Embedded audio coding (EAC) with implicit auditory masking

  • Authors:
  • Jin Li

  • Affiliations:
  • Microsoft Research, Redmond, WA

  • Venue:
  • Proceedings of the tenth ACM international conference on Multimedia
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

An embedded audio coder (EAC) is proposed with compression performance rivals the best available non-scalable audio coder. The key technology that empowers the EAC with high performance is the implicit auditory masking. Unlike the common practice, where an auditory masking threshold is derived from the input audio signal, transmitted to the decoder and used to quantize (modify) the transform coefficients; the EAC integrates the auditory masking process into the embedded entropy coding. The auditory masking threshold is derived from the encoded coefficients and used to change the order of coding. There is no need to store or send the auditory masking threshold in the EAC. By eliminating the overhead of the auditory mask, EAC greatly improves the compression efficiency, especially at low bitrate. Extensive experimental results demonstrate that the EAC coder substantially outperforms existing scalable audio coders and audio compression standards (MP3 and MPEG-4), and rivals the best available commercial audio coder. Yet the EAC compressed bitstream is fully scalable, in term of the coding bitrate, number of audio channels and audio sampling rate.