Speech recognition using equal division quantization

A speech processing apparatus comprising:

• input means for inputting speech data;
• quantization range obtaining means for obtaining a quantization range of the input speech data;
• integral value obtaining means for obtaining an integral value of an output probability of the obtained quantization range by using an output probability distribution of statistical models of speech data which has previously been obtained;
• and
• quantization code book forming means for determining a quantum so as to equally divide the integral value, thereby forming a quantization code book.

A storage medium which can be read out by a computer, for storing the programs instructing the computer to perform the steps of:

• inputting speech data;
• obtaining a quantization range of the input speech data;
• obtaining an integral value of an output probability of the obtained quantization range by using an output probability distribution of statistical models of speech data which has previously been obtained;
• and
• determining a quantum so as to equally divide the integral value and forming a quantization code book.

A program which is read out and is executed by a computer, which instructs the computer to perform the steps:

• inputting speech data;
• obtaining a quantization range of the input speech data;
• obtaining an integral value of an output probability of the obtained quantization range by using an output probability distribution of statistical models of speech data which has previously been obtained;
• and
• determining a quantum so as to equally divide the integral value and forming a quantization code book.