Granted Patent B2

US 7,962,933 · App. 11/694,546 · Granted Jun 14, 2011

Abstract

Claims (38)

1. A method, comprising

opening, and reading into memory, structural information from a first media stream;

opening, and reading into memory, structural information from a second media stream, wherein both the first and second media streams comprise one or more chunks and each of the chunks comprise one or more media samples;

identifying a nearest prior key frame in the second media stream corresponding to an insertion time;

identifying a chunk and a media sample corresponding to the identified key frame;

splitting the identified chunk into first and second split chunks so that the second split chunk begins with the media sample corresponding to the identified key frame, wherein this splitting occurs if the media sample corresponding to the identified key frame is not at the beginning of the identified chunk;

adjusting the structural information for the first and second media streams to effect insertion of the first media stream into the second media stream before the second chunk, wherein the structural information for each media stream comprises a sample-to-chunk table and a chunk-to-offset table, and wherein adjusting the structural information comprises modifying the sample-to-chunk table and the chunk-to-offset table of the second media stream to include entries for the one or more chunks and media samples of the first media stream and the first and second split chunks of the first media stream; and

writing onto an output media stream the adjusted structural information for the first and second media streams.

2. A method as in claim 1 , further comprising the step of piping, onto the output media stream, media data for the first and second media streams.

3. A method as in claim 1 , wherein the structural information for each media stream comprises a header for the media stream and one or more tracks, wherein each track further comprises a header, one or more edit lists, and one or more media, wherein the media further comprises a header and one or more sample tables specifying the locations and durations of sample data.

4. A method as in claim 1 , wherein the first media stream and the second media stream are pre-normalized to avoid transcoding.

5. A method as in claim 1 , wherein the steps are performed by an insertion server.

6. A method as in claim 1 , wherein there are more than two input media streams.

7. A method as in claim 1 , wherein the first media stream comprises an advertisement and the second media stream comprises content.

8. A method as in claim 1 , wherein the piping step uses Java file channels.

9. A method as in claim 1 , wherein the first, second, and output media streams are in container file format.

10. A method as in claim 1 , wherein the first, second, and output media streams are in MP4 file format.

11. An apparatus, comprising

a memory,

a network interface;

one or more processors; and

logic encoded in one or more persistent storage media for execution and when executed operable to cause the one or more processors to:

open, and read into the memory, structural information from a first media stream;

open, and read into the memory, structural information from a second media stream, wherein both the first and second media streams comprise one or more chunks and each of the chunks comprise one or more media samples;

identify a nearest prior key frame in the second media stream corresponding to an insertion time;

identify a chunk and a media sample corresponding to the identified key frame;

split the identified chunk into first and second split chunks so that the second split chunk begins with the media sample corresponding to the identified key frame, wherein this splitting occurs if the media sample corresponding to the identified key frame is not at the beginning of the identified chunk;

adjust the structural information for the first and second media streams to effect insertion of the first media stream into the second media stream before the second chunk, wherein the structural information for each media stream comprises a sample-to-chunk table and a chunk-to-offset table, and wherein adjusting the structural information comprises modifying the sample-to-chunk table and the chunk-to-offset table of the second media stream to include entries for the one or more chunks and media samples of the first media stream and the first and second split chunks of the first media stream; and

write onto an output media stream the adjusted structural information for the first and second media streams.

12. An apparatus as in claim 11 , further comprising encoded logic to pipe onto the output media stream, media data for the first and second media streams.

13. An apparatus as in claim 11 , wherein the structural information for each media stream comprises a header for the media stream and one or more tracks, wherein each track further comprises a header, one or more edit lists, and one or more media, wherein the media further comprises a header and one or more sample tables specifying the locations and durations of sample data.

14. An apparatus as in claim 11 , wherein the first media stream and the second media stream are pre-normalized to avoid transcoding.

15. An apparatus as in claim 11 , wherein the encoded logic is executed by an insertion server.

16. An apparatus as in claim 11 , wherein there are more than two input media streams.

17. An apparatus as in claim 11 , wherein the first media stream comprises an advertisement and the second media stream comprises content.

18. An apparatus as in claim 11 , wherein the piping operation uses Java file channels.

19. An apparatus as in claim 11 , wherein the first, second, and output media streams are in container file format.

20. An apparatus as in claim 11 , wherein the first, second, and output media streams are in MP4 file format.

Assignments (17)