| 主权项 |
1. A method comprising:
from a processor within a device, querying a content recognition database to determine program identification; obtaining through a network connection and caching locally in a memory of the device a local fingerprint database for a program associated with the program identification, the local fingerprint database including audio fingerprints for segments of the program and program times for associated audio fingerprints in the local fingerprint database; with a processor operating on an electronic audio signal sensed from ambient audio by a microphone of the device, extracting audio features from incoming segments of the electronic audio signal, and forming fingerprints for the incoming segments from audio features extracted from corresponding segments; with a processor, matching the fingerprints for the incoming segments with audio fingerprints for the program in the local fingerprint database in the memory of the device, and responsive to detecting matches between fingerprints for incoming segments and matching audio fingerprints in the local fingerprint database, obtaining from the local fingerprint database program times corresponding to the matching audio fingerprints; and with a processor, tracking the program times relative to a time reference in the device to keep track of where a user is within the program as the device is exposed to audio of the program. |
