Encoding apparatus, decoding apparatus, information processing system, encoding method, and program

A decoding apparatus comprising:

• a code-table storage unit that stores a code table in which a variable-length code corresponding to a plurality of quantized values that are all zero is not included;
• a decoding unit that decodes a variable-length code into the plurality of quantized values in accordance with the code table generated in accordance with the probability of occurrence of the plurality of quantized values;
• and
• an inverse-normalization unit that generates a plurality of output values, as a plurality of input values in an input signal, by using a plurality of normalized values generated by dequantizing the plurality of quantized values obtained as a result of decoding performed by the decoding unit and a normalization coefficient corresponding to the plurality of normalized values.

An information processing system comprising:

• an encoding apparatus that includes
• a normalization unit that calculates a plurality of normalized values by dividing a plurality of input values in an input signal by either a normalization coefficient that is closest to a maximum value of absolute values of the plurality of input values or a normalization coefficient that is closest to the maximum value from among normalization coefficients that are larger than the maximum value,
• a quantization unit that generates a plurality of quantized values by quantizing the plurality of normalized values,
• a code-table storage unit that stores a code table in which the smaller the probability of occurrence of the plurality of quantized values, the longer the code length of a variable-length code allocated to the plurality of quantized values, and
• a code output unit that outputs, when the plurality of the quantized values are all zero, a variable-length code allocated to a combination of a plurality of quantized values, which are different from the plurality of quantized values, in accordance with the code table;
• and
• a decoding apparatus that includes
• a code-table storage unit that stores a code table the same as that stored in the encoding apparatus,
• a decoding unit that decodes the variable-length code output from the encoding apparatus into the plurality of quantized values in accordance with the code table, and
• an inverse-normalization unit that generates a plurality of output values, as the plurality of input values, by using a plurality of normalized values generated by dequantizing the plurality of quantized values obtained as a result of decoding performed by the decoding unit and the normalization coefficient corresponding to the plurality of normalized values.

An encoding method for an encoding apparatus that includes a code-table storage unit that stores a code table in which the smaller the probability of occurrence of a plurality of quantized values, the longer the code length of a variable-length code allocated to the plurality of quantized values, the encoding method comprising the steps of:

• calculating a plurality of normalized values by dividing a plurality of input values in an input signal by either a normalization coefficient that is closest to a maximum value of absolute values of the plurality of input values or a normalization coefficient that is closest to the maximum value from among normalization coefficients that are larger than the maximum value;
• generating a plurality of quantized values by quantizing the plurality of normalized values;
• and
• outputting, when the plurality of quantized values are all zero, a variable-length code allocated to a combination of a plurality of quantized values, which are different from the plurality of quantized values, in accordance with the code table.

A program causing a computer for an encoding apparatus that includes a code-table storage unit that stores a code table in which the smaller the probability of occurrence of a plurality of quantized values, the longer the code length of a variable-length code allocated to the plurality of quantized values, to perform the steps of:

• calculating a plurality of normalized values by dividing a plurality of input values in an input signal by either a normalization coefficient that is closest to a maximum value of absolute values of the plurality of input values or a normalization coefficient that is closest to the maximum value from among normalization coefficients that are larger than the maximum value;
• generating a plurality of quantized values by quantizing the plurality of normalized values;
• and
• outputting, when the plurality of quantized values are all zero, a variable-length code allocated to a combination of a plurality of quantized values, which are different from the plurality of quantized values, in accordance with the code table.