Scalable real-time video compositing systems and methods
In some embodiments, a server system composites in real-time, in response to a user video search query, a standard-compliant (e.g. MPEG-4/H.264) SD or HD video stream encoding a rectangular (x-y) composite video preview panel array (grid) of video search results. Each panel/tile in the rectangular panel array displays a temporal section (e.g. the first 90 seconds, looped-back) of a video identified in response to the user query. Generating the composite video panel array in real-time is achieved by compositing the component video panels in the compressed domain, after each panel undergoes off-line a compressed-domain pre-compositing preparation process that facilitates dynamic compositing of the panels into a single video stream. The pre-compositing preparation includes transcoding to a format having a down-scaled common resolution, common GOP structure, and one-slice-per-row slice structure.
1. A computer system comprising at least one memory connected to at least one processor configured to:
in response to a user video search query, composite a plurality of transcoded video streams in the transform domain into a dynamically-generated composite video stream encoding a preview video panel array responsive to the user video search query, wherein:
the plurality of transcoded video streams are generated by transcoding a corresponding plurality of independent input video streams, wherein at least one of the input independent video streams includes slices encoding more than one row of video blocks, wherein transcoding the plurality of independent input video streams comprises
downscaling each of the independent video streams,
enforcing a common group-of-picture (GOP) structure across the plurality of transcoded video streams, wherein a GOP structure of the composite video stream coincides with the common GOP structure of the plurality of transcoded video streams, and
enforcing a one-slice-per-row slice structure for each picture of the plurality of transcoded video streams; and
transmit the composite video stream for generating a display of the preview video panel array.
2. The computer system of claim 1 , wherein the at least one processor is further configured to encapsulate the composite video stream in a container prior to transmitting the composite video stream for generating the display of the preview video panel array.
3. The computer system of claim 1 , wherein the at least one processor is further configured to identify a display hardware configuration used to generate the display of the preview video panel array, and to select an array size of the preview video panel array according to the display hardware configuration.
4. The computer system of claim 3 , wherein the at least one processor is configured to select the array size of the preview video panel array according to a screen diagonal size.
5. The computer system of claim 1 , wherein the at least one processor is further configured to identify a display hardware configuration used to generate the display of the preview video panel array, and to select, according to the display hardware configuration, a pre-constructed video library from a plurality of video libraries, the pre-constructed video library comprising the plurality of input video streams, for retrieving the input video streams from the pre-constructed video library.
6. The computer system of claim 1 , wherein the at least one processor is further configured to transmit a plurality of audio streams responsive to the user search query, each audio stream corresponding to one of the panels of the preview video panel array.
7. The computer system of claim 1 , wherein the at least one processor is further configured to selectively transmit an audio stream corresponding to a selected panel of the preview video panel array in response to an indicator of a user selection from the preview video panel array of the selected panel.
8. The computer system of claim 1 , wherein the preview video panel array shows a loopback display of a predetermined temporal section of each of the input video streams.
9. The computer system of claim 1 , wherein the composite video stream is encoded in an H.264 format.
10. A method comprising employing a computer system comprising at least one memory connected to at least one processor to:
in response to a user video search query, composite a plurality of transcoded video streams in the transform domain into a dynamically-generated composite video stream encoding a preview video panel array responsive to the user video search query, wherein:
the plurality of transcoded video streams are generated by transcoding a corresponding plurality of independent input video streams, wherein at least one of the input independent video streams includes slices encoding more than one row of video blocks, wherein transcoding the plurality of independent input video streams comprises
downscaling each of the independent video streams,
enforcing a common group-of-picture (GOP) structure across the plurality of transcoded video streams, wherein a GOP structure of the composite video stream coincides with the common GOP structure of the plurality of transcoded video streams, and
enforcing a one-slice-per-row slice structure for each picture of the plurality of transcoded video streams; and
transmit the composite video stream for generating a display of the preview video panel array.
11. A computer system comprising:
means for, in response to a user video search query, compositing a plurality of transcoded video streams in the transform domain into a dynamically-generated composite video stream encoding a preview video panel array responsive to the user video search query, wherein:
the plurality of transcoded video streams are generated by transcoding a corresponding plurality of independent input video streams, wherein at least one of the input independent video streams includes slices encoding more than one row of video blocks, wherein transcoding the plurality of independent input video streams comprises
downscaling each of the independent video streams,
enforcing a common group-of-picture (GOP) structure across the plurality of transcoded video streams, wherein a GOP structure of the composite video stream coincides with the common GOP structure of the plurality of transcoded video streams, and
enforcing a one-slice-per-row slice structure for each picture of the plurality of transcoded video streams; and
means for transmitting the composite video stream for generating a display of the preview video panel array.