Music-Separation-as-a-service (MSaaS)
Created a kubernetes cluster that provides a REST API for automatic music separation service and prepares the different tracks for retrieval.
Containers deployed provide the following services:
- rest - the REST frontend will accept API requests for analysis and handle queries concerning MP3's. The REST worker will queue tasks to workers using
redisqueues. - worker - Worker nodes will receive work requests to analyze MP3's and cache results in a cloud object store (here, Min.io).
- redis - a Redis deployment and service to provide a redis database server.
The worker uses open source waveform source separation analysis software from Facebook. Our reason for turning this into a micro-service is because it takes a long time to run the source separation (about 3-4x the running time of a song).
Rather than send the large MP3 files through Redis, I am using the Min.io object storage system to store the song contents and the output of the waveform separation.
In my solution, I have a single 'bucket' called "queue" that holds the objects containing MP3 songs to process and another bucket called "output" to hold the results ready for download:
The "output" bucket has objects named <songhash>-<track>.mp3 that contain the separate tracks of a song:
Two programs sample-requests.py and short-sample-requests.py were produced that make some sample requests.