Processing audio signals with adaptive time or frequency resolution

摘要

In one aspect, an audio processing apparatus is disclosed. The apparatus includes an audio decoder, a filterbank, and a processor. The audio decoder decodes an encoded audio signal to obtain a time-domain audio signal, the encoded audio signal including a plurality of spectral components. The filterbank splits the time-domain audio signal to obtain a plurality of complex-valued subband samples in a first frequency region. The processor generates a plurality of subband samples in a second frequency region based at least in part on the complex-valued subband samples in the first frequency region, adaptively groups at least some of the plurality of subband samples in the second frequency region with an adaptive time resolution or an adaptive frequency resolution, and determines a spectral profile of at least some of the subband samples in the second frequency region based on the groups.

主权项

1. An audio processing apparatus comprising:
an audio decoder that decodes an encoded audio signal to obtain a time-domain audio signal, the encoded audio signal including a plurality of spectral components from at least two channels of audio content; a filterbank that splits the time-domain audio signal to obtain a plurality of complex-valued subband samples in a first frequency region for each of the at least two channels of audio content; and one or more processors that for each of the at least two channels of audio content:
generate a plurality of subband samples in a second frequency region based at least in part on the complex-valued subband samples in the first frequency region,group at least some of the plurality of subband samples in the second frequency region with an adaptive time resolution and an adaptive frequency resolution to obtain an adaptive grouping, anddetermine a spectral profile of at least some of the subband samples in the second frequency region based at least in part on the adaptive grouping,wherein at least one of the audio decoder, the filterbank, and the one or more processors are implemented in hardware, andwherein a parameter in the encoded audio signal indicates the adaptive frequency resolution for each of the at least two channels of audio content by specifying either a first frequency resolution or a second frequency resolution for each of the at least two channels of audio content and wherein the first frequency resolution is finer than that second frequency resolution.