Compressing English Speech Data with Hybrid Methods without Data Loss

Understanding the mechanism of speech formation is great importance in successful coding signal. It also used for various applications, from authenticating audio files to connecting recording data acquisition device (e.g. microphone). Speech vital acquisition, analysis and evaluation sound, investigation criminal events forensics. For collection, processing, analysis, extraction or sounds recorded as files, which play an important role crime detection, it necessary compress without loss. Since there are many voice changing software available today, number their correct interpretation detecting originality. Using techniques such signal noise extraction, filtering on incomprehensible recording, improving speech, making them comprehensible, determining whether any manipulation understanding original, methods addition subtraction used, sounds, code must be decoded transcribed. In this study, first all, what sound is, its purposes, areas use, classification according some features given. Moreover, our study was done English data. This dataset real consists approximately 100000 recordings. using waveform, vocoders hybrid success all system we created measured. Hybrid models gave more results than others. The obtained will set example future work.