Low power and cost effective VLSI design for an MP3 audio decoder using an optimized synthesissubban

1
Low power and cost effective VLSI design for an
MP3 audio decoder using an optimized
synthesis-subband approach

• T.-H. Tsai and Y.-C. Yang
• Department of Electrical Engineering and National
  Central University, Taiwan ROC
• IEE Proceedings on Computers and Digital
  Techniques

2
Abstract

• An optimized approach to MPEG layer-3(MP3) audio
  decoding is presented, with the main theme
  focused on the synthesis subband. Since the
  synthesis subband is the most power-consuming
  component in decoding, a cost-effective
  architecture is proposed based on a system-design
  consideration. By means of an algorithm and
  architecture, the synthesis subband archives a
  high throughput with reduced memory requirements
  and hardware complexity. With a two-stage
  pipeline architecture, it allows 100 hardware
  utilization and is suitable for low-power
  implementation. In addition, the chip design in a
  0.35um process is also accomplished. It occupies
  a die area of about 2.7 3.2 mm2 with a
  transistor count of 157,469 and a low-power
  dissipation of only 2.92mW

3
Whats the problem

• MPEG layer-3(MP3) coding has been widely applied
  to current digital audio broadcasting and
  multimedia application
• A cost-effective and low-power implementation
  will largely reduce the hardware and computation
  complexity
• From the MP3 decoder point of view, the
  computational load depends on the realization of
  a synthesis subband

4
Outline

• Introduction of synthesis subband
• Implementation considerations and analysis
• Proposed method and architecture
• Results and comparison
• Conclusion

5
Introduction(1)

• Elementary concept of MP3
• Multirate subband-based coding techniques
• In the encoder, it performs analysis subband
  filtering with 32 equally spaced filterbanks
  based on a psychoacoustical model
• In the decoder, it performs synthesis subband
  filtering
• Most fast algorithms techniques interpret
  synthesis subband filtering as a modified
  discrete cosine transform (MDCT) with some
  additional windowing operations

6
Introduction(2)

• One of the popular method
• Translate DCT into a FFT kernel
• AdvantageBecause of FFT equations specific
  symmetric and recursive property, we can reduce
  the number of multiplications and additions
• Disadvantagethese methods have complex control
  and irregular data flow which will introduce a
  high hardware cost
• The proposed design
• reduced memory requirements and hardware
  complexity
• High efficiency with 100 hardware utilization
  using a two-stage pipeline architecture

7
Introduction(3)

• MP3 decoding flow
• Hybrid filter bank divided into inverse modified
  discrete cosine transfer with dynamic windowing
  and overlap (DWIMDCT), and the synthesis subband
  filterbank

8
Introduction(4)

• Synthesis-subband decoding flow

9
Implementation analysis

• Design target
• Delivering the required high performance at the
  minimum cost and the smallest silicon area
• The performance is determined by real-time
  constraints

10
Implementation analysis (cont.)

• MOPS Fs SC N
• FsSample frequency
• SCTotal number of numerical calculations per
  sample
• Nnumber of audio channel

11
Implementation consideration

• In synthesis subband, IMDCT can be broken into an
  FFT, a data shift, preprocessing and
  post-processing
• Three considerations
• The initial transformer, the real-number
  computation is also translated into the complex
  number computation
• Data shift, preprocessing and post-processing
  still contain complex multiplications
• FFT algorithms always need many multipliers, and
  the butterfly recursive process leads to some
  complex interconnection and routing

12
Proposed method

• Normal IMDCT
• Proposed IMDCT
• Require about ¼ amount of multiplier-accumulate
  computations
• Required size for the ram buffer can be reduced
  to only 512 words per channel(½ amount of
  original)

13
Architecture
IPQMF
IMDCT
14
Architecture (cont.)

• Pipeline architecture

15
Memory configuration (1)
16
Memory configuration (2)

• Data conflicts in IMDCT and IPQMF

17
Memory configuration (3)

• Memory data access with pipeline operation

18
Results and comparison (1)
19
Results and comparison (2)
20
Results and comparison (3)
21
Conclusion

• By means of novel algorithm and architecture, the
  synthesis subband has a better performance
• It also archives a high throughput, with a
  low-cost memory requirement and hardware
  complexity

22
(No Transcript)