Audio and speech coding device, audio and speech decoding device, method for coding audio and speech, and method for decoding audio and speech
This invention introduces audio/speech encoding apparatus audio/speech decoding apparatus, audio/speech encoding method and audio/speech decoding method to efficiently encode the quantization parameters of split multi-rate lattice vector quantization. In this invention, the position of the sub-vector whose codebook indication consumes the most bits is firstly located, and then the value of the codebook is estimated based on the total number of bits available and the bits usage information for other sub-vectors. The difference value is calculated between the actual value and estimated value. Finally, instead of transmitting the codebook indication which consumes the most bits, the position of the sub-vector whose codebook indication consumes the most bits and the difference value between the actual value and the estimated value are transmitted. By applying of the invented method, bits can be saved by the codebook indications.
1. An audio/speech encoding apparatus that receives audio/speech signals that include a time domain input signal, comprising:
a time to frequency domain transformation section configured to transform the time domain input signal to a frequency spectrum;
a vector quantization section configured to perform quantization splitting the input signal of the frequency spectrum into sub-vectors and generating codebook indications for the sub-vectors, wherein the greater codebook indication, the greater codebook indication bit consumption; and
a codebook indication transformation section configured to transform the codebook indications,
wherein when total bits available is equal to or greater than total bits consumption for the quantization, a position of a sub-vector whose codebook indication consumes the most bits is identified,
when the codebook indication consuming the most bits is greater than a threshold, the identified position, the codebook indications for all the sub-vectors except the codebook indication consuming the most bits, and a difference between bit consumption of the codebook indication consuming the most bits and its estimate are encoded,
when the codebook indication consuming the most bits is not greater than the threshold, the codebook indications for all the sub-vectors except the sub-vector with a pre-identified position and a difference between bit consumption of the codebook indication of the sub-vector with the pre-identified position and its estimate are encoded, and
when the total bits available is less than the total bits consumption for the quantization, the codebook indications are encoded until there are no available bits left; and
a transmitter configured to transmit encoded audio/speech signals over a communication channel to an audio/speech decoding apparatus.
2. The audio/speech encoding apparatus according to claim 1 ,
wherein the estimate of the bit consumption of the codebook indication consuming the most bits is based on the total bits available and bits usage for other sub-vectors.
3. The audio/speech encoding apparatus according to claim 1 ,
wherein the estimate of the bit consuming of the codebook indication of the sub-vector with the pre-identified position is based on the total bits available and bits usage for other sub-vectors.
4. An audio/speech encoding method, which uses an audio/speech encoding apparatus, comprising:
receiving audio/speech signals, including a time domain input signal;
transforming the time domain input signal to a frequency spectrum;
performing quantization splitting the input signal of the frequency spectrum into sub-vectors and generating codebook indications for the sub-vectors, wherein, the greater codebook indication, the greater codebook indication bit consumption;
transforming the codebook indications,
wherein when total bits available is equal to or greater than total bits consumption for the quantization, a position of a sub-vector whose codebook indication consumes the most bits is identified;
when the codebook indication consuming the most bits is greater than a threshold, the identified position, the codebook indications for all the sub-vectors except the codebook indication consuming the most bits and a difference between bit consumption of the codebook indication consuming the most bits and its estimate are encoded; and
when the codebook indication consuming the most bits is not greater than the threshold, the codebook indications for all the sub-vectors except the sub-vector with a pre-identified position and a difference between bit consumption of the codebook indication of the sub-vector with the pre-identified position and its estimate are encoded;
when the total bits available are less than the total bits consumption for the quantization, the codebook indications are encoded until there are no available bits left; and
transmitting encoded audio/speech signals over a communication channel to an audio/speech decoding apparatus.