Spectral translation/folding in the subband domain
The present invention relates to a new method and apparatus for improvement of High Frequency Reconstruction (HFR) techniques using frequency translation or folding or a combination thereof. The proposed invention is applicable to audio source coding systems, and offers significantly reduced computational complexity. This is accomplished by means of frequency translation or folding in the subband domain, preferably integrated with spectral envelope adjustment in the same domain. The concept of dissonance guard-band filtering is further presented. The proposed invention offers a low-complexity, intermediate quality HFR method useful in speech and natural audio coding applications.
1. An apparatus for reconstructing a high frequency portion of an audio signal, the apparatus comprising:
a complex exponential modulated analysis filterbank for filtering a low frequency portion of the audio signal to produce a plurality of low frequency complex-valued subband signals, wherein the complex exponential modulated analysis filterbank includes a plurality of decimators;
a high frequency reconstructor that reconstructs the high frequency portion of the audio signal by patching both a real and an imaginary part of a consecutive number of the plurality of low frequency complex-valued subband signals to consecutive subbands of the high frequency portion; and
a complex exponential modulated synthesis filterbank for generating a wideband audio signal by combining the reconstructed high frequency portion of the audio signal with the low frequency portion of the audio signal, wherein the complex exponential modulated synthesis filterbank includes a plurality of interpolators,
wherein the high frequency reconstructor uses a parameter indicating a quantity of the consecutive number of the plurality of low frequency complex-valued subband signals.
2. The apparatus of claim 1 wherein the complex exponential modulated analysis filterbank and the complex exponential modulated synthesis filterbank have L channels.
3. The apparatus of claim 1 wherein the high frequency reconstructor is configured to reconstruct the high frequency portion of the audio signal with multiple patches.
4. The apparatus of claim 1 wherein the plurality of decimators each have a decimation factor of M.
5. The apparatus of claim 1 wherein the plurality of interpolators each have an interpolation factor of M.
6. The apparatus of claim 2 wherein the plurality of decimators and the plurality of interpolators each have an interpolation factor of M, which is equal to L.
7. A method for reconstructing a high frequency portion of an audio signal, the method comprising:
filtering a low frequency portion of the audio signal with a complex exponential modulated analysis filterbank to produce a plurality of low frequency complex-valued subband signals, wherein the filtering includes decimating the plurality of low frequency subband signals;
reconstructing the high frequency portion of the audio signal by patching both a real and an imaginary art of a consecutive number of the plurality of low frequency complex-valued subband signals to consecutive subbands of the high frequency portion; and
generating a wideband audio signal with a complex exponential modulated synthesis filterbank by combining the reconstructed high frequency portion of the audio signal with the low frequency portion of the audio signal, wherein the generating includes interpolating the plurality of low frequency subband signals,
wherein the reconstructing uses a parameter indicating a quantity of the consecutive number of the plurality of low frequency complex-valued subband signals.