Switched filterbank for use in audio signal coding
AssigneeLucent Technologies Inc.
InventorsJames David Johnston, Deepen Sinha
Life of this patent
Estimating the term…
Abstract
An audio coding technique which utilizes a signal adaptive switched filterbank having a first filterbank and a wavelet filterbank. The filterbank switches between the first filterbank and the wavelet filterbank to filter an input signal as a function of the stationarity of the input signal. The first filterbank is utilized to filter stationary signal components. The wavelet filterbank is utilized to filter non-stationary signal components (e.g., attacks).
Claims20
A method for encoding an audio signal, said method comprising the steps of:
- sampling said audio signal;
- alternatively filtering said sampled audio signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said audio signal;
- and
- encoding said filtered signal to provide a compressed output signal.
A method of encoding an audio signal, said method comprising the steps of:
- generating a plurality of noise threshold values as a function of the frequency characteristics of said audio signal;
- alternatively filtering said audio signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said audio signal;
- quantizing said filtered signal, the coarseness of said quantizing being determined by said noise threshold values;
- and
- perceptually encoding said quantized signal.
A method for encoding a digital audio signal to generate a compressed output signal, said method comprising the steps of:
- generating a plurality of noise threshold values as a function of the frequency characteristics of said digital signal;
- alternatively filtering said digital signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said audio signal;
- and
- perceptually encoding said filtered signal to provide said compressed output signal.
An apparatus for encoding an audio signal, said apparatus comprising:
- means for sampling said audio;
- means for alternatively filtering said sampled audio signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said audio signal;
- and
- means for encoding said filtered signal to produce a compressed output signal.
An apparatus for encoding an audio signal, said apparatus comprising:
- means for generating a plurality of noise threshold values as a function of the frequency characteristics of said audio signal;
- means for sampling said audio signal;
- means for alternatively filtering said sampled audio signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said audio signal;
- means for quantizing said filtered signal, the coarseness of said quantizing being controlled by said noise threshold values;
- and
- means for perceptually encoding said quantized signal.
An apparatus for processing a perceptually encoded audio signal, said perceptually encoded audio signal being produced by generating a plurality of noise threshold values as a function of the frequency characteristics of an input signal;
- sampling said input signal;
- alternatively filtering said sampled signal by switching between a first filterbank and a wavelet filterbank to produce a filtered signal, said wavelet filterbank being a tree-structured non-uniform filterbank, said first filterbank being independent from said wavelet filterbank, and said switching occurring in response to the stationarity of said input signal;
- quantizing said filtered signal, the coarseness of said quantizing being determined by said noise threshold values;
- perceptually encoding said quantized signal to produce said perceptually encoded audio signal;
- and applying said perceptually encoded audio signal to a communications channel, said apparatus comprising:
- means for receiving said perceptually encoded audio signal from said communications channel;
- means for decoding said received perceptually encoded audio signal;
- and
- means for recovering said input signal from said decoded audio signal.
Dependencies are inferred from the claim text.