Method and apparatus for embedding watermarks

• US 20070300066A1
• Filed: 12/09/2005
• Published: 12/27/2007
• Est. Priority Date: 06/13/2003
• Status: Active Grant

• Current Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Original Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Inventors: Srinivasan, Venugopal
• Granted Patent: US 7,460,684 B2
• Time in Patent Office: Days
• US Class Current: 713/176
• CPC Class Codes: G06T 1/0021 Image watermarking
• CPC Class Codes: G10L 19/018 Audio watermarking, i.e. embedding inaudible data in the audio signal
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 25/27 characterised by the analysis technique

Abstract

Methods and apparatus for embedding a watermark are disclosed. In an example method, one or more frames associated with a compressed digital data stream (240) is identified. Each of the one or more frames is unpacked to determine a plurality of transform coefficient sets (320). The plurality of transform coefficient sets (320) is modified to embed the watermark (230).

Claims

1. A method of embedding a watermark comprising:

identifying one or more frames associated with a compressed digital data stream;

unpacking each of the one or more frames to identify a plurality of transform coefficient sets; and

modifying the plurality of transform coefficient sets to embed the watermark.

2. A method as defined in claim 1, wherein modifying the plurality of transform coefficient sets comprises substituting a watermarked transform coefficient set for at least one of a plurality of transform coefficient sets.

3. A method as defined in claim 1, wherein modifying the plurality of transform coefficient sets comprises:

determining a mantissa code associated with a watermarked transform coefficient of one of a plurality of watermarked transform coefficient sets; and

substituting the mantissa code associated with the watermarked transform coefficient for a mantissa code associated with a corresponding transform coefficient of one of the plurality of transform coefficient sets.

4. A method as defined in claim 3, wherein determining the mantissa code associated with the watermarked transform coefficient of the one of the plurality of watermarked transform coefficient sets comprises:

selecting a code signal frequency to encode to the plurality of transformed coefficient sets based on data to be embedded;

determining a masking energy associated with the code signal frequency to be encoded to the plurality of transformed coefficient sets;

selecting a magnitude for the watermarked transform coefficient based on the masking energy; and

determining the mantissa code associated with the watermarked transform coefficient based on the magnitude.

5. A method as defined in claim 4, wherein the code signal frequency comprises a frequency corresponding to one of a plurality of high resolution frequency domain representation.

6. A method as defined in claim 4, wherein the code signal comprises one or more sinusoidal components, and wherein each sinusoidal component having a frequency based on a desired code.

7. A method as defined in claim 1, wherein modifying the plurality of transform coefficient sets comprises:

generating a plurality of time-domain audio blocks based on the plurality of transform coefficient sets;

generating a plurality of reconstructed audio blocks based on the plurality of time-domain audio blocks; and

generating a plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks.

8. A method as defined in claim 7, wherein generating the plurality of time-domain audio blocks comprises generating a first time-domain audio block and a second time-domain audio block associated with an audio block.

9. A method as defined in claim 7, wherein generating the plurality of reconstructed audio blocks based on the plurality of time-domain audio blocks comprises generating a reconstructed time-domain audio block corresponding to an audio block based on a first time-domain audio block and a second time-domain audio block.

10. A method as defined in claim 9, wherein generating the first reconstructed time-domain audio block corresponding to the first audio block based on the first time-domain audio block and the second time-domain audio block comprises adding the first and second time-domain audio blocks.

11. A method as defined in claim 7, wherein generating the plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks comprises:

generating a modifiable time-domain audio block based on the plurality of reconstructed audio blocks; and

generating a first watermarked audio block and a second watermarked audio block based on the modifiable time-domain audio block and the watermark.

12. A method as defined in claim 11, wherein generating the modifiable time-domain audio block based on the plurality of reconstructed audio blocks comprises concatenating a first reconstructed audio block and a second reconstructed audio block to form a 512-sample audio block.

13. A method as defined in claim 1, wherein modifying the plurality of transform coefficient sets comprises modifying the plurality of transform coefficient sets based on a plurality of watermarked transform coefficient sets.

14. A method as defined in claim 1, wherein modifying the plurality of transform coefficient sets comprises generating a plurality of watermarked transformed coefficient sets based on a first watermarked audio block and a second watermarked audio block.

15. A method as defined in claim 14, wherein generating the plurality of watermarked coefficient sets based on the first and second watermarked audio blocks comprises determining a mantissa code associated with a watermarked transform coefficient of each of the plurality watermarked transform coefficient sets based on compression information associated with the compressed digital data stream.

16. A method as defined in claim 1, wherein each of the plurality of transform coefficient sets comprises one or more modified discrete cosine transform coefficients.

17. A method as defined in claim 1, wherein the compressed digital data stream is compressed in accordance with an audio compression standard.

