SoundWave

Real-Time Audio Processing and Speech-to-Text with Kafka and Flink

Table of Contents

1. Project Overview
2. Architecture
3. Audio Processing
   • Kafka Setup
   • Flink Setup

Project Overview

SoundWave is a real-time audio processing pipeline that ingests audio streams from multiple sources, performs audio processing tasks (primarily pitch shift), and stores the processed audio, which can be further processed to generate a transcript. Built using Apache Kafka, Apache Flink and GlusterFS, SoundWave is optimized for fault tolerance, data redundancy, and efficient storage, using a distributed setup to ensure scalability and reliability.

Architecture

Audio Processing

Kafka Setup

download the latest kafka release and extract it from here here

tar -xzf kafka_2.13-3.8.1.tgz
cd kafka_2.13-3.8.1

run the following commands in separate sessions in order to start all services in the correct order:

bin/zookeeper-server-start.sh config/zookeeper.properties
bin/kafka-server-start.sh config/server.properties

now, your kafka broker is up and running.

Flink Setup

download flink sql connector jar from here and correctly replace the JAR path in flink.properties.