Locating points of interest using references to media frames within a packet flow

<u>CROSS REFERENCE TO RELATED APPLICATIONS</u>

Not applicable.

<u>TECHNICAL FIELD</u>

The present disclosure relates generally to digital multimedia streams.

<u>BACKGROUND</u>

Many consumers receive entertainment programming carried via digital signals. Programming (e.g., television programs, movies, sports events, music, etc.) is encoded at a head-end using a compression standard such as Motion Pictures Experts Group (MPEG) or H.264. The resulting stream of frames is transmitted to a media receiver, which may decode and display the stream, store the stream for later viewing, and/or further process the encoded stream.

The digitally encoded stream is carried over a core network (such as the Internet) downstream to the edge of the network, where the core network interfaces with a subscriber access network. The digitally encoded stream may be transformed or consumed by one or more intermediate processing functions located at the edge. Examples of such intermediate processing functions include network-based digital video recording, video on demand, fast channel change, video error repair, encryption/key generation, and digital program insertion. Some of these intermediate processing functions operate by locating key frames or other "points of interest" within the encoded stream. Conventional solutions place the functionality of locating points of interest within the intermediate processing function. However, locating points of interest is relatively compute-intensive, since this involves examining the structural information in each encoded elementary stream and in the multiplexed transport stream when transport streams are used. When such conventional solutions also encrypt the transport stream and/or the elementary stream, then locating points of interest also requires first decrypting the transport stream packets. Thus, a more flexible technique for locating points of interest, and for communicating this information to intermediate processing functions, is desired.

<nplcit id="ncit0001" npl-type="b"><text>Jack Brassil et al.: "Structuring Internet Media Streams With Cueing Protocols" IEEE/ACM TRANSACTIONS ON NETWORKING, IEEE/ACM New York, NY, US, vol. 10, no. 4, 1 August 2002, pages 1-11</text></nplcit>, describes a method of notifying events in a media stream by use of cueing messages.

<patcit id="pcit0001" dnum="US2005169174A"><text>US 2005/169174</text></patcit> relates to a method of creating a transport stream for a media stream using cueing data to generate the transport schedule for the transport stream.

"<nplcit id="ncit0002" npl-type="s"><text>MPEG-7 Overview" VIDEO STANDARDS AND DRAFTS, XX, XX, no.N6828, 5 February 2005, pages 1-76</text></nplcit>, describes a standard for annotating media streams using an extension of XML.

<u>BRIEF DESCRIPTION OF THE DRAWINGS</u>

Many aspects of the disclosure can be better understood with reference to the following drawings. The components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present disclosure. <ul id="ul0001" list-style="none"><li><figref idref="f0001">FIG. 1</figref> is a block diagram of an environment in which one embodiment of a system and method for utilizing locating points of interest is located.</li><li><figref idref="f0002">FIG. 2A</figref> is a block diagram of an embodiment in which a media stream is received by a point-of-interest producer and by a point-of-interest consumer.</li><li><figref idref="f0003">FIG. 2B</figref> is a block diagram of another embodiment in which a point-of-interest producer parses a media stream to locate points-of-interest, and produces a combined stream.</li><li><figref idref="f0004">FIG. 2C</figref> is a block diagram of yet another embodiment in which a point-of-interest producer parses a media stream to locate points-of-interest, and produces a new stream, which is an annotated version of the original media stream.</li><li><figref idref="f0005 f0007">FIGs. 3A-C</figref> are block diagrams showing how syntax and encoding of points-of-interest information is performed in an embodiment of a system and method for utilizing locating points of interest.</li><li><figref idref="f0008">FIG. 4</figref> is a block diagram showing how syntax and encoding of points-of-interest information is performed in another embodiment of a system and method for utilizing locating points of interest.</li><li><figref idref="f0009">FIG. 5</figref> is a flowchart of a process implemented by one embodiment of the point-of-interest producer from <figref idref="f0001">FIG. 1</figref>.</li><li><figref idref="f0010">FIG. 6</figref> is a flowchart of a process implemented by one embodiment of the point-of-interest consumer from <figref idref="f0001">FIG. 1</figref>.</li><li><figref idref="f0011">FIGs. 7A</figref> and <figref idref="f0012">7B</figref> are blocks diagram of various embodiments of point-of-interest producer 150 and point-of-interest consumer 160.</li><li><figref idref="f0013">FIG. 8</figref> is a block diagram showing selected components of a point-of-interest consumer or producer from <figref idref="f0001">FIG. 1</figref>.</li></ul>

<u>DESCRIPTION OF EXAMPLE EMBODIMENTS</u>

OVERVIEW

The invention is defined in the appended independent claims. In one embodiment, a method comprises receiving a stream of transport packets encapsulating media packets, and generating a reference to the point-of-interest media packet relative to the location of the point-of-interest within the transport packet stream. The reference is generated for each media packet that is identified as a point-of-interest.

