Apparatus and method applying adaptive spectral whitening in a high-frequency reconstruction coding system
AssigneeCoding Technologies AB
InventorsKristofer Kjörling, Per Ekstrand, Fredrik Henn, Lars Villemoes
Life of this patent
Estimating the term…
Abstract
The present invention proposes a new method and a new apparatus for enhancement of audio source coding systems utilizing high frequency reconstruction (HFR). It utilizes adaptive filtering to reduce artifacts due to different tonal characteristics in different frequency ranges of an audio signal upon which HFR is performed. Tie present invention is applicable to both speech coding and natural audio coding systems.
Claims19
An apparatus for estimating a level of spectral whitening to be applied to a signal prior to a high-frequency regeneration step or after the high-frequency regeneration step to be performed when generating a high-frequency regenerated signal having a highband which is based on a lowband signal, wherein the spectral whitening is obtained by filtering using a spectral whitening filter, the spectral whitening filter being an adaptive filter being adaptable by means of a filter parameter, the apparatus comprising:
- an estimator for estimating a tonal character of an original signal to be encoded, at a given time, wherein the original audio signal is to be encoded by an audio coder to obtain an encoded audio signal representing only a lowband of the original audio signal, the estimated tonal character including an estimated tonal character of a highband of the original audio signal, which is not included in the encoded audio signal;
- a determinator for determining a varying filter parameter of the spectral whitening filter based on the estimated tonal character;
- and
- an associator for associating the varying filter parameter to the encoded audio signal to obtain a bit stream having the encoded audio signal having the varying filter parameter, the varying filter parameter being dependent on the encoded audio signal.
An apparatus for producing an output signal based on a decoded version of an encoded audio signal representing a lowband of an original audio signal, the encoded audio signal having associated therewith a varying filter parameter for a spectral whitening filter, the varying filter parameter depending on a tonal character of a highband of the original audio signal at a given time, the apparatus comprising:
- a demultiplexer for obtaining the varying filter parameter associated with the encoded audio signal;
- a high-frequency reconstructor for performing a high frequency reconstruction step on a decoded version of the encoded audio signal to produce a high-frequency reconstructed signal;
- and
- an adaptive spectral whitening filter for filtering the decoded version or the high-frequency regenerated signal;
- wherein the adaptive spectral whitening filter has a variable parameter, the variable parameter being set in accordance with the varying filter parameter associated with the encoded audio signal.
A method for estimating a level of spectral whitening to be applied to a signal prior to a high-frequency regeneration step or after the high-frequency regeneration step to be performed when generating a high-frequency regenerated signal having a highband which is based on a lowband signal, wherein the spectral whitening is obtained by filtering using a spectral whitening filter, the spectral whitening filter being an adaptive filter being adaptable by means of a filter parameter, the method comprising:
- estimating a tonal character of an original audio signal to be encoded, at a given time, wherein the original audio signal is to be encoded by an audio coder to obtain an encoded audio signal representing only a lowband of the original audio signal, the estimated tonal character including an estimated tonal character of a highband of the original audio signal, which is not included in the encoded audio signal;
- determining a varying filter parameter of the spectral whitening filter based on the estimated tonal character;
- and
- associating the varying filter parameter to the encoded audio signal to obtain a bit stream having the encoded audio signal having the varying filter parameter, the varying filter parameter being dependent on the encoded audio signal.
Method for producing an output signal based on a decoded version of an encoded audio signal representing a lowband of an original audio signal, the encoded audio signal having associated therewith a varying filter parameter for a spectral whitening filter, the varying filter parameter depending on a tonal character of a highband of the original audio signal at a given time, the method comprising the following steps:
- obtaining the varying filter parameter associated with the encoded audio signal;
- performing a high-frequency regeneration step on a decoded version of the encoded audio signal to produce a high frequency regenerated signal;
- and
- filtering the decoded version or the high-frequency regenerated signal using an adaptive spectral whitening filter;
- wherein the adaptive spectral whitening filter has a variable parameter, the variable parameter being set in accordance with the varying filter parameter associated with the encoded audio signal.
