Patent Grant
7673063
This invention relates generally to a streaming process for media data over a communication network.
The transmission of media data, such as audio and video data, over a network is becoming more common. Users can now watch all types of media data, from a pre-recorded show to a live concert, online over the Internet or on their wireless cellular networks. Because the media data tend to take up large amounts of memory, their transmission requires vast resources. As a result, the media data are generally streamed over the network. Streaming enables the media data to be played in real time as the data are being downloaded over the network as opposed to storing the entire file first to permanent memory. Basically, media data is first divided into a sequence of frames at the sender. Each frame in the sequence contains a small portion of the media data and each frame is assigned a timestamp to indicate its position in the media data, which is generally relative to the beginning of the media. Media data frames are then sent one or a few at a time, and the receiver, such as a media player, buffers the sent frames and outputs the frames according to a timeline reconstructed based on the timestamps carried in the received frames. Streaming of the media data generally avoids the delay entailed in downloading an entire file and then playing it with a helper application at a later time.
Streaming of the media data is also becoming more prevalent within a cellular wireless setting. In fact, wireless broadcast and/or multicast services are now one of the major new features in the Third Generation Partnership Project (3GPP) and the third Generation Partnership Project 2 (3GPP2) communication networks. A major technical problem, however, arises with streaming in a wireless system. Specifically, one problem is that wireless networks tend to be more unstable than landline networks, because the wireless signal can be temporarily blocked or shadowed as the receiver moves between different environments. For example, the wireless signal can be temporarily blocked or shadowed by tall buildings in a city or tunnels through mountains or under rivers. Sometimes, the receiver can miss several minutes of the media data that may not be recoverable as the receiver may be unable to receive for up to several minutes. Although this is far less of an issue with wired Internet Protocol multimedia services, they can nevertheless suffer from similar but minor shadowing problems due to such conditions as transit congestion in a local network. All this, in turn, often causes degradation of the service quality or interruptions of the service in both wired and wireless networks.
To address these problems, one proposed method, known as Forward Error Correction, adds extra bits specifically for error correction to any character or code block of the data prior to transmission. If the transmission is received in error, the correction bits are used to check and repair the data. This method, however, only addresses data corruption caused by bit errors during data transfer or the loss of a few frames of data; it does not correct a total signal blockage.
Another proposed solution is a technique known as jitter buffers, which is widely used in many real-time media receivers such as RealPlayer® and QuickTime®. A jitter buffer is a shared data area where voice packets can be collected, stored, and sent to the voice processor in correctly spaced intervals according to their originally assigned sequence numbers. Variations in packet arrival time, called jitter, can occur because of network congestion, timing drift, or route changes. The jitter buffer, which is located at the receiving end of the voice connection, intentionally delays processing of the arriving packets so that the end user experiences a clear connection with very little sound distortion. Although this method can combat minor temporal jitters, it is impractical for correcting large delays (e.g., more than one minute). This method is also incapable of compensating for the data loss. Moreover, the use of jitter buffers tends to force the receiver to wait at the start of the media stream until the jitter buffer has been filled, but users of the wireless broadcast or multicast services may have very little tolerance for this start latency.
Still another solution, which is generally done in a point-to-point communication scenario to compensate for data loss, is to retransmit the lost data. This method, however, is impractical in a broadcast or multicast scenario due to the complexity of end-to-end synchronization between the sender and multiple receivers. Moreover, the retransmission method requires means for sending feedback from the data receiver to the sender that may not exist in most broadcast or multicast scenarios. Another proposed method is to use interpolation, which is generally used with wireless networks, but this is also not workable because it is ineffective to combat losses of consecutive frames.
The above needs are at least partially met through provision of the streaming process of media data described in the following detailed description, particularly when studied in conjunction with the drawings, wherein:
Skilled artisans will appreciate that elements in the figures are illustrated for simplicity and clarity and have not necessarily been drawn to scale. For example, the dimensions of some of the elements in the figures may be exaggerated relative to other elements to help improve understanding of various embodiments of the present invention. Also, common and well-understood elements that are useful or necessary in a commercially feasible embodiment are typically not depicted in order to facilitate a less obstructed view of these various embodiments of the present invention.
Generally speaking, pursuant to these various embodiments, a streaming process of media data has been provided, which includes an anti-shadow stream that represents a backup copy of a main media stream and an output media stream that represents an output copy of the main media stream. The anti-shadow stream, in one embodiment, is formatted according to a predefined configuration that minimizes memory usage. As such, depending on the implementation, the predefined configuration can be a lower resolution version of the main media stream, a reduced number of data frames version of the main media stream, a compressed version of the main media stream, or simply an unaltered version of the main media stream. The output copy of the main media stream, on the other hand, represents a delayed version of the anti-shadow stream according to a time offset.
