DSPRelated.com

Yaakov Stein (@Y(J)S)


Re: Design FIR filters using DRHS-OLL algorithm

Reply posted 3 years ago (08/26/2021)
Filter design is an optimization problem, and as such many optimization methods may be applied.HS-based optimization is yet another attempt at producing an optimization...

Re: Sub-band processing

Reply posted 3 years ago (08/05/2021)
This is called mel scale SBC, and is used in many codecs.The easiest way to implement it is using a tree structure. First divide your signal into 2 half-bands that...

Re: CRC field in the middle of a packet

Reply posted 3 years ago (07/28/2021)
The best reference I know for this is "Reversing CRC – Theory and Practice."You need to compute a regular CRC up to the insertion point, and a "backwards" CRC...
Why do you want to compute an FFT? Do you really need the entire spectrum?If you only need the spectrum in a known area you can perform a zoom-FFT which has much...

Re: DTFT optimization question

Reply posted 3 years ago (06/04/2021)
MA filters can be formally written in ARMA format, but that doesn't make them full ARMA. I hope that you agree that the simple causal 3-point average with gain...

Re: DTFT optimization question

Reply posted 3 years ago (06/03/2021)
Goertzel isn't an IIR filter. It is a recusive method of evaluating a polynomial in W_N. Polynomial evaluation is FIR (the coefficients - the powers of W_N are...

Re: DTFT optimization question

Reply posted 3 years ago (06/03/2021)
If I understood correctly you are given 10 frequencies and want to calculate the (complex?) FT of a signal of length 8192. I am not sure what "break the signal into...

Re: Off by one windowing

Reply posted 3 years ago (05/05/2021)
What you say makes sense, but as I said in my previous reply, for any reasonable N the difference in straightforward usage is completely negligible (for any window...

Re: Off by one windowing

Reply posted 3 years ago (05/04/2021)
OK, now I understand what you are saying. You assumed that there is overlap between the last point and the first point, so that there are not 2 zeroed-out points...

Re: Off by one windowing

Reply posted 3 years ago (05/04/2021)
Triangle windows are always symmetric around their midpoint (that is the reason they are used).So for n=3 0 1 0 and for n=4 0 2/3 2/3 0 (or 1/2 1 1/2 and 1/4 3/4...
It is easiest to think of this in terms of the basic LPC model for speech.Consonants come in pairs which share the same AR synthesis filter (the poles of which are...

Re: SDR reception of legacy AM, FM waves

Reply posted 3 years ago (04/19/2021)
An SDR would not usually do envelope detection at all - that is a primitive method used because of limitations of analog electronics.If the input is IQ then the...

Re: SDR reception of legacy AM, FM waves

Reply posted 3 years ago (04/18/2021)
I'll try to give the simplest possible answer.Modulation means changing the parameters of a signal in order to carry information. For example, in audio AM one starts...

Re: Counteracting Beamforming Filter

Reply posted 4 years ago (03/30/2021)
HenryThanks for confirming that this is a microphone array (rather than a loudspeaker array). I understand that it is only 2 microphones (I am used to a much larger...

Re: Counteracting Beamforming Filter

Reply posted 4 years ago (03/30/2021)
How far away from the array are you from the microphone array? (I am assuming that this is a microphone array and by 0 degrees you mean perpendicular, but it really...

Re: Band-variable signals filtration - best approach

Reply posted 4 years ago (02/12/2021)
Sorry for not following you. Filtering out means to remove a signal. So you want to filter out the red signal and leave all the others? And then filter out the green...

Re: Band-variable signals filtration - best approach

Reply posted 4 years ago (02/04/2021)
No, I don't understand. If the signal of interest is in the bandpass then it can be demodulated. If the interfering signals are also in the bandpass, then there...

Re: Band-variable signals filtration - best approach

Reply posted 4 years ago (02/04/2021)
If the signal of interest is narrowband then why are you notching out the interferers? You should be bandpassing the signal of interest. That is indeed best done...

Re: Band-variable signals filtration - best approach

Reply posted 4 years ago (02/04/2021)
Yes, I am talking about filtering in the frequency domain which involves FFT, multiplying frequency by frequency, and the iFFTing back into the time domain. (You...

Re: Band-variable signals filtration - best approach

Reply posted 4 years ago (02/04/2021)
Have you considered filtering in the frequency domain using overlap-add or overlap-save?Since you simply zero out the bins you don't want, this will be one operation...
I don't know this book, but the answers to your questions are the very heart of DSP.When sampling an analog signal in the time domain the continuous function s(t)...