An encoder for encoding an original audio signal to obtain an encoded version thereof, comprising:
- an apparatus for estimating a level of spectral whitening to be applied to a signal prior to a high-frequency regeneration step or after the high-frequency regeneration step to be performed when generating a high-frequency regenerated signal having a highband which is based on a lowband signal, wherein the spectral whitening is obtained by filtering using a spectral whitening filter, the spectral whitening filter being an adaptive filter being adaptable by means of a filter parameter, the apparatus comprising:
- an estimator for estimating a tonal character of an original signal to be encoded, at a given time, wherein the original audio signal is to be encoded by an audio coder to obtain an encoded audio signal representing only a lowband of the original audio signal, the estimated tonal character including an estimated tonal character of a highband of the original audio signal, which is not included in the encoded audio signal;
- a determinator for determining a varying filter parameter of the spectral whitening filter based on the estimated tonal character;
- and
- an associator for associating the varying filter parameter to the encoded audio signal to obtain a bit stream having the encoded audio signal having the varying filter
- parameter, the varying filter parameter being dependent on the encoded audio signal;
- an audio encoder for encoding the original audio signal to obtain the encoded version thereof;
- an estimator for estimating a spectral envelope of the original audio signal to obtain an estimated spectral envelope;
- and
- a multiplexer for multiplexing the encoded version of the original audio signal, the filter parameter of the spectral whitening filter and the estimated spectral envelope for obtaining a bit stream.
A decoder for decoding a bit stream including an encoded version of an original audio signal, an estimated spectral envelope and a filter parameter to be applied to a spectral whitening filter, the decoder comprising:
- a bit stream demultiplexer for extracting the encoded version of the original audio signal, the estimated spectral envelope and the filter parameter;
- an audio decoder for decoding the encoded version of the original audio signal to obtain a lowband signal;
- an envelope decoder for decoding the estimated spectral envelope;
- an apparatus for producing an output signal based on a decoded version of an encoded audio signal representing a lowband of an original audio signal, the encoded audio signal having associated therewith a varying filter parameter for a spectral whitening filter, the varying filter parameter depending on a tonal character of a highband of the original audio signal at a given time, the apparatus comprising:
- a demultiplexer for obtaining the varying filter parameter associated with the encoded audio signal;
- a high-frequency reconstructor for performing a high frequency reconstruction step on a decoded version of the encoded audio signal to produce a high-frequency reconstructed signal;
- and
- an adaptive spectral whitening filter for filtering the decoded version or the high-frequency regenerated signal, wherein the adaptive spectral whitening filter has a variable parameter, the variable parameter being set in accordance with the varying filter parameter associated with the encoded audio signal;
- and
- a summer for summing an adaptively spectral whitened high frequency regenerated signal and a delayed version of the decoded audio signal to obtain a wideband output signal.
Method for encoding an original audio signal to obtain an encoded version thereof, comprising the following steps:
- estimating a level of spectral whitening to be applied to a signal prior to a high-frequency regeneration step or after the high-frequency regeneration step to be performed when generating a high-frequency regenerated signal having a highband which is based on a lowband signal, wherein the spectral whitening is obtained by filtering using a spectral whitening filter, the spectral whitening filter being an adaptive filter being adaptable by means of a filter parameter, the step of estimating including:
- estimating a tonal character of an original audio signal to be encoded, at a given time, wherein the original audio signal is to be encoded by an audio coder to obtain an encoded audio signal representing only a lowband of the original audio signal, the estimated tonal character including an estimated tonal character of a highband of the original audio signal, which is not included in the encoded audio signal;
- determining a varying filter parameter of the spectral whitening filter based on the estimated tonal character;
- and
- associating the varying filter parameter to the encoded audio signal to obtain a bit stream having the encoded audio signal having the varying filter parameter, the varying filter parameter being dependent on the encoded audio signal;
- encoding the original audio signal to obtain the encoded version thereof;
- estimating a spectral envelope of the original audio signal to obtain an estimated spectral envelope;
- and
- multiplexing the encoded version of the original audio signal, the filter parameter of the spectral whitening filter and the estimated spectral envelope for obtaining a bit stream.
A method for decoding a bit stream including an encoded version of an original audio signal, an estimated spectral envelope and a filter parameter to be applied to a spectral whitening filter, the method comprising:
- extracting the encoded version of the original audio signal, the estimated spectral envelope and the filter parameter;
- decoding the encoded version of the original audio signal to obtain a lowband signal;
- decoding the estimated spectral envelope;
- producing an output signal based on a decoded version of an encoded audio signal representing a lowband of an original audio signal, the encoded audio signal having associated therewith a varying filter parameter for a spectral whitening filter, the varying filter parameter depending on a tonal character of a highband of the original audio signal at a given time, the step of producing comprising:
- obtaining the varying filter parameter associated with the encoded audio signal;
- performing a high-frequency regeneration step on a decoded version of the encoded audio signal to produce a high-frequency regenerated signal;
- and
- filtering the decoded version or the high-frequency regenerated signal using an adaptive spectral whitening filter, wherein the adaptive spectral whitening filter has a variable parameter, the variable parameter being set in accordance with the varying filter parameter associated with the encoded audio signal;
- and
- summing an adaptively spectral whitened high-frequency regenerated signal and a delayed version of the decoded audio signal to obtain a wideband output signal.
Dependencies are inferred from the claim text.