Bit allocation method in subband coding

A bit allocation method in a subband coding method including the steps of dividing digital data with a sampling frequency into a plurality of M subband signals SB.sub.m, where m=1 . . . M, said M subband signals SB.sub.m having successive frequency bands, dividing each subband signal SB.sub.m into a plurality of blocks having q samples each, and allocating ab[m] bits to each block for quantizing each sample, said bit allocation method comprising:

• a first step of calculating a needed number of bits bf[m] required for quantization of the q samples for each block of each subband signal SB.sub.m ;
• a second step of selecting a subband SB.sub.jm from among the subband signals SB.sub.m whose needed number of bits bf[m] is largest among the result of said first step, replacing a needed number of bits bf[jm] for the selected subband SB.sub.jm with a value obtained by subtracting one from the needed number of bits bf[jm], replacing an allocated number of bits ab[jm]for the selected subband SB.sub.jm with a value obtained by adding one to the allocated number of bits ab[jm], and replacing an available total number of bits TB allocable to all blocks together with a value obtained by subtracting one from the total number of bits TB;
• a third step of returning to said second step when the available total number of bits TB is greater than or equal to one, by checking the available total number of bits TB;
• and
• a fourth step of replacing the allocated number of bits ab[m] with a value obtained by subtracting one from the allocated number of bits ab[m], replacing the available total number of bits TB with a value obtained by adding a number of subbands K whose allocated number of bits ab[m] is one to the available total number of bits TB, and returning to said second step, said fourth step being performed when a presence of a subband whose allocated number of bits ab[m] equals one is ascertained under a condition that the available total number of bits TB is below one.