In another embodiment, a system comprises memory with logic, and a processor. The processor is configured with the logic to receive a stream of transport packets encapsulating media packets, and to identify one or more of the media packets as a point-of-interest. The processor is further configured with the logic to generate a reference to the point-of-interest (POI) media packet relative to the location of the POI within the transport packet stream. The reference is generated for each media packet identified as a point-of-interest.

In another embodiment, a system comprises means for receiving a stream of transport packets encapsulating media packets, and identifying one or more of the media packets as a point-of-interest. The system further comprises means for generating a reference to the point-of-interest (POI) media packet relative to the location of the POI within the transport packet stream.

The reference is generated for each media packet identified as a point-of-interest.

<u>EXAMPLE EMBODIMENTS</u>

<figref idref="f0001">FIG. 1</figref> is a block diagram of an environment in which one embodiment of a system and method for utilizing locating points of interest using references to media frames within a packet flow is located. A system 100 delivers digital multimedia, video and/or television services to subscribers over an Internet Protocol (IP) network 110.

Media source 120 encodes, formats, and transmits a digital media stream. The encoded media stream is then communicated to a media stream receiver 130, which stores, decodes, and/or renders the media stream. Media stream receiver 130 may take the form of (for example) a digital set-top box, a digital video recorder, a home media server, a personal computer, a personal digital assistant, a mobile phone, etc. In this disclosure, the term "media stream" refers to a stream that includes video frames, audio frames, multimedia, or any combination thereof. Common encoding formats for digital media streams include MPEG-2, MPEG-4, H.264, and VC-1. In some environments, the encoded media stream represents a single program, and thus contains a video and an audio stream multiplexed together into a single program transport stream (SPTS).

System 100 also contains one or more components that are downstream of media source 120, and which perform intermediate processing of the encoded media stream. <figref idref="f0001">FIG. 1</figref> contains one example of such a component, a video-on-demand (VOD) server 140. VOD server 140 ingests a media stream from media source 120 and plays the stream out at a later time in response to a user request ("on demand"). VOD server 140 also processes the ingested stream to provide "trick mode" capabilities, such as fast-forward, rewind, slow-motion, and pause. Another example of an intermediate processing component is a fast channel change server (not shown), which ingests a media stream from media source 120, and stores, into a cache, a moving "window" through the stream. This cache represents the last few seconds of the stream contents. On request, the fast channel change server plays out a portion ("burst") of the cached window, in order to accelerate the channel change. Without such processing by a fast channel change server, a user typically experiences a delay in displaying the new stream that results from a channel change, during which time the screen might be frozen or blank.

These and other intermediate processing units operate by locating key frames or points of interest (POI) within the encoded stream. Key frames are video frames which can be decoded without reference to another frame. VOD server 140 provides some trick-mode functions by creating a trick mode stream which contains only key frames, or references to such frames. A fast channel change server uses key frames to resynchronize a decoder after a channel change. In addition to key frames, other points of interest in the encoded stream include program-specific information (PSI). Examples of PSI are program map tables (PMTs), program allocation tables (PATs), and entitlement control messages (ECM). PATs and PMTs both provide fundamental structural information about a stream, which a decoder uses to find and process the elements of that stream. ECMs contain keying information that is used to decrypt encrypted content appearing later in the stream. Having these elements explicitly identified as points of interest allows intermediate processing functions to determine which pieces to cache, without having to parse the stream. In addition to the two examples of intermediate processing discussed here (fast channel change and video-on-demand), a person of ordinary skill in the art will be aware of other intermediate processing functions which utilize points of interest, and the principles disclosed herein apply to these as well.

