https://github.com/worldveil/dejavu
python
45 songs = 377 MB db
https://willdrevo.com/fingerprinting-and-audio-recognition-with-python/
"the fingerprints take up a surprising amount of space (slighty more than raw MP3 files)"
needs mysql or postgresql
pr for sqlite and mongo ignored
maintenance iffy

musly
1.1 KiB per song

chromaprint
3.4 KiB per song
acoustid-based, "not a general purpose audio fingerprinting solution"
"target use cases are full audio file identifcation, duplicate audio file detection and long audio stream monitoring"
https://oxygene.sk/2011/01/how-does-chromaprint-work/
"Chromaprint works by analyzing the first two minutes of a track, detecting the strength in each of 12 pitch classes, storing these 8 times per second"
yes

https://github.com/spotify/echoprint-server
https://github.com/spotify/echoprint-codegen
c++
abandoned 2016
27 KiB per song

https://github.com/JorenSix/Olaf
C, AGPL
"mainly targets 32-bit ARM devices such as some Teensy’s, some Arduino’s and the ESP32"
"On traditional computers fingerprints are stored in a high-performance key-value-store: LMDB"

https://github.com/gurushida/mnemophonix
C, MIT
130 songs = 75 MB db
haar
"simplified version in C of this work [how to build a Shazam-like system]"
"built from scratch without any dependency"
"only moderately optimized with multithreading"
"no attempt at storing the signatures in an optimized way"
"2 hour movie produces a 16Mb signature and takes about 55 seconds on a MacBook Pro"

https://github.com/dpwe/audfprint
python, MIT
hardcapped at 2^20 (~1M) distinct fingerprints
"times larger than 2^14*0.023 = 380 sec, or about 6 mins, are aliased"
"by default, the database can only remember 2^18 = 262k tracks"

https://github.com/methi1999/Findit
anothe shazam impl
python3, sox, PyDub, SciPy, NumPy and Matplotlib
meh results

https://github.com/a-gram/audioneex
looks sick but shady deps

others
Shazam
MusicRetrieval