Speaker: Darcy Tsai Advisor: Prof. An-Yeu Wu Date: 2013/10/31

Slides:



Advertisements
Similar presentations
DFT & FFT Computation.
Advertisements

David Hansen and James Michelussi
Digital Kommunikationselektronik TNE027 Lecture 5 1 Fourier Transforms Discrete Fourier Transform (DFT) Algorithms Fast Fourier Transform (FFT) Algorithms.
DFT and FFT FFT is an algorithm to convert a time domain signal to DFT efficiently. FFT is not unique. Many algorithms are available. Each algorithm has.
1 Final project Speaker: Team 5 電機三 黃柏森 趙敏安 Mentor : 陳圓覺 Adviser: Prof. An-Yeu Wu Date: 2007/1/22.
Lecture #17 INTRODUCTION TO THE FAST FOURIER TRANSFORM ALGORITHM Department of Electrical and Computer Engineering Carnegie Mellon University Pittsburgh,
Sampling, Reconstruction, and Elementary Digital Filters R.C. Maher ECEN4002/5002 DSP Laboratory Spring 2002.
Introduction to Fast Fourier Transform (FFT) Algorithms R.C. Maher ECEN4002/5002 DSP Laboratory Spring 2003.
Introduction SYSC5603 (ELG6163) Digital Signal Processing Microprocessors, Software and Applications Miodrag Bolic.
Lecture #18 FAST FOURIER TRANSFORM INVERSES AND ALTERNATE IMPLEMENTATIONS Department of Electrical and Computer Engineering Carnegie Mellon University.
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU A New Algorithm to Compute the Discrete Cosine Transform VLSI Signal Processing 台灣大學電機系.
Fast Fourier Transforms
Discrete-Time and System (A Review)
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU Adaptive Signal Processing -Equalization- Ben Advisor: Prof. An-Yeu Wu 2008/01/22.
Processor Architecture Needed to handle FFT algoarithm M. Smith.
Under-Graduate Project Mid-Term Paper Reading Presentation Adviser: Prof. An-Yeu Wu Mentor: 詹承洲 第二組 溫仁揚 溫昌懌.
Student : Andrey Kuyel Supervised by Mony Orbach Spring 2011 Final Presentation High speed digital systems laboratory High-Throughput FFT Technion - Israel.
Jeff Wang Kay-Won Chang March 18, DEMO Harmonic Product Spectrum (HPS) pitch detection: obtain fundamental frequency from FFT Fast Fourier Transform.
Hossein Sameti Department of Computer Engineering Sharif University of Technology.
Fast Memory Addressing Scheme for Radix-4 FFT Implementation Presented by Cheng-Chien Wu, Master Student of CSIE,CCU 1 Author: Xin Xiao, Erdal Oruklu and.
Radix-2 2 Based Low Power Reconfigurable FFT Processor Presented by Cheng-Chien Wu, Master Student of CSIE,CCU 1 Author: Gin-Der Wu and Yi-Ming Liu Department.
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU Undergraduate Projects Speaker: Wes Adviser: Prof. An-Yeu Wu Date: 2015/09/22 Lab.
Digital Signal Processing Chapter 3 Discrete transforms.
Paper Reading - A New Approach to Pipeline FFT Processor Presenter:Chia-Hsin Chen, Yen-Chi Lee Mentor:Chenjo Instructor:Andy Wu.
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU Under-Graduate Project Improving Timing, Area, and Power Speaker: 黃乃珊 Adviser: Prof.
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU CORDIC (Coordinate rotation digital computer) Ref: Y. H. Hu, “CORDIC based VLSI architecture.
ACCESS IC LAB Graduate Institute of Electronics Engineering, NTU Under-Graduate Project Case Study: Single-path Delay Feedback FFT Speaker: Yu-Min.
