Cambridge Silicon Radio Limited

Esfandiar Zavarehei

David Richard Hargreaves

David Hargreaves

A method and apparatus are provided for estimating the noise spectrum of an audio signal occupying a certain frequency range. The amplitude of the audio signal is measured at only a subset of the frequency range having a low expected signal to noise ratio at the time of measuring such that the measured amplitude is indicative of noise amplitude at that subset of the frequency range. A predefined noise spectral shape is then adapted by fitting to the noise amplitude so as to form the estimated noise spectrum. The noise spectrum so estimated is used to remove interference noise from the audio signal.

Noise estimation

Methods of encoding a signal using a perceptual model are described in which a signal to mask ratio parameter within the perceptual model is tuned. The signal to mask ratio parameter is tuned based on a function of the bitrate of the part of the signal which has already been encoded and the target bitrate for the encoding process. The tuned signal to mask ratio parameter is used to compute a masking threshold for the signal which is then used to quantize the signal.

Adaptive tuning of the perceptual model

Methods of spectral partitioning which may be implemented in an encoder are described. The methods comprise determining an estimate of bit requirements for each of a plurality of spectral sub-bands. These estimates are then used to group the sub-bands into two or more regions by minimizing a cost function. This cost function is based on the estimates of bit requirements for each sub-band and the estimates may include estimates of code bit requirements and/or additional code bit requirements for each sub-band. These estimates may be determined in many different ways and a number of methods are described.

Fast spectral partitioning for efficient encoding

Methods of encoding a signal using a perceptual model are described in which a signal to mask ratio parameter within the perceptual model is tuned. The signal to mask ratio parameter is tuned based on a function of the bitrate of the part of the signal which has already been encoded and the target bitrate for the encoding process. The tuned signal to 5 mask ratio parameter is used to compute a masking threshold for the signal which is then used to quantise the signal.

Methods and apparatus for iteratively encoding a portion of a signal are described in which the portion of the signal is quantised and an output bit count is estimated based on the sum of logarithms to base n of values of each sample in the plurality of quantised samples and the total number of samples. The output bit count corresponds to an estimate of the output bit count for the portion of the signal once encoded using a code, such as a Huffman code.

Growing community of inventors

Cambridge, United Kingdom

Esfandiar Zavarehei