Hi folks , hope you guys are doing great , i just want to share project i have built with you guys .it is an audio recognition engine implementation base on the Industrial-Strength Audio Search Algorithm with the backend in rust (axum) and the frontend in react.
for audio decoding i use symphonia and rustfft for the fingerprinting, the hashes are stored in a postgres database and the raw audio file in rustfft (s3 storage).
to prevent the server to get overloaded and reduce latency , the track decode and fingerprinting is handle by a work who pull the jobs from the database.
in release move the lantency for identification on a database of 100 tacks (i know it is small) is around 175 ms for me. i have also added a python script to help you generate the sample programmatically.
here is the repo link: https://github.com/lessan-cyber/Auris
Note: this is my first major project in rust , for those you will read the code let me know if there possible improvements. also the frontend have some burgs i am trying to figure out