System and methods for offline audio recognition
View Patent ↗In one implementation, a method is described of retrying matching of an audio query against audio references. The method includes receiving a follow-up query that requests a retry at matching a previously submitted audio query. In some implementations, this follow-up query is received without any recognition hint that suggests how to retry matching. The follow-up query includes the audio query or a reference to the audio query to be used in the retry. The method further includes retrying matching the audio query using retry matching resources that include an expanded group of audio references, identifying at least one match and transmitting a report of the match. Optionally, the method includes storing data that correlates the follow-up query, the audio query or the reference to the audio query, and the match after retrying.
1. A computer-implemented method of retrying matching of an initial audio query against audio references, the method including:
receiving an initial audio query submitted by a user device, the initial audio query comprising a recorded audio sample or audio fingerprints derived from the recorded audio sample, the audio sample having been recorded by the user device from ambient sound, and the audio sample representing a portion of, and less than the entirety of, an audio recording;
attempting to match the initial audio query to an audio reference in a core set of audio references;
sending to the user device zero or more potential matches to the initial audio query;
receiving from the user device a follow-up query that requests retrying to match the initial audio query previously submitted by the user device, the follow-up query including the initial audio query or a reference to the initial audio query;
processing the follow-up query by retrying to match the initial audio query against a second set of audio references different from the core set of audio references;
identifying at least one match and transmitting a report of the match.
2. The computer-implemented method of claim 1 , wherein:
one or more potential matches are sent to the user device; and
the follow-up query received from the user device indicates that none of the one or more potential matches is a correct match.
3. The computer-implemented method of claim 1 , wherein the second set of audio references used to process the follow-up query is an updated core set of audio references; and
further comprising the follow-up query successfully identifying at least one match in the updated core set of audio references.
4. The computer-implemented method of claim 1 , further including, responsive at least in part to finding a match against a reference not in the core set of audio references, adding the matching audio reference to the core set of audio references.
5. The computer-implemented method of claim 1 , further including, increasing a popularity score for a matching audio reference not in the core set of audio references; and
responsive to determining that the popularity score exceeds a configured threshold, adding the matching audio reference to the core set of audio references.
6. The computer-implemented method of claim 1 , wherein receiving the initial audio query further comprises storing the initial audio query and transmitting to the user device a reference to the stored initial audio query.
7. The computer-implemented method of claim 1 , wherein the follow-up query includes a recognition hint that includes at least one of:
an artist name, song title, album name, or genre;
a set of audio references in which to search for a match; or
an indication that the recorded audio sample is not well known.
8. A non-transitory computer-readable storage medium storing instructions, which when executed by a processor, cause the processor to perform the operations of claim 1 .
9. An article of manufacture comprising:
a processor; and
the non-transitory computer-readable storage medium of claim 8 .
10. A computer-implemented method of retrying matching of an initial audio query against audio references, the method including:
transmitting, by a user device, an initial audio query that requests identification information for a recorded audio sample, the initial audio query comprising the recorded audio sample or audio fingerprints derived from the recorded audio sample, the audio sample having been recorded by the user device from ambient sound, and the audio sample representing a portion of, and less than the entirety of, an audio recording;
receiving a report that identifies zero or more audio references that match the initial audio query;
receiving from a user interface an indication that the received report does not include a correct match for the initial audio query; and
creating a follow-up query that includes the previously submitted initial audio query or a reference to the previously submitted initial audio query;
transmitting the follow-up query;
in response to transmitting the follow-up query, receiving a report that identifies one or more audio reference that correctly match the initial audio query.
11. The computer-implemented method of claim 10 , wherein the previously submitted initial audio query is a first initial audio query, and the follow-up query is a first follow-up query; and further comprising:
transmitting a second initial audio query; and
transmitting a second follow-up query that includes the second initial audio query before creating the first follow-up query.
12. The computer-implemented method of claim 11 , further comprising storing a plurality of initial audio queries on the user device, and
providing a user interface for selecting one of the stored plurality of initial audio queries for inclusion in a follow-up query.
13. The computer-implemented method of claim 11 , further comprising:
receiving from the user interface a selection of one of a plurality of previously submitted initial audio queries stored on the user device; and
creating and transmitting a follow-up query including the selection of the one previously submitted initial audio queries.
14. The computer-implemented method of claim 10 , wherein the user device waits synchronously for the response to the follow-up query.
15. The computer-implemented method of claim 10 , wherein the response to the follow-up query is received by way of an asynchronous notification.
16. The computer-implemented method of claim 10 , wherein the created follow-up query includes a recognition hint that includes at least one of:
an artist name, song title, album name, or musical genre;
a set of audio references in which to search for a match; or
an indication that the recorded audio sample is not well known.
17. A non-transitory computer-readable storage medium storing instructions, which when executed by a processor, cause the processor to perform the operations of claim 10 .
18. An article of manufacture comprising:
a processor; and
the non-transitory computer-readable storage medium of claim 17 .
19. A computer-implemented method of retrying matching of an initial audio query against audio references, the method including:
receiving an initial audio query submitted by a user device, the initial audio query comprising a recorded audio sample or audio fingerprints derived from the recorded audio sample, the audio sample having been recorded by the user device from ambient sound, and the audio sample representing a portion of, and less than the entirety of, an audio recording;
attempting to match the initial audio query to an audio reference in a core set of audio references;
finding no potential matches to the initial audio query;
creating a follow-up query that requests retrying to match the initial audio query for which no potential matches were found;
processing the follow-up query by retrying to match the initial audio query against a second set of audio references different from the core set of audio references; and
identifying at least one match and transmitting a report of the match.
20. The computer-implemented method of claim 19 , further comprising queuing the follow-up query for deferred processing.
21. The computer-implemented method of claim 20 , wherein processing the queued follow-up query is deferred to a time of reduced query load.
22. A non-transitory computer-readable storage medium storing instructions, which when executed by a processor, cause the processor to perform the operations of claim 19 .
23. An article of manufacture comprising:
a processor; and
the non-transitory computer-readable storage medium of claim 22 .