Inverse DFT. Frequency to time domain Sometimes calculations are easier in the frequency domain then later convert the results back to the time domain.
Reconfigurable FFT architecture
DEPARTMENTT OF ECE TECHNICAL QUIZ-1 AY Sub Code/Name: EC6502/Principles of digital Signal Processing Topic: Unit 1 & Unit 3 Sem/year: V/III.
Copyright ©2010, ©1999, ©1989 by Pearson Education, Inc. All rights reserved. Discrete-Time Signal Processing, Third Edition Alan V. Oppenheim Ronald W.
Professor A G Constantinides 1 Discrete Fourier Transforms Consider finite duration signal Its z-tranform is Evaluate at points on z-plane as We can evaluate.
Under-Graduate Project Adviser: Prof. An-Yeu Wu Mentor: 詹承洲 第二組 溫仁揚 溫昌懌.
A New Class of High Performance FFTs Dr. J. Greg Nash Centar ( High Performance Embedded Computing (HPEC) Workshop.
Fast Fourier Transforms. 2 Discrete Fourier Transform The DFT pair was given as Baseline for computational complexity: –Each DFT coefficient requires.
The Discrete Fourier Transform
An FFT for Wireless Protocols Dr. J. Greg Nash Centar ( HAWAI'I INTERNATIONAL CONFERENCE ON SYSTEM SCIENCES Mobile.
FFT VLSI Implementation
EC1358 – DIGITAL SIGNAL PROCESSING
Eeng Chapter 2 Discrete Fourier Transform (DFT) Topics:  Discrete Fourier Transform. Using the DFT to Compute the Continuous Fourier Transform.
EE345S Real-Time Digital Signal Processing Lab Fall 2006 Lecture 17 Fast Fourier Transform Prof. Brian L. Evans Dept. of Electrical and Computer Engineering.
بسم الله الرحمن الرحيم Digital Signal Processing Lecture 14 FFT-Radix-2 Decimation in Frequency And Radix -4 Algorithm University of Khartoum Department.
Low Power Design for a 64 point FFT Processor
Chapter 9. Computation of Discrete Fourier Transform 9.1 Introduction 9.2 Decimation-in-Time Factorization 9.3 Decimation-in-Frequency Factorization 9.4.
CORDIC Based 64-Point Radix-2 FFT Processor
Discrete Fourier Transform
1 Paper reading A New Approach to FFT Processor Speaker: 吳紋浩 第六組 洪聖揚 吳紋浩 Adviser: Prof. Andy Wu Mentor: 陳圓覺.
Chapter 4 Discrete-Time Signals and transform
Final Project Report 64 points FFT
DIGITAL SIGNAL PROCESSING ELECTRONICS
Fixed-pointed FFT model
Fast Fourier Transforms Dr. Vinu Thomas
Real-time double buffer For hard real-time
A New Approach to Pipeline FFT Processor
Lecture #17 INTRODUCTION TO THE FAST FOURIER TRANSFORM ALGORITHM
4.1 DFT In practice the Fourier components of data are obtained by digital computation rather than by analog processing. The analog values have to be.
Fast Fourier Transformation (FFT)
Chapter 9 Computation of the Discrete Fourier Transform
Chapter 2 Discrete Fourier Transform (DFT)
C Model Sim (Fixed-Point) -A New Approach to Pipeline FFT Processor
Chapter 19 Fast Fourier Transform
Speaker: Yumin Adviser: Prof. An-Yeu Wu Date: 2013/10/24
95-1 Under-Graduate Project Fixed-point Analysis
FFT VLSI Implementation
95-1 Under-Graduate Project Paper Reading Presentation
Introduction SYSC5603 (ELG6163) Digital Signal Processing Microprocessors, Software and Applications Miodrag Bolic.
Introduction SYSC5603 (ELG6163) Digital Signal Processing Microprocessors, Software and Applications Miodrag Bolic.
Fast Fourier Transform (FFT) Algorithms
Speaker: Chris Chen Advisor: Prof. An-Yeu Wu Date: 2014/10/28
Lecture #17 INTRODUCTION TO THE FAST FOURIER TRANSFORM ALGORITHM
Presentation transcript:

