sound compression by optimums

compress and reconstruct, about 4 times copression for speech sound.

http://simulations.narod.ru/

run test1.m It record sound from microphone then apply 10 kHz lowpass filer, then find optimums. In compressed data there are optimums shifts and optimums value. The method is works if signal has 90 - 10000 Hz frequency range. about 4 times compression for speech sound. Compression ratio for continues a-sound is about 20 times. At the end it plot graphs and plays reconstructed signal. Of cause mp3 is much better method then this.