Sound representation via winner-take-all coding of auditory spectra

1. A system, comprising:

a memory that stores computer executable components;

a processor that executes the following computer executable components stored in the memory:

an auditory transformation component that converts a set of sound frames from audio content into a set of spectral slices;

a spectral encoder component that encodes the set of spectral slices into winner-take-all codes with a winner-take-all hash function; and

an identification component that identifies which spectral dimension of a subset of spectral dimensions within at least one spectral slice has a highest spectral value according to the winner-take-all codes, that identifies subsets of the spectral dimensions of the at least one spectral slice based on a random function, and that identifies at least one reference audio content that is similar to the audio content based on the winner-take-all codes.

2. The system of claim 1 , further comprising a key component that concatenates the winner-take-all codes into index keys.

3. The system of claim 1 , comprising a histogram component that histograms the winner-take-all codes into histogrammed winner-take-all codes.

4. The system of claim 1 , wherein the winner-take-all codes comprise auditory data based on position of signal peaks.

5. The system of claim 1 , comprising an indexing component that indexes index keys based on the winner-take-all codes into a set of hash tables.

6. The system of claim 5 , comprising a matching component that identifies the at least one reference audio content that is similar to the audio content based in part on the hash tables.

7. The system of claim 1 , wherein the set of spectral slices include stabilized auditory images.

8. A method comprising:

employing a processor to execute computer executable instructions stored in a memory to perform the following acts:

converting sound frames into spectral slices from a set of audio content;

encoding the spectral slices into winner-take-all codes with a winnertake-all hash function;

identifying which spectral dimension of a subset of spectral dimensions within at least one spectral slice has a highest spectral value according to the winner-take-all codes;

identifying subsets of the spectral dimensions of the at least one spectral slices based on a random function; and

identifying at least one reference audio content that is similar to the audio content based on the winner-take-all codes.

9. The method of claim 8 , comprising:

concatenating the winner-take-all codes into index keys.

10. The method of claim 8 , further comprising:

processing the winner-take-all codes into histogrammed winner-take-all codes.

11. The method of claim 8 , wherein the winner-take-all codes comprise auditory data based on a position of signal peaks.

12. The method of claim 8 , comprising identifying the at least one reference audio content that is similar to the audio content according to hash tables based on the winner-take-all codes.

13. A non-transitory computer readable storage medium comprising computer-executable instructions, that in response to execution, cause a computing system to perform operations, comprising:

converting sound frames into spectral slices;

generating winner-take-all codes;

encoding the spectral slices with the winner-take-all codes to indicate a dimension within a group of a set of groups of spectral dimensions that has a highest spectral value;

identifying subsets of the spectral dimensions of the at least one spectral slices based on a random function; and

identifying at least one reference audio content that is similar to the audio content based on the winner-take-all codes.

14. The operations of claim 13 , comprising:

converting the winner-take-all codes into index keys.

15. The operations of claim 14 , further comprising:

processing the winner-take-all codes into histogrammed winner-take-all codes with the index keys.

16. The operations of claim 13 , comprising indexing a set of index keys based on the winner-take-all codes into a set of hash tables.