Computing FFT Twiddle Factors
Rick Lyons gives two compact algorithms to compute individual twiddle factors for radix-2 DIF and DIT FFTs, handy when you need only a subset of outputs such as in pruned FFTs. He explains stage indexing, provides closed-form formulas including the bit-reversal step for DIT, and walks through N=8 examples so you can implement the twiddle-angle calculations directly.
Knowledge Mine for Embedded Systems
A little-known interactive portal makes learning embedded systems surprisingly practical and visual. The site is organized into four main areas: embedded systems design, design lifecycle, design methods, and design tools. Each section uses clickable system block diagrams so you can jump from a block, for example a MAC unit, to a focused page with detailed explanations. It’s a handy, ready reference for DSP and embedded engineers.
Hidden Linear Algebra in DSP
Linear algebra is hiding in plain sight inside many DSP techniques, not just abstract theory. By treating linear systems as matrix operators y = A x you reveal Toeplitz structure in LTI systems, connect to covariance matrices, and gain geometric intuition via eigenvalues and eigenvectors. This matrix viewpoint complements convolution-based thinking and offers practical tools for filter and channel analysis.
Accelerating Matlab DSP Code on the GPU
Seth Benton spent a few days testing Jacket to accelerate MATLAB on NVIDIA GPUs, and found it surprisingly easy to speed up DSP code. He ran 2D FFT and interp2 benchmarks on a MacBook Air with a GeForce 9400M, seeing impressive speedups for large images while hitting GPU memory and precision limits at high sizes. The post shares practical tips on casting to GPU types, minimizing CPU-GPU transfers, and when GPU acceleration is most useful.
Googling: a now-required skill
How many times has Google saved you? If you're a dsp programmer, I'll bet A LOT. These days, there are simply so many answers out there (and a bigillion more added daily), that for any given problem, the solution is out there. Or at least information to point you in the right direction.
I won't claim being an expert. There are other blogs for that. But I thought I'd share a few dsp-related insights that have helped me out immensely.
-
It's not all...
OpenCV for DSP/GPU, MSDN equivalent for CCS, and more
Porting OpenCV to DSPs could be a real business opportunity, but it is far from trivial, writes Shehrzad Qureshi. He highlights major obstacles: the engineering scale, mixed open-source licenses, and hard-to-parallelize primitives like connected components. He also criticizes Code Composer Studio's help system compared with MSDN, notes an ATI Stream talk, and announces a CUDA walkthrough on FFT-based image filtering.
Computing an FFT of Complex-Valued Data Using a Real-Only FFT Algorithm
Rick Lyons shows a compact trick to get an N-point complex FFT using only real-input FFT routines by transforming the real and imaginary parts separately and recombining their outputs. The post presents a one-line recombination formula, Xc(m) = real[Xr(m)] - imag[Xi(m)] + j{imag[Xr(m)] + real[Xi(m)]}, and an algebraic derivation based on the two-real-in-one-complex FFT identity. Useful for systems that only provide real-only FFTs.
Random GPGPU Musings
Shehrzad Qureshi argues that general-purpose GPU computing is poised to reshape engineering workloads, and contrasts Nvidia's CUDA ecosystem with ATI's Stream and OpenCL. He points out that GPU architectures and programming models are similar across vendors, but Nvidia's head start in sample code and developer community gives CUDA a practical advantage. Read for a concise industry perspective on choosing a GPGPU platform.
GPGPU DSP
Shehrzad Qureshi kicks off his DSP blog by championing GPGPU, focusing on Nvidia's CUDA and real-product experience. He argues that with CPU clock speeds stalled, large-scale parallelism on GPUs is the practical path forward for many signal-processing tasks. The post traces GPGPU history from shader 'hacks' to modern APIs and previews future posts comparing CUDA vs OpenCL, Intel's Larrabee, and Nvidia Fermi.
