of the media file. These parity packets can often be used to recover or regenerate lost data packets by using the received data packets along with the parity packets to recreate lost packets. Such schemes provide for a fairly reliable delivery of streaming media where the packet loss ratio is low. However, as the packet loss ratio increases, the ability of FEC schemes to recover lost packets quickly degrades, thereby also causing the playback of the media file to degrade.
Related schemes for addressing the packet loss problem go a step further by using interleaving and buffer management to disperse burst errors caused by a lost packet into random errors in the bitstream which is then further corrected by using an FEC scheme. As with the aforementioned FEC schemes, these schemes ensure a fairly reliable delivery of streaming media where the packet loss ratio is low. However, as with the previous schemes, as the packet loss ratio increases or fluctuates widely, the ability of these schemes to correct for lost packets quickly degrades, thereby again causing the playback of the media file to be degraded.
Still other schemes have achieved even better results for streaming media files over a network such as the Internet by using the concept of scalable audio or video coding. With scalable coding of audio or video, the compressed bitstream is comprised of a number of layers of decreasing importance level. As the bandwidth between the server and the client increases, packets representing more layers are transmitted. Conversely, as the bandwidth decreases, fewer packets representing layers are transmitted. Decoding of the media file can be achieved using only a subset of the layers. However, only switching among layers does not achieve an optimum transmission performance for the scalable coded media. Since there is no special processing of the lost packets, the quality of the decoded media will decrease rapidly as the packet loss ratio increases.
Therefore, what is needed is a system and method for reliably delivering streaming audio or video media content or some combination thereof over a network such as the Internet. Such a system should automatically account for fluctuations in available bandwidth between the server and client while maximizing the quality of streamed media files during client playback. Further, such a system should automatically minimize any degradation of streamed media files caused b
Li Jin
Zhou Jianping
Cuchlinski Jr. William A.
Lyon & Harr LLP
Phan Tam T.
Watson Mark A.
LandOfFree
System and method for delivery of dynamically scalable... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for delivery of dynamically scalable..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for delivery of dynamically scalable... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3189233