Put differently, the anti-shadow stream is forward-shifted from the output media stream. According to one embodiment, the delay of the output media stream is done by sending sequenced frames of the anti-shadow stream ahead of sequenced frames of the output media stream according to the time offset. Another embodiment is to format the anti-shadow stream forward-shifted from the output media stream, while keeping the output media stream the same version as the main media stream. The main media stream, in various teachings, includes both pre-recorded and live media stream.
According to different embodiments, the anti-shadow stream and the output media stream can be sent simultaneously using two separate logical communication channels or two different radio frequencies. In one embodiment, both streams are sent simultaneously using a single sequence of data frames. As an output process, according to various embodiments, it is determined whether a selected media stream packet is stored in a main media buffer that stores the output copy of the main media stream. In another embodiment, it is also determined whether the selected media stream packet is stored in the anti-shadow buffer that stores the backup copy of the main media stream. The media stream packet is retrieved from either the main media buffer or the anti-shadow buffer and outputted. If the selected media stream packet cannot be found in either of these buffers, instructions for loss concealment of the selected media stream packet are optionally sent, according to one embodiment.
In one preferred embodiment, the media streams that are stored in the anti-shadow buffer are purged to save memory in the buffer. According to one preferred embodiment, it is determined whether an oldest packet stored in the anti-shadow buffer is older than a last packet that was outputted. If so, the oldest media stream from the anti-shadow buffer is removed from the buffer. According to various teachings, the output copy of the main media stream stored in the main media buffer is delayed from the backup copy stored in the anti-shadow buffer according to the time offset. Using the main media buffer and the anti-shadow buffer, according to an embodiment, it is determined whether a received media stream packet is part of the anti-shadow stream. If so, the received media stream packet is stored in the anti-shadow buffer. Otherwise, the received media stream is stored in the main media buffer.
Through the embodiments of various teachings, a streaming method for media data has been provided that is better suited for wireless networks and offers added flexibility in a wired network. Since the receiver is no longer required to wait at the start of the media stream for the buffer to be filled, the service start latency of the streaming process has been substantially eliminated. Moreover, long signal blockages, such as ones greater than a minute, can now be effectively compensated for through the various teachings. In fact, the entire media stream may be fully recoverable from repeated occurrences of signal blockage, since the anti-shadow buffer can be refilled once each signal blockage is over. Because the various embodiments enable the service quality during signal blockage to be controlled by the service provider, this creates new opportunities for layered pricing of the broadcast service. The various teachings further offer backward compatibility, and thus existing receiver devices that do not support the new mechanism will still work with the embodiments described below. These various embodiments also offer great flexibility since they are not platform dependent. As a result, the multiple embodiments are applicable to multiple networks, including both wired and wireless environments.
Referring now to the drawings, and in particular to
Pursuant to this example, a broadcast and multicast service (BCMCS) 12 is offered to a mobile station 14 via a mobile network 16. As with most typical wireless networks, the mobile station 14 generally requests the BCMCS 12 through multiple base stations, specifically as shown, a base station A 18 and a base station B 20. The base stations, in turn, send the mobile station's request to the BCMCS 12. To provide service to the mobile station, the BCMCS 12 generally includes a BCMCS content provider 22 that stores the main media streams 24, a BCMCS controller 26 that controls the streaming of the stored media files, and a BCMCS server 28 that transmits the media streams. These various teachings can be implemented in any of the components shown in the system 10. Any components within the system can be used to implement the various teachings described below. As an example, the description below is based on an implementation using specifically the BCMCS 12 and the mobile station 14.
Turning now to
In this particular embodiment, an anti-shadow buffer 48 is included to buffer data frames of the anti-shadow stream. As shown, the controller 46 accordingly forwards a data frame 50 bearing sequence number 108, which is identified as a next frame of the anti-shadow stream, to the anti-shadow buffer 48. The buffer 48 contains other previously sent data frames 52 (e.g., data frame sequence number 100), 54 (e.g., data frame sequence numbers 102-106), and they are in the sequential order of the data frames of the main media stream. Please note that besides examining the sequence number of the frames, other techniques to identify the next frame in the anti-shadow buffer are contemplated. For example, a timestamp of the frame can be examined, or the controller can readily determine that a frame is an anti-shadow frame because it came from the logical channel assigned for transmitting the anti-shadow stream. However, these other implementations are readily appreciated by one skilled in the art, and thus they are within the scope of the various teachings.
To make space for the newly received data frame 50, a stale frame 56 is dropped from the buffer 48, and in this case the stale frame bears the sequence number 98. Since the anti-shadow stream represents a backup copy of the main media stream, the data frames stored in the buffer 48 are generally not used unless there is a signal blockage. To provide for this scenario, the buffer 48 is directly connected a media decoder 58, which decodes the data frames and output them to the user. As noted, the data frames can be media outputs of audio or video data.
