Method and apparatus for compressing a digital input signal with block floating applied to blocks corresponding to fractions of a critical band or to multiple critical bands

A method for compressing a digital input signal to provide a compressed signal, the method comprising the steps of:

• deriving spectral components from the digital input signal;
• dividing the spectral components into critical bands;
• dividing the spectral components into a block having a bandwidth corresponding to plural critical bands;
• applying block floating processing to the block to generate a floating coefficient;
• and
• quantizing the spectral components in each critical band using an adaptively-allocated number of bits.

A method for compressing a digital input signal to provide a compressed signal, the method comprising the steps of:

• deriving spectral components from the digital input signal;
• dividing the spectral components into critical bands;
• dividing the spectral components in a critical band into blocks having a bandwidth narrower than the critical band;
• applying block floating processing to the blocks to generate a floating coefficient for each block;
• and
• quantizing the spectral components in each block using an adaptively-allocated number of bits.

A method for compressing a digital input signal to provide a compressed signal, the method comprising the steps of:

• deriving spectral components from the digital input signal;
• dividing the spectral components into critical bands;
• dividing the spectral components in a critical band into blocks having a bandwidth narrower than the critical band;
• applying block floating processing to the blocks to generate a floating coefficient for each block;
• quantizing the spectral components in each block using a number of bits allocated to the block according to an allowed noise level for the critical band, the number of bits allocated being indicated by a word length;
• and
• including the compressed signal, instead of the floating coefficient for each block, the allowed noise level with a range that is shifted downwards in level by a predetermined level from a signal level range in the critical band.

An apparatus for compressing a digital input signal to provide a compressed signal, the apparatus comprising:

• an orthogonal transform circuit whereto the input signal is connected;
• means for dividing spectral components from the orthogonal transform circuit into critical bands;
• means for dividing the spectral components in a critical band into blocks having a bandwidth narrower than the critical band;
• means for applying block floating processing to the blocks to generate a floating coefficient for each block;
• means for quantizing the spectral components in each block using a number of bits allocated to the block according to an allowed noise level for the critical band, the number of bits being indicated by a word length;
• and
• means for including in the compressed signal the word length for each block, and, instead of the floating coefficient for each block, the allowed noise level for the critical band.

An apparatus for compressing a digital input signal to provide a compressed signal, the apparatus comprising:

• an orthogonal transform circuit whereto the digital input signal is connected;
• means for dividing spectral components from the orthogonal transform circuit into critical bands;
• means for dividing the spectral components into a block having a bandwidth corresponding to plural critical bands;
• means for applying block floating processing to the block to generate a block floating coefficient;
• and
• means for quantizing the spectral components in each critical band using an adaptively-allocated number of bits.