Processing Audio Signals with Adaptive Time or Frequency Resolution

Abstract

In some aspects, a method performed in an audio processing apparatus for decoding an audio signal encoded with variable time-frequency resolution is disclosed. The method includes receiving the encoded audio signal, extracting from the encoded audio signal a frame of spectral coefficients, dividing the frame of spectral coefficients into a plurality of frequency bands, and extracting from the encoded audio signal a time-frequency resolution parameter for each of the plurality of frequency bands. The time-frequency resolution parameter indicates a resolution for a current band. The method further includes decoding the plurality of spectral coefficients with a variable time-frequency resolution based at least in part on the time-frequency resolution parameters.

1 . A method performed in an audio processing apparatus for encoding a time domain audio signal with variable time-frequency resolution, the method comprising:
 receiving the time domain audio signal;   transforming the time domain audio signal into a frequency domain representation comprising a frame of spectral coefficients;   dividing the frame of spectral coefficients into a plurality of frequency bands;   generating a time-frequency resolution parameter for each of the plurality of frequency bands, the time-frequency resolution parameter indicating a resolution for a current band; and   encoding the plurality of spectral coefficients and the time-frequency resolution parameters into an encoded audio bitstream,   wherein and the audio processing apparatus is implemented at least in part with some hardware.   
     
     
         2 . The method of  claim 1  wherein the at least some of the time-frequency resolution parameters indicate a change in resolution of the current band from a previous band. 
     
     
         3 . The method of  claim 1  further comprising generating and encoding a second time-frequency resolution parameter into the encoded audio bitstream, the second time-frequency resolution parameter indicating a degree or amount of resolution change. 
     
     
         4 . The method of  claim 1  further comprising generating and encoding a level parameter for each of the plurality of frequency bands, wherein the level parameters represent a spectral profile of the audio signal. 
     
     
         5 . The method of  claim 1  wherein the plurality of frequency bands are critical bands approximating the Bark scale. 
     
     
         6 . The method of  claim 1  wherein the frame of spectral coefficients were generated by overlapping successive windows. 
     
     
         7 . The method of  claim 3  further comprising generating and encoding a transient parameter indicating whether a frame of audio contains a transient. 
     
     
         8 . The method of  claim 7  wherein both the transient parameter and the second time-frequency resolution parameter are used to determine the degree or amount of resolution change. 
     
     
         9 . The method of  claim 1  wherein a temporal resolution is increased for bands including at least one transient. 
     
     
         10 . A method performed in an audio processing apparatus for decoding an audio signal encoded with variable time-frequency resolution, the method comprising:
 receiving the encoded audio signal;   extracting from the encoded audio signal a frame of spectral coefficients;   dividing the frame of spectral coefficients into a plurality of frequency bands;   extracting from the encoded audio signal a time-frequency resolution parameter for each of the plurality of frequency bands, the time-frequency resolution parameter indicating a resolution for a current band; and   decoding the plurality of spectral coefficients with a variable time-frequency resolution based at least in part on the time-frequency resolution parameters,   wherein and the audio processing apparatus is implemented at least in part with some hardware.   
     
     
         11 . The method of  claim 1  further comprising extracting a second time-frequency resolution parameter from the encoded audio signal, the second time-frequency resolution parameter indicating a degree or amount of resolution change. 
     
     
         12 . The method of  claim 1  further comprising extracting a level parameter for each of the plurality of frequency bands, wherein the level parameters represent a spectral profile of the audio signal. 
     
     
         13 . The method of  claim 1  wherein the plurality of frequency bands are critical bands approximating the Bark scale. 
     
     
         14 . The method of  claim 1  wherein the frame of spectral coefficients were generated by overlapping successive windows. 
     
     
         15 . The method of  claim 11  further comprising extracting a transient parameter indicating whether a frame of audio contains a transient. 
     
     
         16 . The method of  claim 15  wherein both the transient parameter and the second time-frequency resolution parameter are used to determine the degree or amount of resolution change. 
     
     
         17 . The method of  claim 10  wherein a temporal resolution is increased for bands including at least one transient. 
     
     
         18 . The method of  claim 10  wherein a frequency resolution is increased for bands including at least one tone. 
     
     
         19 . The method of  claim 10  wherein at least some of the time-frequency resolution parameters indicate a change in resolution for the current band from a previous band. 
     
     
         20 . An audio processing apparatus for decoding an audio signal encoded with variable time-frequency resolution, the apparatus comprising:
 an input interface for receiving the encoded audio signal;   an extractor for extracting from the encoded audio signal a frame of spectral coefficients;   a separator for dividing the frame of spectral coefficients into a plurality of frequency bands;   an extractor for extracting from the encoded audio signal a time-frequency resolution parameter for each of the plurality of frequency bands, the time-frequency resolution parameter indicating a resolution for a current band; and   a decoder for decoding the plurality of spectral coefficients with a variable time-frequency resolution based at least in part on the time-frequency resolution parameters,   wherein the audio processing apparatus is implemented at least in part with some hardware.