Abstract: In one embodiment an apparatus for splicing an ad into a data stream includes an ad segmenter, an ad variant generator coupled to the ad segmenter, and an ad variant selector coupled to the ad variant generator. The ad segmenter designates end portions, including beginning and ending portions, of the ad and designates an intermediate portion between the end portions. The ad variant generator generates a plurality of variants of at least one end portion. Each variant corresponds to a unique bit-rate. The ad variant selector selects at least one variant for splicing into the data stream based on a buffer level of a data stream buffer.

Abstract: In one embodiment, a network device analyzes an encoded stream to identify a timing value included in a packet before a splice point. Next, the device determines a difference between identified timing value and a timing value included in a packet of the encoded segment to be spliced. In conjunction with splicing, the network device formats timing values of packets according to the determined difference, which improves play out of the spliced stream.

Abstract: A technique for providing program restart functionality in a digital video network involves receiving requests from different end-user devices to restart a program and servicing the program restart requests from the different end-user devices via a single broadcast stream. In an embodiment, program restart requests that are received within an established time interval, referred to as an aggregation interval, are serviced by the same broadcast stream. Additionally, during the aggregation interval and before the program restart begins, personalized content can be provided to the end-user devices via, for example, unicasting. Once the aggregation interval is over, all of the unicast streams are switched over to the same broadcast stream and the program broadcast is restarted. In an embodiment, a switch buffer is provided just before the restarted broadcast stream in order to ensure a smooth transition from each of the multiple unicast streams to the restarted broadcast stream.

Abstract: Black video segments in a video playlist may be minimized. Black detection may be utilized by a network device to identify true starting point and a true ending point of a video asset which includes black video segments in the playlist. Locations associated with the true starting and ending points may be recorded in metadata associated with the video asset. New black video segments may be added at the beginning and at the end of the playlist. Playback of the playlist may then be initiated to play the video asset only between the recorded location associated with the true starting point and the recorded location associated with the true ending point at the network device thereby minimizing the black video segments in the video asset.

Abstract: Techniques are provided to perform seamless or non-seamless ad insertion using a low-complexity process. Delay is introduced into a transport stream to prevent decoder buffer overflow. Delay buildup in the transport stream is reduced by removing pre-identified redundant or near redundant frames. Pre-identification of frames for removal reduces the processing required at the splice points, while maintaining a minimum overall delay in the transport stream.

Abstract: Switching a client from unicasting back to multicasting involves simultaneously providing the digital video content to the client via unicasting and multicasting until the client has buffered duplicate frames (i.e., frames that contain the same digital video content). Once the client has buffered duplicate frames, the client can transition from playing out of a buffer that holds frames received via unicasting to playing out of a buffer that holds frames received via multicasting without skipping a frame. Once the transition back to multicasting is complete, unicasting is terminated and any frames remaining in the unicast buffer are flushed.

Abstract: In a digital video network that is capable of distributing digital video content to a client via multicasting and unicasting, servicing a channel change request from a client involves switching from providing the digital video content to the client via multicasting to providing the digital video content to the client via unicasting and continuing to provide digital video content to the client via unicasting until a pre-established condition is met. Continuing to provide digital video content to the client via unicasting until a pre-established condition is met allows the network to opportunistically switch the client from unicasting back to multicasting.

Abstract: Black video segments in a video playlist may be minimized. Black detection may be utilized by a network device to identify true starting point and a true ending point of a video asset which includes black video segments in the playlist. Locations associated with the true starting and ending points may be recorded in metadata associated with the video asset. New black video segments may be added at the beginning and at the end of the playlist. Playback of the playlist may then be initiated to play the video asset only between the recorded location associated with the true starting point and the recorded location associated with the true ending point at the network device thereby minimizing the black video segments in the video asset.

Abstract: Switching a client from unicasting back to multicasting involves accumulating enough digital video content at the client to bridge the time it takes to transition from receiving the digital video content via unicasting to receiving the digital video content via multicasting. Digital video content is accumulated at the client by temporarily increasing the stream rate of the unicast stream above the playout rate. While the digital video content is being streamed at the increased rate, the amount of digital video content stored in the client's stream buffer grows. The stream rate is held at the increased rate until the client accumulates enough frames in its buffer to be able to bridge the time it takes to transition from receiving the digital video content via unicasting to receiving the digital video content via multicasting.

Abstract: Techniques are provided to perform seamless or non-seamless ad insertion using a low-complexity process. Delay is introduced into a transport stream to prevent decoder buffer overflow. Delay buildup in the transport stream is reduced by removing pre-identified redundant or near redundant frames. Pre-identification of frames for removal reduces the processing required at the splice points, while maintaining a minimum overall delay in the transport stream.

