Objective: - To downsample stereo audio & preserve channel identity, then upsample mono audio back to original stereo formApproach:- Analyzed the working of VoLTE calls and VoIP calls, SIP protocol, IMS framework and codecs like Opus- Designed an upsampling & downsampling algorithm for stereo audio, further extended it to multichannel audio- Implemented and tested the performance of the designed algorithm using Librosa library in Python programming- Explored multiple sample reconstruction techniques and analyzed their performance for audio upsampling taskResult: - Proposed algorithm supports stereo sound effect over VoLTE calls which further enhances the user experience

• Android Audio Processing (SURGE'22)Mentor: Prof. Vipul Arora, Department of Electrical Engineering IITK– Thoroughly analyzed and implemented the C++ Oboe library for high-performance audio on a sample android app– Reduced audio latency from a range of 10ms to 200ms to a range of 0ms to 25ms, which is nearly 88% optimized– Coordinated with the Madhav Lab IITK team and resolved significant issues in the student-end of the Narottam Android Application, a Government of India funded project– Integrated the optimized audio playing and recording mechanism of the sample app into the Narottam Application