Automatic speech segmentation to improve speech synthesis performance

摘要: Emerging growth of information and communication technologies has influenced the research trends to focus on speech technologies. Pre-processing signal serves various purposes in any processing application. It includes noise removal, endpoint detection, pre-emphasis, framing, windowing, echo canceling etc. Out these, automatic word/sentence boundary detection is fundamental step for applications like recognition synthesis. This paper expose problem words sentences silent noisy situations. study proposes an algorithm segmentation Indian languages voiced speech. A modified data based scheme finding entropy placed better performance. The method good performance features than energy-based methods. To determine candidates segments, adaptive threshold used which are related sentences. Simulation results that this will provide energy algorithms.