Methods and apparatus to perform audio watermark detection and extraction

• US 9,368,123 B2
• Filed: 10/16/2012
• Issued: 06/14/2016
• Est. Priority Date: 10/16/2012
• Status: Active Grant

• Current Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Original Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Inventors: Srinivasan, Venugopal, Topchy, Alexander
• Primary Examiner(s): Azad, Abul
• Application Number:
  US13/653,001
• Publication Number:
  US 20140105448A1
• Time in Patent Office: 1,337 Days
• Field of Search: 704/273, 704500-504
• US Class Current: 1/1
• CPC Class Codes: G10L 19/018 Audio watermarking, i.e. embedding inaudible data in the audio signal

Abstract

Methods and apparatus to perform audio watermark detection and extraction are disclosed. An example method includes sampling a media signal to generate samples, wherein the media signal includes an embedded message, determining a first symbol value for a first block of the samples, determining a second symbol value for a second block of the samples, and determining, using a processor, a resulting symbol value, representative of a part of the embedded message, based on the first symbol value and the second symbol value for the first block of samples and the second block of samples, wherein the first block and the second block partially overlap.

Claims

1. A method for determining information embedded in media signals, the method comprising:

sampling, using a processor, a media signal to generate samples, wherein the media signal includes an embedded message;

determining, using the processor, a first symbol value based on a first frequency spectrum determined for a first block of the samples;

determining, using the processor, a second symbol value based on a second frequency spectrum determined for a second block of the samples; and

determining, using the processor, a resulting symbol value, representative of a part of the embedded message, by voting based on the first symbol value for the first block of samples and the second symbol value for the second block of samples, wherein the first block and the second block partially overlap in time in the media signal.

2. The method as defined in claim 1, further including determining a third symbol value for a third block of the samples, the third block of samples being located a multiple of a length of an embedded message prior to the first block of samples, and wherein determining the resulting symbol value is also based on the third symbol value.

3. The method as defined in claim 2, further including determining a fourth symbol value for a fourth block of the samples, the fourth block of samples and the third block of samples partially overlap, and determining the resulting symbol value is also determined based on the third symbol value.

4. The method as defined in claim 1, further including determining a first plurality of symbol values for a first plurality of blocks of samples, each block in the first plurality of blocks of samples being located a multiple of a length of a message prior to the first block of samples, and determining the resulting symbol value is also determined based on the first plurality of symbol values.

5. The method as defined in claim 4, further including determining a second plurality of symbol values for a second plurality of blocks of samples, each member of the second plurality of blocks of samples partly overlapping a member of the first plurality of blocks of samples, and determining the resulting symbol value is also determined based on the second plurality of symbol values.

6. The method as defined in claim 1, further including determining a third symbol value from a third block of samples, and the resulting symbol value is also determined based on the third symbol value.

7. The method as defined in claim 6, wherein determining the resulting symbol value includes extracting a most occurring symbol value.

8. The method as defined in claim 7, wherein the most occurring symbol value is determined by voting.

9. The method as defined in claim 1, wherein the media signal is embedded with a plurality of messages, each message including a series of symbols.

10. The method as defined in claim 1, wherein the samples are stored in a buffer.

11. The method as defined in claim 10, wherein the buffer is a circular buffer.

12. The method as defined in claim 1, further including storing the first symbol value and the second symbol value in a tangible memory, wherein the processor reads the first symbol value and the second symbol value from the tangible memory when determining the resulting symbol value.

13. The method as defined in claim 12, wherein the tangible memory is a circular buffer.

14. The method as defined in claim 1, wherein the first symbol value and the second symbol value are determined by performing a spectral analysis on, respectively, the first block of samples and the second block of samples to determine the first symbol value and the second symbol value.

15. The method as defined in claim 14, wherein the spectral analysis is performed using a fast Fourier transform.

16. The method as defined in claim 1, wherein the media signal is an audio signal.

17. The method as defined in claim 16, wherein the embedded message is embedded as an audio watermark.

18. A system for identifying messages embedded within media signals, the system comprising:

a sampler to sample a media signal to generate samples, wherein the media signal includes an embedded message;

a first symbol value extractor to determine a first symbol value for a first block of the samples;

a second symbol value extractor to determine a second symbol value for a second block of the samples; and

a processor to determine a resulting symbol value, representative of a part of the embedded message, based on the first symbol value and the second symbol value for the first and second block of the samples, wherein the first block of the samples and the second block of the samples partially overlap.

19. The system as defined in claim 18, further including a third symbol value extractor to determine a third symbol value for a third block of the samples, wherein the third block of the samples is located a multiple of a length of an embedded message prior to the first block of the samples, wherein determining the resulting symbol value is also based on the third symbol value.

20. The system as defined in claim 19, further including a fourth symbol value extractor to determine a fourth symbol value for a fourth block of samples wherein the fourth block of samples and the third block of samples partially overlap, wherein determining the resulting symbol value is also determined based on the third symbol value.

21. The system as defined in claim 18, further including a first plurality of symbol value extractors to determine a first plurality of symbol values for a first plurality of blocks of samples, wherein each block in the first plurality of blocks of samples is located a multiple of a length of a message prior to the first block of samples, wherein determining the resulting symbol value is also determined based on the first plurality of symbol values.

22. A tangible computer readable storage medium comprising instructions, which, when executed, cause a machine to at least:

sample a media signal to generates samples, wherein the media signal includes an embedded message;

determine a first symbol value for a first block of the samples;

determine a second symbol value for a second block of the samples; and

determine a resulting symbol value, representative of a part of the embedded message, based on the first symbol value and the second symbol value for the first and second blocks of samples, wherein the first block and the second block are partially overlapped.

23. The tangible computer readable storage medium as defined in claim 22, wherein the instructions, when executed, further cause the machine to determine a third symbol value for a third block of the samples, wherein the third block of samples is located a multiple of a length of an embedded message prior to the first block of samples, wherein determining the resulting symbol value is also based on the third symbol value.