Speaker: Darcy Tsai Advisor: Prof. An-Yeu Wu Date: 2013/10/31 102-1 Under-Graduate Project Case Study: Single-path Delay Feedback FFT Speaker: Darcy Tsai Advisor: Prof. An-Yeu Wu Date: 2013/10/31

Outline Introduction of Fast Fourier Transform (FFT) DFT/IDFT & FFT/IFFT Flow Graph of FFT Algorithm Hardware Implementation Radix-n FFT Algorithm System Design Flow Floating Point Modeling Fixed Point Modeling Simulation

DFT/IDFT Definition of Discrete Fourier Transform (DFT) and Inverse DFT (IDFT) DFT X[k] Frequency domain spectrum x[n] Time domain sequence IDFT Twiddle factor :

FFT/IFFT Fast Fourier Transform (FFT) is based on the concept of “Divide-and-Conquer” The complexity of DFT: N2 The complexity of FFT: Nlog2N Decimation-in-Time (DIT) FFT Algorithm —

Flow Graph of DIT FFT Algorithm Pre-processing Post-processing

Flow Graph of DIT FFT Algorithm Computation: Nlog2N N N N log2N stages

Flow Graph of DIT FFT Algorithm DFT-4 DFT-2 Bit-reverse order Normal order 000 100 010 110 001 101 011 111 000 001 010 011 100 101 110 111 [1]

Flow Graph of DIF FFT Algorithm DFT-2 DFT-4 Normal order Bit-reverse order 000 001 010 011 100 101 110 111 000 100 010 110 001 101 011 111 [1]

Hardware Implementation Fully Spread Reuse of Single Butterfly Slow  ———— Speed ————  Fast Small  ———— Area ————  Large Complex  ———— Control ————  Simple

Hardware Implementation [2]

Radix-4 FFT Algorithm Radix-4: decimation into 4 groups

Radix-2 Single-path Delay Feedback for N=16 [2]

Radix-n FFT Algorithm For Radix-n FFT, the complexity is NlognN Larger N — Less complex multiplier Less stages More complex butterfly structure Designing at algorithm level outperforms others Pipeline, Parallel, Retiming, Folding/Unfolding

Relationship of Radix-4 & Radix-22 BF4 BF2i BF2ii [2]

Radix-22 Single-path Delay Feedback for N=256 BF2i 1 Xr(n) Xi(n) Xr(n+N/2) Xi(n+N/2) Zr(n) Zi(n) Zr(n+N/2) Zi(n+N/2) - Xi(n+N/2) Xr(n+N/2) t s BF2ii 1 Xr(n) Xi(n) Zr(n) Zi(n) Zr(n+N/2) Zi(n+N/2) - ± [2]

System Design Flow Physical Model MATLAB Floating Point Model Fixed Point Model Optimize Simulation MATLAB Verilog Verification

Floating Point Model Implemented with MATLAB / C code Translate physical structure to high level language Keep original signal flow intact

Floating Point Model -j -j Butterfly(16) Butterfly(8) Butterfly(4)

Fixed Point Model of FFT Simulate truncation due to limited word-length Dynamic range of input is critical Ex: Only 3-bit of fractional part 1.422(10)  1.422(10) (floating point) 1.422(10)  1.011011(2) = 1 + 2-2+ 2-3 = 1.375 Input signal are truncated to limited precision Apply truncation where arithmetic is applied after the multiplier module Twiddle factors are also truncated before introduced to multiplier Fixed Point Model of FFT

Fixed Point Model of FFT -j -j

Simulation Parameterize the word-lengths of input Integer word-length Fractional word-length Twiddle factor word-length Insert randomly generated floating point input Compare with floating point result from MATLAB (SQNR computing)

Calculation of SQNR SQNR: Signal-to-Quantization-Noise Ratio

Optimal set: 2+6 = 8 Integer 2 bits Fractional 6 bits Fixed twiddle

Optimal set: 9+2 = 11 Integer 2 bits Twiddle 9 bits Fix Fractional

Optimal set: 9+7 = 15 Twiddle 9 bits Fractional 7 bits Fix Integer

Verification Word-lengths chosen: Integer 2 bits Fractional 7 bits Twiddle 9 bits Run multiple random tests (105 times) to ensure we have desired results Adjust bit lengths to ensure the SQNR ≧ 50 if necessary

Fractional 7 bits, Twiddle 9+1 bits

References [1] Alan V.Oppenheim, Ronald W. Schafer, “Discrete-time signal processing” 2nd edition. [2] E.H. Wold and A.M. Despain. “Pipelined and parallel-pipelined FFT processors for VLSI implementation.,” IEEE Trans. Comput., May 1984 [3] Shousheng He and Torkelson, M., “A new approach to pipeline FFT processor,” Proceedings of IPPS '96, 15-19 April 1996, pp766 –770.