A real-time audio transcoder with watermark images for content protection and quality of service monitoring

Audio streaming has become a popular application on the Internet over the past few years, and is expected to grow with the advent of high-speed Internet access. Content adaptability, copyright protection, and Quality of Service (QoS) monitoring of these streams are some of the major issues encountered in audio streaming. In this research study, we present a universal media access-compliant (UMA) transcoding module along with a watermarking-based data-hiding module that provides copyright protection as well as data integrity, authenticity assurance, and QoS monitoring. The audio was first adapted to multiple clients' needs on the server side, watermarked, and then sent to the connected clients. The content adaptability was done by down sampling and a sample size reduction of the live microphone-captured audio. The adaptation was done in conformance to the UMA paradigm. The watermarking, on the other hand, was done on different levels in order to provide guaranteed reliability and authenticity of the audio data as well as an efficient lightweight quality of service monitoring. The watermarking scheme was done on three different levels: First, the private watermark was watermarked by a public watermark. Second, the audio was watermarked by the watermarked private watermark using a private key. Third, the watermarked audio was watermarked again by a hashed value or itself using the same private key. The audio contains an image along with encryption and another image protecting it from temperament and allowing users to track the quality of the audio. The client would have the ability to extract the private watermark and the hash value with his private key, check the integrity of the audio using this hash value and the received audio, extract the public watermark from the private watermark, and compare it to its supposed value in order to obtain a broad idea of the quality of the service. The watermarking techniques used throughout this research were least significant bit (LSB) embedding and some variations of this technique. All the audio data that is outputted can be played with legacy audio players. The content adaptation and the watermarking will not change in any way the conformance of the audio to legacy standards.