Turning to the data frames of the output media stream, a data frame 60 bearing a sequence number 100 shown as part of the output media stream 34 is forwarded to the media decoder 58 by the controller 46. In this embodiment shown, a separate main media buffer 62 is included with the media decoder 58 as the data frames 64, 66 (e.g., data frame sequence 98 and 99, respectively) are being outputted. Furthermore, in this embodiment, the data frame 56 of the anti-shadow stream is efficiently being dropped while the data frame 64 of the output media stream is being outputted, and they both represent data sequence number 98 of the main media stream. It should be noted though that the various teachings contemplate an implementation of a separate main media buffer that is not part of the media decoder 58 or without a separate main media buffer, at all. These various alternative embodiments greatly depend upon the configuration of the system and its components, and as such, they are nevertheless within the scope of the present teachings. The general overview of the mobile receiver during an uninterrupted transmission from the transmitter 32 has been shown as one embodiment and as an example.
Turning now to
Referring now to
The time offset between the output media stream and the anti-shadow stream is approximately 250 data frames. Thus, the output media stream is delayed 88 from the anti-shadow stream by approximately 250 data frames, which means that 250 data frames are recoverable at any given time. To put it differently, the anti-shadow stream is forward-shifted in the data frame sequence from that of the output media stream. Specifically, in this embodiment, there are at least 250 and 125 data frames stored in the anti-shadow buffer for the deluxe and decimated anti-shadow streams 82, 84, respectively. The deluxe anti-shadow stream 82 is substantially the same copy of the main media stream without any alternations. As a result, it is more deluxe, because there would particularly be no difference when the anti-shadow stream is used to replace the output media stream. This embodiment, however, takes large amounts of memory storage and more bandwidth overhead, which may be more expensive and undesirable for a mobile receiver but may be preferred for Internet helper applications operating on a general computer.
The decimated anti-shadow stream 84, on the other hand, is a reduced version of the main media stream. This embodiment provides efficiency for memory storage and bandwidth overhead but at the expense of the quality of anti-shadow stream. In order to reduce memory usage and bandwidth overhead, the anti-shadow stream can also be encoded with a lower resolution version or a compressed version of the main media stream while keeping the same number of frames. Of course, any combination of these features can also be implemented. There may be other data structures that have not been specifically shown, but they are within the scope of the various teachings since they are readily appreciated by a skilled artisan.
Referring now to
The transmission process, in this embodiment, is initiated 92 by a request for a pre-recorded media stream being received 94. In response to the request, it is determined 96 whether the anti-shadow stream has been requested. This step is to account for existing mobile stations that may not have been implemented with a new mechanism according to the various embodiments shown. An existing mobile station with legacy technology will likely not request an anti-shadow stream since it is not equipped to handle the two streams. If this is the case, the process sends 98 the pre-recorded main media stream, which brings the process to an end 100. If, however, the anti-shadow stream has been requested, a time offset is configured 102 for the anti-shadow stream. The anti-shadow stream is accordingly formatted 104 according to the configured time offset.
Several embodiments of the anti-shadow stream are contemplated with a pre-recorded main media stream. Since the output stream is delayed from the anti-shadow stream, one embodiment is to send the anti-shadow stream ahead of the output stream according to the configured time offset. In this case, the output stream would likely be the same as the pre-recorded main media stream, because there is no reason to format the output stream for the desired delay. This may not be workable, however, depending on the length of the start latency. When the approach of sending the anti-shadow stream ahead of the output stream is used, the receiver will receive data frames from the anti-shadow stream at the beginning (e.g., for the duration equal to the time offset). The receiver, in this case, will not start the play-out from the anti-shadow frames. Instead, the receiver buffers those arrived anti-shadow frames in its anti-shadow buffer and waits for the arrival of the first output media stream frame to start the play-out.
In another embodiment, the anti-shadow stream can be sent starting with a data frame configured to the time offset instead of starting with the first data frame. In this case, the output stream would start with the first data frame, and the anti-shadow stream that may be sent simultaneously would start with an “N” data frame of the main media stream. In this embodiment, the disadvantage is that any data lost in the first N data frames are not recoverable, but the advantage is that there would be no start latency and the output media stream does not have to be formatted to create the delay. In this embodiment, the output media stream is again preferably the same stream as the main media stream. Other embodiments using an output stream that is different from the pre-recorded main media stream are contemplated, and thus they are within the scope of the various embodiments. With these embodiments shown, the anti-shadow stream is sent 106, which is followed by the pre-recorded main media stream being sent 98. In other words, the anti-shadow stream and the main media stream will be sent simultaneously, but the anti-shadow stream content is time-shifted forward. This concludes 100 the process.