Using inventive features disclosed herein, point-of-interest producer 150 generates references to these points of interest within the media stream, and provides these references to an intermediate component (e.g., a consumer of the points of interest) such as VOD server 140. Point-of-interest consumer 160 uses the references to locate the points of interest within the media stream. An intermediate component associated with point-of-interest consumer 160 uses the points of interest to perform its intermediate function (e.g. produce a trick mode stream). In the example of <figref idref="f0001">FIG. 1</figref>, this intermediate component is VOD server 140, and point-of-interest consumer 160 resides within VOD server 140. However, a person of ordinary skill in the art should appreciate that, in other embodiments, point-of-interest consumer 160 is separate from, but in communication with, an intermediate function such as VOD server 140.

Point-of-interest producer 150 can use different mechanisms to provide point-of-interest references to point-of-interest consumer 160. <figref idref="f0002">FIG. 2A</figref> is a block diagram of an embodiment in which media stream 210 is received by a point-of-interest producer 150' and by a point-of-interest consumer 160'. Media stream 210 is a stream 220 of media packets that are either elementary stream packets, or elementary streams encapsulated within transport layer packets 230. The inventive concepts described herein apply to various types of elementary stream encapsulations, including (but not limited to): raw MPEG2 Transport Stream (TS) over legacy transport; MPEG2 Elementary Stream (ES) over UDP/IP, RTP/UDP/IP and RTP/TCP/IP; MPEG2 TS over UDP/IP, RTP/UDP/IP and RTP/TCP/IP.

Point-of-interest producer 150' examines media stream 210 to identify one or more media packets 220 that contain points-of-interest. In the example embodiment of <figref idref="f0002">FIG. 2A</figref>, POI producer 150' generates a corresponding, but separate, points-of-interest stream 240 that contains references to these "interesting" media packets 220 within media stream 210. Points-of-interest stream 240 is a stream of POI packets 250, where a POI packet 250 contains one or more references to media packets 220, within media stream 210, that were identified as points of interest. (In <figref idref="f0002">FIG. 2A</figref>, these references are represented symbolically by lines 260. More details about mechanisms for implementing references to media packets 220 will be discussed later in connection with <figref idref="f0005 f0006 f0007">FIG. 3</figref>.

Point-of-interest consumer 160' receives points-of-interest stream 240, and also receives the original media stream 210. Point-of-interest consumer 160' uses points-of-interest stream 240 to locate individual points-of-interest (e.g., key frames, program-specific information) within media stream 210. In the example scenario of <figref idref="f0002">FIG. 2A</figref>, four media packets (220A-D) are identified as points-of-interest, and thus points-of-interest stream 240 contains four POI references (260A-D). Once the points-of-interest are located, an intermediate processing function (not shown) uses the points-of-interest in a manner appropriate to its function.

<figref idref="f0003">FIG. 2B</figref> is a block diagram of an alternative embodiment in which point-of-interest producer 150" parses media stream 210 to locate points-of-interest, and produces a combined stream 270. Combined stream 270 contains transport layer packets (not shown) from media stream 210, where transport layer packets contain media packets 220. Combined stream 270 also contains POI packets 250 that refer to those media packets 220 within media stream 210 that are identified as points-of-interest. Point-of-interest consumer 160" examines combined stream 270, and uses the references within POI packets 250 to locate point-of-interest payloads in combined stream 270. In the example scenario of <figref idref="f0003">FIG. 2B</figref>, three media packets (220A-C) are identified as points-of-interest, and thus points-of-interest stream 240 contains three POI references packets 250A-C, each containing a reference 260A-C to a POI packet 230A-C. Once the points-of-interest are located, an intermediate processing function (not shown) uses the points-of-interest in a manner appropriate to its function. In yet another embodiment (not shown), the functionality of point-of-interest producer 150 is integrated with media source 120. In other words, media source 120 generates references to points-of-interest within the media stream as the media stream itself is generated.

<figref idref="f0004">FIG. 2C</figref> is a block diagram of another alternative embodiment in which point-of-interest producer 150'" parses media stream 210 to locate points-of-interest, and produces a new stream 280, which is an annotated version of original media stream 210. Annotated stream 280 contains transport layer packets 230 from media stream 210, where transport layer packets 230 contain media packets 220. Information carried in headers 290 on the transport layer packets 230 identifies whether or not the media packets 220 contained within are points-of-interest, and if so, provides further information about the point-of-interest. Point-of-interest consumer 160'" examines annotated stream 280, and uses the annotations within the transport layer headers 290 to determine which media payloads in annotated stream 280 are points-of-interest. In the example scenario of <figref idref="f0004">FIG. 2C</figref>, two media packets (220A, B) are identified as points-of-interest by information in their respective transport layer headers 290A, 290B.

