Audio coding and quantization method

An audio coding and quantization method comprising the steps of:

• converting each of blocks of an input audio signal into a number of spectral subband components, the blocks being produced from the signal along a time axis;
• converting a related one of the blocks into an input vector of frequency domain values;
• quantizing each subband component whereby the number of quantizer levels for a related one of spectral subbands is obtained from a bit allocation;
• controlling the bit allocation for each subband by using a psychoacoustic model which generates an allowed distortion level of a related one of scalefactor bands corresponding to the subbands;
• calculating, during the controlling step, a quantization of the frequency domain values of the related block through a first control loop, the first control loop being repeated until a bit rate requirement is met that a count of bits needed to encode a bitstream is less than a count of bits available to encode the bitstream;
• and
• calculating, through a second control loop, a quantization noise for each subband, produced by the quantization of the frequency domain values within the first control loop, the second control loop being repeated until a masking requirement is met that a quantization error level of the frequency domain values with scalefactors applied to the values within the scalefactor bands is less than the allowed distortion level,
• wherein the first control loop and the second control loop are alternately performed for the related block such that both the bit rate requirement and the masking requirement are met, and, after both the requirements are met, an output vector of quantized frequency domain values is produced;
• and
• wherein said method further comprises the steps of:
• calculating a parameter of the audio signal based on the scalefactors applied to the frequency domain values within the scalefactor bands every time the calculation of the quantization of the frequency domain values of each block through the first control loop is completed;
• and storing the parameter after completion of the calculation thereof.