1 / 12

LNS Arithmetic for MPEG Encoding Using a Fast DCT

LNS Arithmetic for MPEG Encoding Using a Fast DCT. Jie Ruan and Mark G. Arnold. LNS (Logarithmic Number System). Represents a number by a sign bit and an exponent to a certain base (b). S. Exponent (n-1 bits). F (Precision). Properties of LNS. Large dynamic range

albert
Download Presentation

LNS Arithmetic for MPEG Encoding Using a Fast DCT

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LNS Arithmetic for MPEG Encoding Using a Fast DCT Jie Ruan and Mark G. Arnold

  2. LNS (Logarithmic Number System) • Represents a number by a sign bit and an exponent to a certain base (b) S Exponent (n-1 bits) F (Precision)

  3. Properties of LNS • Large dynamic range • Easy for multiplications, divisions and exponentiations • Additions are not linear operations for LNS • Cost of adders is exponential to word length • Has advantages at low precisions.

  4. Multiplication Logb(XY)=logbX+logbY The cost is a fixed-point adder Addition More complex process than multiplication E.g., when calculating logb(X+Y), (x=logbX, y=logbY) Calculate z=x-y Table-lookup sb(z)=logb(1+2z) Logb(X+Y)=y+sb(z) LNS Arithmetic Units

  5. Major parts of MPEG Encoding • RGBYUV • Motion Estimation • Discrete Cosine Transform (DCT) • Variable Length Coding

  6. 2-D Discrete Cosine Transform 2 Dimensional 8x8 DCT

  7. Berkeley MPEG Encoder • Berkeley MPEG Tool is a software to perform MPEG encoding and decoding, written in C language. • The DCT function is modified to accommodate LNS features.

  8. Visual Effect of MPEG Encoding with LNS DCT • At F=4, a visually acceptable precision 10-bit LNS 16-bit Fixed-point 64-bit Floating-point

  9. Numerical results (SNR)

  10. Hardware implementation for DCT • NxN 2 Dimensional DCT • Usually performed through 2 rounds of 1-D N-point DCT, with an NxN buffer • Less hardware and simpler data input arrangement than direct 2-D implementations • Chen’s 1-D DCT algorithm • Directly factorizes the DCT matrix • 16 multiplications • 26 additions

  11. Compares LNS DCT with Fixed-point DCT Comparison Hardware cost for Chen’s algorithm, at F=4

  12. Conclusion • At low precisions LNS is a good way for DCT in MPEG, compared to Fixed-point • shorter word length • less area • Comparable visual result • LNS is a good alternative for MPEG encoding

More Related