In some embodiments, RTP is used as the transport layer protocol, and the point-of-interest information is conveyed in the RTP header using RTP header extensions. The RTP header extension is formed as a sequence of extension elements, with possible padding. Each extension element has a local identifier and a length. Further details on RTP header extensions can be found in Internet draft-ietf-avt-rtp-hdrext-12.txt ("A general mechanism for RTP Header Extensions"). Examples of local identifiers in this context include point-of-interest classifier, SMPTE time code information, or MPEG stream hints that can be used for decoder optimization, trick play state generation, etc.

The concept of a frame or PSI as being a "point-of-interest" should be familiar to a person of ordinary skill in the art, as will various mechanisms used to select "points of interest". Commonly used selection criteria include selecting the start of all I-frames, all PSIs, all ECMs, or various combinations of these criteria. In one embodiment, selection criteria involves selecting anchor frames. In this disclosure, an anchor frame is an I-frame, instantaneous decoder refresh frame (IDR-frame), or a frame that depends only on a past single reference frame that is the most-recently decoded anchor frame.

As described above in connection with <figref idref="f0002 f0004">FIGs. 2A-C</figref>, the same technique for identifying and encoding the POI information described herein can be used in many different ways: the POI information can be included in the steam as originated (e.g. included by the encoder); the media stream can be processed to produce a new stream of the same or a different protocol with POI annotations; or a separate POI annotation stream can be produced, leaving the original stream untouched. Having described in general the relationship between media stream 210 and the other streams (210, 220, 270, 280) the syntax and encoding of points-of-interest information will now be described in more detail in connection with the diagrams of <figref idref="f0005 f0007">FIGs. 3A-C</figref>.

As can be seen in <figref idref="f0005 f0007">FIGs. 3A-C</figref>, media stream 210 is formatted as a stream of application-layer packets, referred to earlier as media packets 220. A media packet 220 encapsulates items such as a video frame, PSI, audio segments, etc. In some embodiments, these items can be split across media packets 220, so that one media packet 220 contains only a portion of a video frame, PSI, or audio segment.

Media packets 220 are encapsulated within transport layer packets 230, where each transport layer packet 230 includes a transport layer header 230H and a transport layer payload 230P. Transport layer packets 230 are in turn encapsulated by additional protocol layers, shown here as an additional header 310. In the embodiments described herein, Internet Protocol (IP) is used as the network layer and Real-time Transport Protocol (RTP) in combination with User Datagram Protocol (UDP) is used as the transport layer, with the media packets multiplexed into an MPEG2 Transport Stream. A person of ordinary skill in the art should be familiar with MPEG2 transport, IP, UDP and RTP packet formats, so headers and payloads have been simplified in <figref idref="f0005 f0007">FIGs. 3A-C</figref>.

As described earlier, point-of-interest producer 150 locates key frame media packets and program-specific information (PSI) media packets, and generates a reference to each of these points of interest. The first level reference to a particular media packet 220 identifies the containing transport layer packet 230. A particular transport layer packet 230 is identified by a flow identifier 320 and a flow-unique packet identifier 330. A person of ordinary skill in the art should be familiar with the concept of a packet flow, which consists of a series of packets between two endpoints, where the two endpoints are identified by information contained in each packet header. An RTP flow is defined by a destination address and a destination port (both part of IP/UDP header 310), and a synchronization source identifier (SSRC, part of transport layer header 230H). Thus, the combination of destination address, destination port, and SSRC serves as a flow identifier 320 for an RTP flow. In the case of RTP, the packet identifier 330 is a sequence number 330.

In the example scenario shown in <figref idref="f0005">FIG. 3A</figref>, point-of-interest producer 150 processes transport layer packet 230-A, and identifies two points of interest, namely, 220-A1 and 220-A2. References to these two points-of-interest are then added to points-of-interest stream 240 (shown in <figref idref="f0002 f0004">FIGs. 2A-C</figref>). Since both points of interest are located in the same transport layer packet (230-A), the first reference begins with a flow reference 350 and a transport packet reference 360, which together identify a particular transport layer packet (here, 230-A). In <figref idref="f0005 f0006 f0007">FIG. 3</figref>, flow reference 350 "points" to the flow identifier fields of transport layer packet 230-A, but a person of ordinary skill in the art should appreciate that this implies that flow reference 350 has the same values as the flow fields (destination address, destination port, SSRC) in transport layer packet 230-A. Similarly, transport packet reference 360-A "points" to the packet identifier field 330 of transport layer packet 230-A, which implies that transport packet reference 360-A is set to the same value as the packet identifier field 330.