Some Thoughts on a German Mathematician
Rick Lyons revisits the remarkable career of Carl Friedrich Gauss, mixing memorable anecdotes with technical highlights. The post links Gauss’s work on the Gaussian curve, complex-plane representation, orbit prediction, and early telegraph experiments to ideas familiar to DSP engineers, and notes historical evidence that he developed trigonometric series before Fourier. It’s a short, engaging reminder of Gauss’s broad influence.
A Table of Digital Frequency Notation
Rick Lyons compiles a compact, practical table that untangles the many algebraic frequency notations used in DSP. The reference lines up continuous and discrete sinusoid forms, shows the frequency variable names and units, and lists valid ranges and conversions like Ω = 2πf and normalized forms with fs. A printable PDF of the table is available for easy desk reference.
A Useful Source of Signal Processing Information
I just discovered a useful web-based source of signal processing information that was new to me. I thought I'd share what I learned with the subscribers here on DSPRelated.com.
The Home page of the web site that I found doesn't look at all like it would be useful to us DSP fanatics. But if you enter some signal processing topic of interest, say, "FM demodulation" (without the quotation marks) into the 'Search' box at the top of the web page
and click the red 'SEARCH...
Exact Near Instantaneous Frequency Formulas Best at Peaks (Part 1)
Cedron Dawg presents a new family of exact time-domain formulas to estimate the instantaneous frequency of a single pure tone. The methods generalize a known one-sample formula into k-degree neighbor-pair sums with spacing d, giving exact results in the noiseless case and tunable robustness in noise. The paper explains why real-tone estimates must be taken at peaks and shows the formulas also work for complex tones.
Improved Three Bin Exact Frequency Formula for a Pure Real Tone in a DFT
Cedron Dawg extends his two-bin exact frequency formulas to a three-bin DFT estimator for a pure real tone, and presents the derivation in computational order for practical use. The method splits complex bin values into real and imaginary parts, forms vectors A, B, and C, applies a sqrt(2) variance rescaling, and computes frequency via a projection-based closed form. Numerical tests compare the new formula to prior work and show improved accuracy when the tone lies between bins.
Orfanidis Textbooks are Available Online
Two classic signal processing textbooks by Sophocles J. Orfanidis are now available for download from his Rutgers webpages. The first, Introduction to Signal Processing, includes errata and a homework solutions manual. The second, Optimum Signal Processing, includes a solutions manual plus MATLAB, C and Fortran code. Note that Prof. Orfanidis retains copyright on both books, All Rights Reserved.
Do you like the new Comments System?
I have just finished implementing a new comments system for the blogs. Do you like it?
Please share your thoughts with me by adding a comment.
I'll wait a few days and make sure it works properly and then I'll port it to the code snippets and papers section.
Thanks!
DSPRelated Finally on Twitter!
After resisting social networks, Stephane Boucher announces DSPRelated's move to Twitter and a few site improvements. Users can now sign in once to access DSPRelated, FPGARelated and EmbeddedRelated with the same account, and the site will post updates from @dsprelated, @embeddedrelated and @fpgarelated. To encourage followers, Boucher will occasionally tweet links that award prizes to the first visitors.
Implementing Simultaneous Digital Differentiation, Hilbert Transformation, and Half-Band Filtering
Recently I've been thinking about digital differentiator and Hilbert transformer implementations and I've developed a processing scheme that may be of interest to the readers here on dsprelated.com.
50,000th Member Announced!
In my last post, I wrote that DSPRelated.com was about to reach the 50,000 members mark. Well, I am very happy to announce that it happened during the holidays, and the lucky person is Charlie Tsai from Taiwan. Charlie is an assistant professor in the Department of Electrical Engineering at the National Central University in Taiwan where he teaches the "Biomedical Signal Processing" class. He is also the advisor of the
Engineering the Statistics
Do you remember the probability course you took in undergrad? If you were like me, you would consider it one of those courses that you get out of confused. But maybe a time will come where you regret skipping class because of the lecturer's persisting attempts to scare you with mathematical involved nomenclature.As you might have guessed, I had this moment few months back where I had to go deep into statistical analysis. I learned things the hard way, or maybe it is the right way. I mean...