18. A method as defined in claim 1, wherein identifying the one or more frames associated with the compressed digital data stream comprises identifying audio blocks associated with at least one of a plurality of audio channels.

19. A method as defined in claim 1, wherein unpacking each of the one or more frames to identify the plurality of transform coefficient sets comprises identifying compression information associated with the compressed digital data stream.

20. A method as defined in claim 1 further comprising repacking the one or more frames based on a plurality of watermarked transform coefficient sets.

21. A method as defined in claim 1, wherein the watermark is associated with one of a media source and a media program.

22. An apparatus of embedding a watermark comprising:

an identifier to identify one or more frames associated with a compressed digital data stream;

an unpacker to unpack each of the one or more frames to identifying a plurality of transform coefficient sets; and

a modifier to modify the plurality of transform coefficient sets to embed the watermark.

23. An apparatus as defined in claim 22, wherein the modifier determines a mantissa code associated with a watermarked transform coefficient of one of a plurality of watermarked transform coefficient sets, and substitutes the mantissa code associated with the watermarked transform coefficient for a mantissa code associated with a corresponding transform coefficient of one of the plurality of transform coefficient sets.

24. An apparatus as defined in claim 23, wherein the modifier selects a code signal frequency to encode to the plurality of transformed coefficient sets based on data to be embedded, determines a masking energy associated with the code signal frequency to be encoded to the plurality of transformed coefficient sets, selects a magnitude for the watermarked transform coefficient based on the masking energy, and determines the mantissa code associated with the watermarked transform coefficient based on the magnitude.

25. An apparatus as defined in claim 24, wherein the code signal frequency comprises a frequency corresponding to one of a plurality of high resolution frequency domain representation.

26. An apparatus as defined in claim 24, wherein the code signal comprises one or more sinusoidal components, and wherein each sinusoidal component having a frequency based on a desired code.

27. An apparatus as defined in claim 22, wherein the modifier generates a plurality of time-domain audio blocks, generates a plurality of reconstructed audio blocks based on the plurality of time-domain audio blocks, and generates a plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks.

28. An apparatus as defined in claim 27, wherein the modifier generates a prime audio block and a double-prime audio block associated with an audio block.

29. An apparatus as defined in claim 27, wherein the modifier generates a reconstructed time-domain audio block corresponding to an audio block based on a first time-domain audio block and a second time-domain audio block.

30. An apparatus as defined in claim 29, wherein the modifier adds the first and second time-domain audio blocks.

31. An apparatus as defined in claim 27, wherein the modifier generates a modifiable time-domain audio block based on the plurality of reconstructed audio blocks, and generates a first watermarked audio block and a second watermarked audio block based on the modifiable time-domain audio block and the watermark.

32. An apparatus as defined in claim 31, wherein the modifier concatenates a first reconstructed audio block and a second reconstructed audio block to form a 512-sample audio block.

33. An apparatus as defined in claim 32, wherein the modifier determines a mantissa code associated with each watermarked MDCT coefficients of the one of the plurality watermarked coefficient sets based on compression information of the compressed digital data stream.

34. An apparatus as defined in claim 22, wherein the modifier modifies the plurality of transform coefficient sets based on a plurality of watermarked transform coefficient sets.

35. An apparatus as defined in claim 34, wherein the modifier generates one of the plurality of watermarked transform coefficient sets based on a first watermarked audio block and a second watermarked audio block.

36. An apparatus as defined in claim 34, wherein the modifier substitutes one of the plurality of watermarked transform coefficient sets for one of the plurality of transform coefficient sets.

37. An apparatus as defined in claim 22, wherein each of the plurality of transform coefficient sets comprises one or more modified discrete cosine transform coefficients.

38. An apparatus as defined in claim 22, wherein the compressed digital data stream is compressed in accordance with an audio compression standard.

39. An apparatus as defined in claim 22, wherein the identifying unit identifies audio blocks associated with a plurality of audio channels.

40. An apparatus as defined in claim 22, wherein the unpacking unit identifies compression information associated with the compressed digital data stream.

41. An apparatus as defined in claim 22, wherein the watermark comprises a watermark associated with one of a media source and a media program.

42. An apparatus as defined in claim 22 further comprising a frame repacker to repack the one or more frames based on a plurality of watermarked transform coefficient sets.

43. A machine accessible medium having instructions, which when executed, cause a machine to:

identify one or more frames associated with a compressed digital data stream;

unpack each of the one or more frames to identify a plurality of transform coefficient sets; and

modify the plurality of transform coefficient sets to embed a watermark.

44. A machine accessible medium as defined in claim 43, wherein the instructions, which when executed, cause the machine to modify the plurality of transform coefficient sets by substituting a watermarked transform coefficient set for at least one of the plurality of transform coefficient sets.

