Multi-rate system for audio processing

Patent metadata

Field	Value
Publication number	US-2016240183-A1
Application number	US-201514824998-A
Country	US
Kind code	A1
Filing date	Aug 12, 2015
Priority date	Feb 12, 2015
Publication date	Aug 18, 2016
Grant date	—

Abstract

A multi-rate audio processing system and method provides real-time measurement and processing of amplitude/phase changes in the transition band of the lowest frequency subband caused by the audio processing that can be used to apply amplitude/phase compensation to the higher subband(s). Tone signals may be injected into the transition band to provide strong tonal content for measurement and processing. The real-time measurement and compensation adapts to time-varying amplitude/phase changes regardless of the source of the change (e.g. non-linear time-varying linear or user control parameters) and provides universal applicability for any linear audio processing.

First claim

1 . A method of audio reproduction, comprising:
splitting a first digital audio signal into at least a first and a second frequency subbands to generate at least a first and a second subband signal, said first and second frequency subbands separated by a transition frequency band; generating a first downsampled subband signal by downsampling or decimating the first subband signal; injecting at least one tone signal that lies within the transition frequency band into the first downsampled subband signal; generating a first processed subband signal by performing an audio processing on the first downsampled subband signal; measuring in real-time amplitude and phase in the transition frequency band of the first downsampled subband signal and the first processed subband signal; processing in real-time the measurements to calculate a change in amplitude and phase in the transition frequency band; calculating in real-time both an amplitude compensation and a phase compensation for the second frequency subband to approximately match the measured change in amplitude and phase in the first frequency subband; generating in real-time a second processed subband signal by applying said amplitude and phase compensation to the transition band of the second subband signal; removing said at least one tone signal from the first processed subband signal; and combining the first and second processed subband signals to reconstruct an output audio signal.

2 . The method of claim 1 , further comprising:
determining whether there is sufficient amplitude of the first downsampled subband signal in the transition band to calculate the change in amplitude and phase; determining from the measured amplitude and phase whether the first downsampled signal has sufficient tonal content in the transition band to calculate the change in amplitude and phase; and selectively enabling or disabling the injection of the one or more tone signals.

3 . The method of claim 1 , further comprising:
measuring the amplitude and phase in the full first frequency subband of the first downsampled subband signal; determining the amplitude of the one or more tone signals to be higher than a noise signal in the transition band but low enough to make minimal change to the amplitude in the full first frequency subband; and selecting the phase and frequency of the one or more tone signals to avoid cancellation of the first downsampled subband signal in the transition band.

4 . The method of claim 1 , wherein the audio processing is performed on blocks of samples of the first downsampled subband signal, wherein the real-time amplitude and phase measurements and the application of said amplitude and phase compensation are performed on each block of samples.

5 . The method of claim 1 , wherein the at least one tone signal comprises multiple tone signals spaced at different frequencies in the transition band.

6 . The method of claim 5 , wherein calculating the amplitude and phase compensation comprises calculating a weighted average of the measurements for the multiple tone signals to produce a single amplitude compensation value and a single delay compensation value.

7 . The method of claim 5 , wherein calculating the amplitude and delay compensation comprises calculating one or more compensation filters whose amplitude and phase characteristics approximately match amplitude and phase characteristics of the audio processing in the transition band.

8 . The method of claim 1 , further comprising:
processing amplitude measurements of the first downsampled subband signal and the first processed subband signal over the first frequency subband to calculate a change in amplitude over the first frequency subband caused by the audio processing, wherein the amplitude compensation is calculated as a function of the amplitude change over the first frequency subband and the amplitude change over the transition frequency band to balance perfect reconstruction of the output audio signal and preservation of an intended gain or signal level of the audio processing.

9 . The method of claim 8 , wherein
m coef =ƒ( m δtr ,m δRMS ),
ƒ( m δtr ,m δRMS )= G ( d δrms _ tr ,τ),
d δrms _ tr =m δRMS (dB)− m δtr (dB),

m coef is the amplitude compensation value to be applied to the second subbands, ƒ(m δtr , m δRMS ) is a function that calculates the amplitude compensation value m coef based on the transition band amplitude change (m δtr ) and the first frequency subband change (m δRMS ), G( ) determines a balance between the transition band and the first frequency band as a function of amplitude difference d δrms _ tr and target threshold τ.

10 . The method of claim 1 , wherein the audio processing is one of a non-linear process, a linear but time-variant process or a linear or non-linear process subject to user selectable control parameters whereby said audio processing produces amplitude and phase changes that are time-variant, wherein the calculation and application of the amplitude and phase compensation is time-variant.

11 . The method of claim 1 , wherein the real-time amplitude and phase measurements and the application of said amplitude and phase compensation are performed only in the transition band.

12 . The method of claim 1 , wherein the method is implemented in a quadrature mirror filter (QMF) bank, said first frequency subband signal is maximally decimated to the sampling rate of the audio processing, and wherein combining the processed signals includes expanding or upsampling the processed signals to an output sampling rate.

13 . The method of claim 1 , further comprising:
converting the output audio output signal to an analog audio signal; and transducing via one or more speakers the analog audio signal to sound.

14 . The method of claim 1 , wherein the method is implemented in one or more digital processors and memory.

15 . The method of claim 1 , further comprising:
removing the signal in the transition band from the first downsampled subband signal such that the measurements of amplitude or phase measure only the at least one tone signal before and after audio processing, and injecting the signal in the transition band to the first processed subband signal.

16 . A method of audio reproduction, comprising:
splitting a first digital audio signal into at least a first and a second frequency subbands to generate at least a first and a second subband signal, said first and second frequency subbands separated by a transition frequency band; generating a first downsampled subband signal by downsampling or decimating the first subband signal; generating a first processed subband signal by performing an audio processing on the first downsampled subband signal; measuring in real-time amplitude and phase in the transition frequency band of the first downsampled subband signal and the first processed subband signal; processing in real-time the measurements to calculate a change in amplitude and phase in the transition frequency band; calculating in real-time both an amplitude compensation and a phase compensation for the second frequency subband to approximately match the measured change in amplitude and phase in the first frequency subband; generating in real-time a second processed subband signal by applying said amplitude and phase compensation to the transition band of the second subband signal; and combining the first and second processed subband signals to reconstruct an output audio signal.

17 . The method of claim 16 , further compr

Assignees

• Dts Inc

Inventors

• Noh Daekyoung
• Katsianos Themis

Classifications

• H04R3/04

  for correcting frequency response · CPC title

• G10L21/0388 Primary

  Details of processing therefor · CPC title

• H04R3/00

  Circuits for transducers (arrangements for producing a reverberation or echo sound G10K15/08; amplifiers H03F) · CPC title

• H03G5/005

  of digital signals · CPC title

• H03H17/0272

  Quadrature mirror filters · CPC title