Since a transport layer packet can carry multiple media packets, the reference to media packet 220-A1 continues with a media packet reference 370-A1. In <figref idref="f0005">FIG. 3A</figref>, media packet reference 370-A1 "points" to media packet 220-A1 as the first media packet within the transport layer packet 230. A person of ordinary skill in the art should appreciate that the symbolic representation used here (arrows between packet fields) can be implemented by setting media packet reference 370 to any value that specifies one packet in the sequence of media packets conveyed in transport layer packet 230. For example, media packet reference 370 can be a media packet number (e.g., first, second, etc.) or a byte offset (e.g., 0, 1 x media packet size, etc.). In this example embodiment, the reference to media packet 220-A1 concludes with a point-of-interest classifier 380-A1 or type, further describing the point-of-interest. For example, a classifier 380 can describe a key frame, a particular type of key frame (e.g. I-frame), a PSI, or a particular type of PSI.

Point-of-interest producer 150 then adds a second reference to points-of-interest stream 240, the reference to point-of-interest 220-A2. Since this point-of-interest is contained in the same transport layer packet (230-A), it is unnecessary to repeat the flow reference (350) and the transport reference (360-A). The second point-of-interest (220-A2) is simply identified as the last media packet (370-A2) in transport layer packet 230-A.

<figref idref="f0006">FIG. 3B</figref> continues the example scenario. At this time, point-of-interest producer 150 processes a second transport layer packet 230-B, but finds no points of interest within transport layer packet 230-B. Therefore, no new references are added to points-of-interest stream 240.

<figref idref="f0007">FIG. 3C</figref> completes the example scenario. At this time, point-of-interest producer 150 processes a third transport layer packet 230-C, and finds one point-of-interest. Transport layer packet 230-C belongs to the same flow as the other two transport packets. Therefore, it is unnecessary to repeat the flow reference (350). The third point-of-interest (220-C) is identified by a new transport reference (360-C) and a new media packet reference (370-C).

The embodiment of <figref idref="f0005 f0006 f0007">FIGs. 3A-3C</figref> uses a combination of RTP and UDP as a transport layer. Another embodiment that uses UDP without RTP is now described in connection with <figref idref="f0008">FIG. 4</figref>. Media packets 220 are encapsulated within UDP packets 410, where each UDP packet 410 includes a UDP header 410H and a UDP payload 410P. UDP packet 410 is in turn encapsulated by additional protocol layers, shown here as an additional IP header 420. A particular UDP packet 410 belongs to a flow 430, identified by source IP address, destination IP address, protocol identifier (UDP), source UDP port, and destination UDP port.

In the example scenario shown in <figref idref="f0008">FIG. 4</figref>, point-of-interest producer 150 processes transport layer packet 230-J, and identifies one point-of-interest, namely 220-J1. References to this point-of-interest are then added to points-of-interest stream 240. The first reference begins with a flow reference 440 that partially identifies transport layer packet 230-J. In <figref idref="f0005 f0006 f0007">FIG. 3</figref>, flow reference 350 "points" to the flow identifier fields of transport layer packet 230-J, but a person of ordinary skill in the art should appreciate that this implies that flow reference 350 has the same values as the flow fields (source and destination address, source and destination port) in transport layer packet 230-J.

The partial identification of transport layer packet 230-J is completed by transport packet reference 450. In the embodiment described above, RTP sequence numbers are used as transport packet references. However, UDP does not use sequence numbers, and there is nothing that is conveyed in the UDP packet itself which uniquely distinguishes one UDP packet from another. The embodiment of <figref idref="f0008">FIG. 4</figref> uses an inventive technique whereby UDP packets are identified by a checksum. This checksum is computed by point-of-interest producer 150 and used as transport packet reference 450 within points-of-interest stream 240. Point-of-interest consumer 160 then computes a checksum on incoming packets, and uses this computed checksum as a transport packet identifier. In this manner, point-of-interest consumer 160 is able to distinguish among UDP packets belonging to the same flow, even though the packets themselves do not carry a unique identifier. In these embodiments, the checksum can be computed over any combination of the UDP header and UDP payload (which encapsulates an MPEG Transport Packet). In one example, the identifier is a checksum for UDP packet 410.

