search

worldveil / dejavu

Audio fingerprinting and recognition in Python
MIT License
6.34k stars 1.43k forks source link

Which field in the results indicates the best match with the original song #282

Open 2954456878 opened 1 year ago

2954456878 commented 1 year ago

Which field in the results indicates the best match with the original song

2954456878 commented 1 year ago

{'song_id': 1, 'song_name': b'close', 'input_total_hashes': 214, 'fingerprinted_hashes_in_db': 130, 'hashes_matched_in_input': 17, 'input_confidence': 0.08, 'fingerprinted_confidence': 0.13, 'offset': -17, 'offset_seconds': -0.78948, 'file_sha1': b'FF3DD189441C50496EFCBB93498C03967F6DABB8'}

2954456878 commented 1 year ago

If two results are matched, how do you know which result matches the original song best

alffykasum commented 1 year ago

Hello @2954456878 In the results you provided, the field that indicates the best match with the original song is hashes_matched_in_input. This field tells you the number of hashes in the input audio file that matched with hashes in the database. A higher number of matched hashes generally indicates a better match with the original song.