Audio duplicate detector
The present invention relates to a system and methodology to facilitate automatic management and pruning of audio files residing in a database. Audio fingerprinting is a powerful tool for identifying streaming or file-based audio, using a database of fingerprints. Duplicate detection identifies duplicate audio clips in a set, even if the clips differ in compression quality or duration. The present invention can be provided as a self-contained application that it does not require an external database of fingerprints. Also, a user interface provides various options for managing and pruning the audio files.
1. A system for managing audio information, comprising:
a fingerprinting component that maps portions of a plurality of audio files to corresponding fingerprints;
a detector that tags one or more of the audio files for potential removal from a data storage device based in part upon a distance between the fingerprints; and
a database for storing veto fingerprints that are employed to identify noisy audio files.
2. The system of claim 1 , the detector tags the audio files based upon the distance between fingerprints being below a predetermined threshold.
3. The system of claim 1 , the fingerprinting component further producing a plurality of fingerprints for a file, the plurality of fingerprints corresponding to a time window of audio in the file, and wherein the detector tags the audio files based upon a lowest distance between the plurality of fingerprints and one or more stored fingerprints for each file.
4. The system of claim 3 , the fingerprinting component is disposed to accept a time offset into the audio file and a duration of a time window in the files.
5. The system of claim 4 , the detector computes and compares all fingerprints in a time window, in order to find a best matching location in a file that has already been processed.
6. The system of claim 5 , the detector is employed for determining an identity of an audio file.
7. The system of claim 6 , the identity is composed of metadata associated with an audio file.
8. The system of claim 1 , the fingerprint component computes fingerprints that are generated from more than one second of audio, and that consist of about 64 floating point numbers.
9. The system of claim 1 , the detector utilizes at least two internal databases referred to as DB 1 and DB 2 , in DB 1 , a record comprises a fingerprint and associated numerical quantities including a normalization factor, in DB 2 , a record includes at least four objects: a filename, an associated index referred to as an ID index, an ‘offset’ parameter and a ‘distance’ parameter.
10. The system of claim 1 , further comprising a database that is employed to output a list of duplicate or defective audio files to a user interface.
11. The system of claim 10 , the detector logs error conditions while processing the audio files and outputs a list of files associated with the error conditions to the user interface.
12. A computer readable medium having computer readable instructions stored thereon for implementing the fingerprinting component and the detector of claim 1 .