dejavu
dejavu copied to clipboard
Audio fingerprinting and recognition in Python
This will add MongoDB support. You can test this by changing the following files: #### dejavu.cnf.SAMPLE ``` { "database": { "database": "dejavu" }, "database_type": "mongodb" } ``` #### docker-compose.yaml ```...
Let's say I have the following original tracks: Track A: 5 hours Track B: 3 minutes Now I'm trying to match 10 seconds of audio from Track B. The current...
Anyone know a way to fingerprint the audio input from an HDMI to USB capture card? I'm experimenting with a Raspberry Pi 4 and would like the capture card to...
Fingerprinting a folder with ~10000 tracks consuming 45GB of RAM and it's growing. It seems that the memory is leaked even when a song has completed recognition.
I was wondering if there was documentation on how to interpret each individual metric from a fingerprinting task--like for example, what is the difference between `input_confidence` and `fingerprinted_confidence`? Thanks!
input_song is GHOST result is right [{"song_id":"3","song_name":"GHOST","input_total_hashes":"1078","fingerprinted_hashes_in_db":"74033","hashes_matched_in_input":"9655","input_confidence":"8.9564","fingerprinted_confidence":"0.13041","offset":"783","offset_seconds":"36.36245","file_sha1":"96D9C866D5E246CAD131FC1370A5555A6D74BE83"},{"song_id":"1","song_name":"money","input_total_hashes":"1078","fingerprinted_hashes_in_db":"63688","hashes_matched_in_input":"130","input_confidence":"0.12059","fingerprinted_confidence":"0.00204","offset":"1293","offset_seconds":"60.0468","file_sha1":"343F421F6E1DC76BBBF76FE9B25E7F4BF0D8F54C"}] dejavu/dejavu/__init__.py : INPUT_CONFIDENCE: round(hashes_matched / queried_hashes, 2), Why does the confidence exceed 1? Is it because there's a duplicate hash? High confidence is...
The `offset_seconds` calculation uses the DEFAULT_FS even if the fingerprinted tracks have a different framerate (e.g. 48k) ```python nseconds = round(float(offset) / DEFAULT_FS * DEFAULT_WINDOW_SIZE * DEFAULT_OVERLAP_RATIO, 5) ``` Suggested...
I am having issues using dejavu without docker and I don't know why. I'm not being able to trace the error. Could someone help me with this? Here is what...
In dejavu/dejavu/__init__.py, line 202 nseconds = round(float(offset) / DEFAULT_FS * DEFAULT_WINDOW_SIZE * DEFAULT_OVERLAP_RATIO, 5) I believe that the code above should be ... * (1 - DEFAULT_OVERLAP_RATIO) ... for DEFAULT_WINDOW_SIZE...
So I did some poking around while getting this set up using a Postgres RDS instance and I noticed that the biggest bottleneck when fingerprinting an entire directory was inserting...