Turning now to
Different embodiments of transmitting the anti-shadow stream and the output media stream are contemplated in the transmission processes shown in
Turning now to
In this embodiment shown, the process starts 132 with a selection 134 of a media stream packet K that is to be outputted or played. In this example, since the main media stream generally contains packets of order sequenced data frames, the first media stream packet to be outputted would be first data sequence number (e.g., K=1). It should be noted, however, that a media stream packet includes one or more data frames, depending on the specific packet's implementation. Thus, any references to a media stream packet include embodiments of one or more data frames packaged in each media stream packet. After the media stream packet K is selected, it is determined 136 whether the media stream packet K is stored in a main media buffer. If so, the media stream packet K is accordingly retrieved 138 from the main media buffer. Otherwise, the media stream packet K may not have been received because of signal blockage, and in this case, it is determined 140 whether the media stream packet K can be found in the anti-shadow buffer. If so, the media stream packet K is retrieved 142 from the anti-shadow buffer to replace the media stream that is missing from the main media buffer.
Once the media stream packet K is retrieved, either from the main media buffer or the anti-shadow buffer, it is outputted 144, for example, to the media decoder in the receiver. If it turns out that the media stream packet K is not found in the anti-shadow buffer, instructions for the loss concealment for the media stream packet K may be outputted 142 to indicate that the media stream packet K has been lost and cannot be recovered. Once an appropriate output has been sent for the media stream packet K, it is determined whether an oldest media packet stored in the anti-shadow buffer is older than a last packet outputted. Using this step, any old media stream packet that is no longer needed (e.g., since it has already been outputted) is purged from the anti-shadow buffer. This way, the memory usage of the anti-shadow buffer can be conserved for relevant data frames.
Accordingly, if an oldest media packet stored in the anti-shadow buffer is older than the last packet outputted, it is removed 150 from the anti-shadow buffer and loops back to check again until this is no longer the case. Once there are no more oldest media packets in the anti-shadow buffer that is older than a last packet outputted, it is next checked to determine 152 whether the media presentation should continue. If not, the process ends 154. Otherwise, K is set 156 to K+1 to run the process for a next media stream packet, which restarts the process from the selection 134 of the new media stream packet K.
Turning now to
With these various teachings shown, a technique of streaming media data has been provided. As a result of the various teachings shown, long signal blockages, such as ones greater than a minute, can now be effectively compensated since a separate backup copy of the main media stream is transmitted along with the output copy. The entire and/or large portions of the main media stream are fully recoverable, since the anti-shadow buffer can be refilled once the signal blockage is over. As a result, higher quality of service is provided, because the negative affects of the signal blockages have been minimized. Moreover, because the service quality during signal blockage can be specifically controlled by the service provider, new opportunities for layered pricing of the broadcast service are created. The various teachings can easily be implemented with a legacy mobile station since they are not platform-specific, thus offering backward compatibility. The receiver is no longer required to wait at the start of the media stream for the filling of the buffer. Thus, the service start latency of the streaming process has been substantially eliminated as compared to other helper applications. These and other benefits, as a result, provide a streaming process of media data that is better suited for wireless networks while at the same time provide added flexibility in a wired network.
Those skilled in the art will recognize that a wide variety of modifications, alterations, and combinations can be made with respect to the above described embodiments without departing from the spirit and scope of the invention, and that such modifications, alterations, and combinations are to be viewed as being within the ambit of the inventive concept.
| Number | Name | Date | Kind |
|---|---|---|---|
| 6272190 | Campana, Jr. | Aug 2001 | B1 |
| 6442694 | Bergman et al. | Aug 2002 | B1 |
| 6658073 | Grunert | Dec 2003 | B1 |
| 6674755 | Potter et al. | Jan 2004 | B1 |
| 6820133 | Grove et al. | Nov 2004 | B1 |
| 6904059 | Newson et al. | Jun 2005 | B1 |
| 7266609 | Bill et al. | Sep 2007 | B2 |
| 7483487 | Liu et al. | Jan 2009 | B2 |
| 7548995 | Thukral | Jun 2009 | B2 |
| 7577714 | Saunders et al. | Aug 2009 | B2 |
| 20020161847 | Weigand et al. | Oct 2002 | A1 |
| 20030099202 | Lear et al. | May 2003 | A1 |
| 20040093618 | Baldwin et al. | May 2004 | A1 |
| 20040123324 | Sazzad et al. | Jun 2004 | A1 |
| 20040223058 | Richter et al. | Nov 2004 | A1 |
| 20050013249 | Kong et al. | Jan 2005 | A1 |
| Number | Date | Country |
|---|---|---|
| 2004086748 | Oct 2004 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20060085551 A1 | Apr 2006 | US |
