This project covers the development of a self-contained 'system-on-a-chip' (SoC) design which allows the lossy compression of digital audio data. Primarily, this is achieved by the creation of a general purpose extensible SoC framework, based around an off-the-shelf central processing unit (CPU) core. The framework allows extension of the CPU by adding custom instructions and data processors which are supported by a collection of customisable fractional arithmetic units. The goal of this design is to allow easy and rapid exploration of the hardware design space when running and accelerating the open-source 'Ogg Vorbis' audio encoding algorithm. By creating custom acceleration hardware using the framework, a speed increase of around 33%, compared to an unmodified refence encoder, is achieved in an FPGA prototype implementation. This project is the only work we are aware of so far that considers the use of 'Ogg Vorbis' for encoding in an embedded system.
[1]
Wayne Luk,et al.
Autonomous Memory Block for reconfigurable computing
,
2004,
2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512).
[2]
Srivaths Ravi,et al.
Custom-instruction synthesis for extensible-processor platforms
,
2004,
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.
[3]
Kenzo Akagiri,et al.
ATRAC: Adaptive Transform Acoustic Coding for MiniDisc
,
1992
.
[4]
Young-Cheol Park,et al.
Low power MPEG/audio encoders using simplified psychoacoustic model and fast bit allocation
,
2001,
IEEE Trans. Consumer Electron..