45. A machine accessible medium as defined in claim 43, wherein the instructions, which when executed, cause the machine to modify the plurality of transform coefficient sets by:

determining a mantissa code associated with a watermarked transform coefficient of one of a plurality of watermarked transform coefficient sets; and

substituting the mantissa code associated with the watermarked transform coefficient for a mantissa code associated with a corresponding transform coefficient of one of the plurality of transform coefficient sets.

46. A machine accessible medium as defined in claim 45, wherein the instructions, which when executed, cause the machine to determine the mantissa code associated with the watermarked transform coefficient of the one of the plurality of watermarked transform coefficient sets by:

selecting a code signal frequency to encode to the plurality of transformed coefficient sets based on data to be embedded;

determining a masking energy associated with the code signal frequency to be encoded to the plurality of transformed coefficient sets;

selecting a magnitude for the watermarked transform coefficient based on the masking energy; and

determining the mantissa code associated with the watermarked transform coefficient based on the magnitude.

47. A machine accessible medium as defined in claim 46, wherein the code signal frequency comprises a frequency corresponding to one of a plurality of high resolution frequency domain representation.

48. A machine accessible medium as defined in claim 46, wherein the code signal comprises one or more sinusoidal components, and wherein each sinusoidal component having a frequency based on a desired code.

49. A machine accessible medium as defined in claim 43, wherein the instructions, which when executed, cause the machine to modify the plurality of transform coefficient sets by:

generating a plurality of time-domain audio blocks based on the plurality of transform coefficient sets;

generating a plurality of reconstructed audio blocks based on the plurality of time-domain audio blocks; and

generating a plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks.

50. A machine accessible medium as defined in claim 49, wherein the instructions, which when executed, cause the machine to generate the plurality of time-domain audio blocks by generating a first time-domain audio block and a second time-domain audio block associated with an audio block.

51. A machine accessible medium as defined in claim 49, wherein the instructions, which when executed, cause the machine to generate a plurality of reconstructed audio blocks based on the plurality of time-domain audio blocks by generating a reconstructed time-domain audio block corresponding to an audio block based on a first time-domain audio block and a second time-domain audio block.

52. A machine accessible medium as defined in claim 51, wherein the instructions, which when executed, cause the machine to generate the first reconstructed time-domain audio block corresponding to the first audio block based on the first time-domain audio block and the second time-domain audio block comprises adding the first and second time-domain audio blocks.

53. A machine accessible medium as defined in claim 49, wherein the instructions, which when executed, cause the machine to generate a plurality of watermarked audio blocks based on the plurality of reconstructed audio blocks by:

generating a modifiable time-domain audio block based on the plurality of reconstructed audio blocks; and

generating a first watermarked audio block and a second watermarked audio block based on the modifiable time-domain audio block and the watermark.

54. A machine accessible medium as defined in claim 53, wherein the instructions, which when executed, cause the machine the generate the modifiable time-domain audio block based on the plurality of reconstructed audio blocks comprises concatenating a first reconstructed audio block and a second reconstructed audio block to form a 512-sample audio block.

55. A machine accessible medium as defined in claim 49, wherein the instructions, which when executed, cause the machine to generate the plurality of watermarked transformed coefficient sets by generating one of a plurality of watermarked transform coefficient sets based on a first watermarked audio block and a second watermarked audio block.

56. A machine accessible medium as defined in claim 55, wherein the instructions, which when executed, cause the machine to generate the one of the plurality of watermarked coefficient sets based on the first and second watermarked audio blocks by determining a mantissa code associated with each watermarked transform coefficient of the one of the plurality watermarked coefficient sets based on compression information associated with the compressed digital data stream.

57. A machine accessible medium as defined in claim 43, wherein each of the plurality of transform coefficient sets comprises one or more modified discrete cosine transform coefficients.

58. A machine accessible medium as defined in claim 43, wherein the compressed digital data stream is compressed in accordance with an audio compression standard.

59. A machine accessible medium as defined in claim 43, wherein the instructions, which when executed, cause the machine to identify the one or more frames associated with the compressed digital data stream by identifying audio blocks associated with at least one of a plurality of audio channels.

60. A machine accessible medium as defined in claim 43, wherein the instructions cause the machine to unpack each of the one or more frames to identify the plurality of transform coefficient sets by identifying compression information associated with the compressed digital data stream.

61. A machine accessible medium as defined in claim 43, wherein the instructions, which when executed, cause a machine to repack the one or more frames based on a plurality of watermarked transform coefficient sets.

62. A machine accessible medium as defined in claim 43, wherein the watermark is associated with one of a media source and a media program.

63. A machine accessible medium as defined in claim 43 is one of a programmable gate array, application specific integrated circuit, erasable programmable read only memory, read only memory, random access memory, magnetic media, and optical media.