The embodiment of <figref idref="f0008">FIG. 4</figref> uses a media packet reference 460 to specify a particular media packet 220 within transport layer packet 230-J. Media packet reference 460 is similar to the one described above for the UDP +RTP embodiment of <figref idref="f0005 f0007">FIGs. 3A-C</figref>, specifying either a packet number offset or a byte offset within the UDP payload 410P. This embodiment can also use a point-of-interest classifier 380 or type, further describing the point-of-interest (e.g., key frame, I-frame, B-frame, P-frame, PSI, PMT, PAT, ECM, etc.)

<figref idref="f0009">FIG. 5</figref> is a flowchart of a process implemented by one embodiment of point-of-interest producer 150. The process 500 begins at block 510, where a stream of media packets encapsulated in a transport stream is received. Next, at block 520, the media packets are examined, and one or more frames or PSIs in the media packets are determined to be points of interest. The concept of a "point-of-interest" in a media stream should be known to a person of ordinary skill in the art, as should be various mechanisms for determining which frames or PSIs are points of interest. A few simple examples of selection criteria are selecting I-frames, selecting anchor frames (where anchor frames are described above), selecting PMTs, selecting PATs, and selecting ECMs, as well as combinations of these criteria.

Processing continues at block 530, where each point-of-interest is identified with a reference to a containing transport packet, and media packet within the transport packet. As described above, in one embodiment the reference is a combination that identifies the flow, the particular transport packet within the flow, and the particular media packet within the transport packet. In some embodiments, the reference also includes a description of the point-of-interest, such as frame or PSI, type of frame (I, B, P, etc.) and type of PSI (PMT, PAT, etc.).

Next, at block 540, a packet or series of packets that contain the references to points-of-interest are generated. As described above (in connection with <figref idref="f0004">FIG. 2C</figref>), some embodiments modify the transport layer headers of the original media packets to include point-of-interest information (such as a point-of-interest classifier), while other embodiments (described above in connection with <figref idref="f0003">FIG. 2B</figref>) generate reference packets or "point-of-interest" packets which are separate and distinct from media packets. A person of ordinary skill in the art should realize that when separate media and reference packets are used, the protocol used to carry the references is not required to be the same as the protocol that encapsulates the media packets (although the same protocol can be used). For example, one embodiment uses UDP to encapsulate MPEG-2 transport packets, but RTP to encapsulate the points-of-interest stream.

Processing continues at block 550, where the stream of packets containing the references to points-of-interest is transmitted to a receiver containing point-of-interest consumer 160 (<figref idref="f0001">FIG. 1</figref>). At block 560, which is optional, the packets containing the multimedia flow (received at block 510) are transmitted to the same receiver. These data flows may be in real time, or may be done in a non-real-time mode.

<figref idref="f0010">FIG. 6</figref> is a flowchart of a process implemented by one embodiment of point-of-interest consumer 160. The process 600 begins at block 610, where a stream of media packets encapsulated in a transport stream, and a stream of corresponding point-of-interest reference packets is received. At block 620, the next point-of-interest reference packet is examined. Block 630 uses the information within to locate the appropriate transport packet in the received media stream. In this example embodiment, the flow reference and transport packet reference fields within the reference packet are used to find a transport packet in the media stream having a matching flow identifier and sequence identifier.

Processing continues at block 640, where additional information within the current point-of-interest reference packet is used to locate the appropriate media packet within the already-identified (at block 630) transport packet. In this example embodiment, the media packet reference field within the reference packet is used to find the appropriate media packet within the series (e.g., the first, second, etc.).

Next, at block 650, the media packet found at block 640 is processed by some intermediate processing function, or is handed off by such an intermediate processing functions. The role of intermediate processing functions (such as trick mode stream generation in video-on-demand servers and locating key frames in fast channel change servers) is discussed above, and should be familiar to a person of ordinary skill in the art.

Block 660 determines whether any reference packets remain to be processed. If Yes, then processing starts for the next reference packet, at block 620. If No, process 600 is complete.

<figref idref="f0011">FIG. 7A</figref> is a block diagram of an embodiment of point-of-interest producer 150 and point-of-interest consumer 160, in which IP multicast is used to deliver both the primary media stream and the points-of-interest stream. Point-of-interest producer 150 produces one points-of-interest stream 710, and transmits this points-of-interest stream 710 to several instances of point-of-interest consumer 160 using a IP multicast flow 720. As explained above, points-of-interest stream 710 is associated with (and generated from) a multimedia transport stream 730. In this example, transport stream 730 is transmitted on a separate IP multicast flow 740. This example also illustrates that transport stream 730 is transmitted by an entity other than point-of-interest producer 150. Other embodiments are contemplated in which point-of-interest producer 150 transmits both streams.