TI DSP Predictions
Jeff Brower lays out two bold predictions for Texas Instruments that could reshape the DSP developer ecosystem. He argues TI will offer a supported real-time Linux on their C6x DSPs now that legal obstacles have eased, and that TI may acquire an FPGA company to own the board space around its chips. Read to weigh the technical and strategic impact.
Improved Three Bin Exact Frequency Formula for a Pure Real Tone in a DFT
Cedron Dawg extends his two-bin exact frequency formulas to a three-bin DFT estimator for a pure real tone, and presents the derivation in computational order for practical use. The method splits complex bin values into real and imaginary parts, forms vectors A, B, and C, applies a sqrt(2) variance rescaling, and computes frequency via a projection-based closed form. Numerical tests compare the new formula to prior work and show improved accuracy when the tone lies between bins.
Update to a Narrow Bandpass Filter in Octave or Matlab
Paul Lovell presents an updated, compact Octave/Matlab implementation of a narrow bandpass FIR that runs about four times faster and uses float32 to cut processing cost. The design combines a single matrix IFIR stage with three moving-sum (RRS) stages per baseband, auto-calculates the IFIR expansion factor, and adds easier parameter setup plus WAV I/O and FFT plots. A TensorFlow Colab demo is also provided.
Exact Near Instantaneous Frequency Formulas Best at Peaks (Part 2)
Cedron Dawg derives a second family of exact time domain formulas for single-tone frequency estimation that trade a few extra calculations for improved noise robustness. Built from [1+cos]^k binomial weighting of neighbor-pair sums, the closed-form estimators are exact and are best evaluated at signal peaks for real tones, while complex tones do not share the zero-crossing limitation. Coefficients up to k=9 are provided.
Project update-2 : Digital Filter Blocks in MyHDL and their integration in pyFDA
This update shows a working integration between Pyfda and MyHDL using a compact API that passes fixed-point coefficients, stimulus data, and returns simulated filter responses. It walks through two usage styles, constructor-based and setter-method-based, and demonstrates a Pyfda workflow from specs to MyHDL simulation and plotting. Future plans include HDL code generation and API extension as filters grow.
Of Forests and Trees and DSP
Too often DSP engineers fixate on algorithms and miss the rest of the product. Tim Wescott uses the humble Korg CA-20 chromatic tuner to show that a great algorithm alone does not make a usable device, you also need good data acquisition, adequate processing, sensible precision, a usable UI, and appropriate casing and cost. The post gives practical do's and don'ts for system-level DSP design.
Engineering the Statistics
Do you remember the probability course you took in undergrad? If you were like me, you would consider it one of those courses that you get out of confused. But maybe a time will come where you regret skipping class because of the lecturer's persisting attempts to scare you with mathematical involved nomenclature.As you might have guessed, I had this moment few months back where I had to go deep into statistical analysis. I learned things the hard way, or maybe it is the right way. I mean...
Digging into an Audio Signal and the DSP Process Pipeline
In this post, I'll look at the benefits of using multiple perspectives when handling signals.A Pre-existing Audio File
Let's say we have an audio file of interest. Let's load it into Audacity and zoom in a little (using View → Zoom → Zoom In, multiple times). The figure illustrates the audio signal: just a basic single-tone signal.
By continuing to zoom into the signal, we eventually get to the point of seeing individual samples as illustrated below. Notice that I've marked one...
Why is Fourier transform broken
Many engineers know the Gibbs phenomenon without grasping its root cause. This post shows that the problem comes from using the incomplete metric space of continuous functions, C[a,b], for Fourier series, and explains how switching to Lp spaces resolves convergence in the mean but allows functions to differ on sets of measure zero. It also reminds readers that Fourier analysis gives no time localization, so be mindful of its limits.
State Space Representation and the State of Engineering Thinking
Most, if not all, textbooks in signal processing (SP) thoroughly covers the frequency analysis of signals and systems alike, including the Fourier and the Z-transform that produce the well known Transfer Function. Another way of signal analysis, not as popular in signal processing though, is State Space representation. State space models describes the internal signals of the system or the process and how it affect the output, in contrast to the frequency representation that only describe the...
