Automatic Technique in Frequency Domain for Near-Lossless Time-Scale Modification of Audio

Abstract

Time-scale modification of sounds has been a topic of interest in computer music since its very beginning. Many different techniques both in time and frequency domain have been proposed to solve the problem. Some frequency domain techniques yield high-quality results and can work with large modification factors. However, they present some artifacts, like phasiness, loss of attack sharpness and loss of stereo image. In this paper we propose a new frequency domain approach for an automatic near-lossless time stretching of audio.