<figref idref="f0012">FIG. 7B</figref> is a block diagram of another embodiment of point-of-interest producer 150 and point-of-interest consumer 160, in which IP multicast is used to deliver the primary media stream and IP unicast is used to deliver the points-of-interest stream. Point-of-interest producer 150 produces one points-of-interest stream 710, and transmits this points-of-interest stream 710 to several instances of point-of-interest consumer 160, using a separate IP unicast flow 750A-C for delivery to each point-of-interest consumer 160. A single IP multicast flow 720 is used to transmit the multimedia transport stream 730 that is associated with the points-of-interest stream 710.

Several other flow variations (not shown) are contemplated. In one, each receiver receives the primary media stream and the points-of-interest stream on a single IP unicast address. The two streams are differentiated by UDP port number. In another variation, different IP multicast addresses are used to deliver points-of-interest stream 710 to different receivers. In yet another variation, the points-of-interest stream is delivered to multiple receivers using the same multicast IP address, but different UDP ports. In still another variation, the points-of-interest stream is delivered to different receivers on a single multicast IP address and a single UDP port, but distinguished by different PIDs.

<figref idref="f0013">FIG. 8</figref> is a block diagram showing selected components of a point-of-interest consumer 160 or producer 150 which implements at least one of the systems and methods disclosed herein. Consumer 160 or producer 150 comprises: a network interface 810; a peripheral I/O interface 820; a processor 830; and memory 840. These components are coupled by a bus 850.

Memory 840 contains instructions that are executed by processor 830 to control operations of consumer 160 or producer 150. Peripheral I/O interface 820 provides input and output signals, for example, user inputs from a remote control or front panel buttons or a keyboard, and outputs such as LEDs or LCD on the front panel. Network interface 810 transmits/receives points-of-interest stream 240, media stream 210, and/or combined stream 270 (depending on the configuration). In some embodiments, network interface 810 is for a local area network (LAN) or a wide area network (WAN) such as the Internet. In other embodiments, this interface is for a radio frequency (RF) network, and so may include a tuner/demodulator (not shown) which processes the digital signals received over the RF network.

Omitted from <figref idref="f0013">FIG. 8</figref> are a number of conventional components, known to those skilled in the art, that are unnecessary to explain the operation of the systems and methods disclosed herein. A person of ordinary skill in the art should understand that software components referred to herein includes executable code that is packaged, for example, as a standalone executable file, a library, a shared library, a loadable module, a driver, or an assembly, as well as interpreted code that is packaged, for example, as a class.

Any process descriptions or blocks in flowcharts should be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps in the process. As would be understood by those of ordinary skill in the art of the software development, alternate implementations are also included within the scope of the disclosure. In these alternate implementations, functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved.

The systems and methods disclosed herein can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device. Such instruction execution systems include any computer-based system, processor-containing system, or other system that can fetch and execute the instructions from the instruction execution system. In the context of this disclosure, a "computer-readable medium" can be any mechanism that can contain, store, communicate, propagate, or transport the program for use by, or in connection with, the instruction execution system. The computer readable medium can be, for example but not limited to, a system or propagation medium that is based on electronic, magnetic, optical, electromagnetic, infrared, or semiconductor technology.

Specific examples of a computer-readable medium using electronic technology may include (but are not limited to) the following: an electrical connection (electronic) having one or more wires; a random access memory (RAM); a read-only memory (ROM); an erasable programmable read-only memory (EPROM or Flash memory). A specific example using magnetic technology includes (but is not limited to) a portable computer diskette. Specific examples using optical technology include (but are not limited to) an optical fiber and a portable compact disk read-only memory (CD-ROM).

The foregoing description has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the disclosure to the precise forms disclosed. Obvious modifications or variations are possible in light of the above teachings. The implementations discussed, however, were chosen and described to illustrate the principles of the disclosure and its practical application to thereby enable a person of ordinary skill in the art to utilize the disclosure in various implementations and with various modifications as are suited to the particular use contemplated. All such modifications and variation are within the scope of the disclosure as determined by the appended claims when interpreted in accordance with the breadth to which they are fairly and legally entitled.