Abstract: A technique for providing program restart functionality in a digital video network involves receiving requests from different end-user devices to restart a program and servicing the program restart requests from the different end-user devices via a single broadcast stream. In an embodiment, program restart requests that are received within an established time interval, referred to as an aggregation interval, are serviced by the same broadcast stream. Additionally, during the aggregation interval and before the program restart begins, personalized content can be provided to the end-user devices via, for example, unicasting. Once the aggregation interval is over, all of the unicast streams are switched over to the same broadcast stream and the program broadcast is restarted. In an embodiment, a switch buffer is provided just before the restarted broadcast stream in order to ensure a smooth transition from each of the multiple unicast streams to the restarted broadcast stream.

Abstract: In one embodiment an apparatus for splicing an ad into a data stream includes an ad segmenter, an ad variant generator coupled to the ad segmenter, and an ad variant selector coupled to the ad variant generator. The ad segmenter designates end portions, including beginning and ending portions, of the ad and designates an intermediate portion between the end portions. The ad variant generator generates a plurality of variants of at least one end portion. Each variant corresponds to a unique bit-rate. The ad variant selector selects at least one variant for splicing into the data stream based on a buffer level of a data stream buffer.

Abstract: In one embodiment, a network device analyzes an encoded stream to identify a timing value included in a packet before a splice point. Next, the device determines a difference between identified timing value and a timing value included in a packet of the encoded segment to be spliced. In conjunction with splicing, the network device formats timing values of packets according to the determined difference, which improves play out of the spliced stream.

Abstract: In one embodiment, a network device analyzes an encoded video stream using black detection before splicing video into the encoded video stream. When black data is detected, the network device determines whether insertion of a black intra frame with the spliced video will improve viewing upon decoding. The network device may also determine whether a splicing boundary indication included in the encoded video stream is accurate based on an observed transition between the detected black data and other data, and compensate the video splice accordingly.

Abstract: In a digital video network that is capable of distributing digital video content to a client via multicasting and unicasting, servicing a channel change request from a client involves switching from providing the digital video content to the client via multicasting to providing the digital video content to the client via unicasting and continuing to provide digital video content to the client via unicasting until a pre-established condition is met. Continuing to provide digital video content to the client via unicasting until a pre-established condition is met allows the network to opportunistically switch the client from unicasting back to multicasting.

Abstract: A technique for supporting trick modes in a streaming digital video environment involves receiving a source stream of digital video content that is encoded for a normal playout rate and generating a root trick mode stream from the source stream, wherein the root trick mode stream is generated for a playout rate that is faster than the normal playout rate. The root trick mode stream is then used to generate subsequent trick mode streams that have faster playout rates than the root trick mode stream. Using a root trick mode stream to generate subsequent trick mode streams requires only one trick mode stream to be generated at ingest and therefore reduces ingest processing requirements and eliminates the need to store a different trick mode stream for every possible trick mode playout rate.

Abstract: Switching a client from unicasting back to multicasting involves simultaneously providing the digital video content to the client via unicasting and multicasting until the client has buffered duplicate frames (i.e., frames that contain the same digital video content). Once the client has buffered duplicate frames, the client can transition from playing out of a buffer that holds frames received via unicasting to playing out of a buffer that holds frames received via multicasting without skipping a frame. Once the transition back to multicasting is complete, unicasting is terminated and any frames remaining in the unicast buffer are flushed.

Abstract: Switching a client from unicasting back to multicasting involves accumulating enough digital video content at the client to bridge the time it takes to transition from receiving the digital video content via unicasting to receiving the digital video content via multicasting. Digital video content is accumulated at the client by temporarily increasing the stream rate of the unicast stream above the playout rate. While the digital video content is being streamed at the increased rate, the amount of digital video content stored in the client's stream buffer grows. The stream rate is held at the increased rate until the client accumulates enough frames in its buffer to be able to bridge the time it takes to transition from receiving the digital video content via unicasting to receiving the digital video content via multicasting.

Abstract: Disclosed is a method for copying a segment from an audiovisual file having a multiplicity of audio frames and a multiplicity of video frames. In a first step, a mark-in location in a video file is selected to correspond to a first video frame in the segment such that the first video frame has an associated start time. Next, a mark-out location in the video file is selected to correspond to a last video frame in the segment, and the last video frame having an associated end time. Once the mark-in video frame is selected, a first audio frame having a first audio frame start time that is at least as early as the first video frame start time is designated as an initial audio frame. A second audio frame having a second audio frame start time that is at least as late as the last video frame end time is designated as the last audio frame.

Abstract: Disclosed is a stitcher implemented to join segments of video glue generated during a first pass of an editing engine. Preferably, the generated glue segments are implemented at the beginning and at the end of a clipped video segment. In the second processing pass, any glue segments generated in the first pass may be stitched to any un-processed portion of the clipped segment. Once any glue segments and unprocessed portions are stitched in a time ordered sequence, the stitched segment is output to a application.