Implementation and optimization of parametric stereo encoding in enhanced aacPlus encoder
The state-of-the-art lowbitrate audio coder is enhanced aacPlus, which is a combination of Parametric Stereo (PS), Spectral Band Replication (SBR) and Advanced Audio Coding (AAC). PS as the newest addition to the coder makes it possible to encode the audio at a bitrate of as low as 24 kbps with acce...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Theses and Dissertations |
Published: |
2008
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/3521 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Summary: | The state-of-the-art lowbitrate audio coder is enhanced aacPlus, which is a combination of Parametric Stereo (PS), Spectral Band Replication (SBR) and Advanced Audio Coding (AAC). PS as the newest addition to the coder makes it possible to encode the audio at a bitrate of as low as 24 kbps with acceptable audio quality. The idea behind PS is to code stereo audio as a monaural downmix signal and a small amount of spatial parameters which describe its spatial image. The monaural signal can then be encoded by any generic audio coder while the spatial parameters are embedded into the resulting mono audio bitstream. The details of PS encoding as well as a general overview of the decoding process are presented in this thesis. An implementation of PS encoder which supports the full MPEG-4 PS configurations is presented. Along with the implementation, two optimizations are proposed: enhanced downmixing scheme and unified transient detector. A subjective listening test to evaluate both optimizations reveals that the optimized encoder is able to perform as well as the reference encoder with a total saving of 4% of the computational complexity. In addition, a concept of an objective method to evaluate spatial image distortion due to audio processing is proposed. The evaluation of the proposed method reveals that the output metrics defined are able to approximate the simulated spatial distortion. |
---|