Victor Company of Japan, Limited

Takafumi Kizuki

Toshihiro Maruyama

Susumu Takahashi

Kyoichi Shimizu

An audio signal is compressed and decompressed. The audio signal is sampled per regular sampling interval to obtain sampled values. The sampled values are orthogonal-transformed per constant period to obtain sequential orthogonal transfer functions. Feature points are extracted that represent change in coefficient intensity of the orthogonal transfer functions per constant frequency interval. Coefficient intensity data, frequency data, and time data of the feature points are obtained. The orthogonal transfer functions are reproduced by interpolating the coefficient intensity of the orthogonal transfer functions on the basis of the coefficient intensity data, frequency data, and time data. The reproduced orthogonal transfer functions are inverse orthogonal-transformed to reproduce the sampled values. And, the reproduced sampled values are converted into the audio signal.

Audio signal compression and decompression utilizing amplitude,

A signal coding system capable of high efficiency, high quality signal coding is provided. Digital signals represented in the time domain are divided into set time interval data units and output. One output is converted to a digital signal represented in the frequency domain, and the other is output as-is. The energy dispersion of the digital signal represented in the frequency domain is compared with that of the digital signal represented in the time domain, and the digital signal having the least energy dispersion is coded. This coded digital signal is then multiplexed with an identification signal to identify it as a frequency domain or time domain signal, and the resulting multiplexed signal is output.

Growing community of inventors

Yokosuka, Japan

Takafumi Kizuki