Re: Clock handling in multi rate systems

Reply posted 4 years ago (09/25/2020)
The standard reference is "Multirate DSP" by Crochiere and Rabiner.There are several implementation philosophies depending on the details. In easy cases (as mentioned...

Re: Simultaneous Static and Dynamic Acclerometer

Reply posted 4 years ago (09/21/2020)
If the vibration has a strong resonant frequency then use a notch filter. If it is white (i.e., noisy) then you have to consider the frequency content of the signal...

Re: Simultaneous Static and Dynamic Acclerometer

Reply posted 4 years ago (09/18/2020)
Are you interested in only one angle (e.g., the angle to up)? How much delay can you tolerate?The obvious answer to you question would be to derive the noisy angle...
You use a notch when you don't want to lose too much of the background signal, only that part of the spectrum which is being wiped out by the interfering tone.You...
No, notch filters needn't be IIR, although in many situations IIR will result in less computation and less delay, at the expense of phase distortion.

Re: Proof of Correlation Theorem for DFT

Reply posted 4 years ago (06/10/2020)
Is this a cyclic convolution? (Probably yes, since this is DSP...) If so, add N to any negative indexes to make them from 0 to N-1. In particular, if m goes from...

Re: Upward Compressor Soft Knee Equation

Reply posted 4 years ago (06/08/2020)
almostthe middle equation should have + (1/R - 1)here's my MATLAB :x = 0:0.001:1 ;T = 0.5 ;  % kick-in point (sample 501)W = 0.2 ;  % soft knee width (10 samples)W2...

Re: Upward Compressor Soft Knee Equation

Reply posted 4 years ago (06/08/2020)
Use almost the same equations. Just reverse the first and last ranges (so that the last range has slope 1) and in the middle range reverse the sign of the 2nd term...
It depends on what you are assuming about your signal. Remember that the phase signal is only half the picture (and if by \phi(t) with a small "p" you mean the...

Re: New IEEE Signal Processing Society Journal

Reply posted 5 years ago (10/31/2019)
Of course it depends what you want to achieve.These open access journals are careful with their processes, and are frequently produced by the same organizations...

Re: New IEEE Signal Processing Society Journal

Reply posted 5 years ago (10/31/2019)
Yes, with few exceptions only peer-reviewed journals count in academia (the major exception being patents).What you are specifically missing is 1) an ISBN number...

Re: New IEEE Signal Processing Society Journal

Reply posted 5 years ago (10/31/2019)
Proceedings of the National Academy of Sciences.

Re: New IEEE Signal Processing Society Journal

Reply posted 5 years ago (10/29/2019)
While not as much as some open journal charges (PNAS charges $1700) this is still quite a high charge.I review for 2 journals, and get paid nothing for my efforts...
To put this paper into perspective, multiplying a general N^2 matrix by an N-dimensional vector requires N^2 multiplications (N times a row of the matrix times the...

Re: channel capacity and data rate

Reply posted 5 years ago (10/14/2019)
The assumptions are that the noise is additive, the spectrum of the signal and noise are flat over the bandwidth, and that there is no correlation between the...

Re: OFF TOPIC: A Question About PI

Reply posted 5 years ago (07/08/2019)
A real number that contains any finite sequence of digits appears at least once is called a "rich number".There are real numbers that have been proven to be rich....

Re: getting started on dereverberation

Reply posted 5 years ago (04/30/2019)
Google for:"Fifty Years of Reverberation Reduction - From analog signal processing to machine learning" Emanuel Habets,  AES 60th Conference on DREAMS.Y(J)S

Re: Phase Locked Loop Books in a time of DSP

Reply posted 6 years ago (01/12/2019)
Actually, it was published 19 years ago and is already sold out (except for the India and China versions).It amazes me how much I would change if I would do a 2nd...

Re: Phase Locked Loop Books in a time of DSP

Reply posted 6 years ago (01/09/2019)
DSPCSP (Digital Signal Processing, a Computer Science Perspective - Wiley 2000)

Re: Phase Locked Loop Books in a time of DSP

Reply posted 6 years ago (01/09/2019)
There is a huge literature base on this. Search for "digital PLL" or more specifically "digital PLL symbol rate recovery".A good book is "Digital Clocks for Synchronization...
Wow - that's a lot of questions.I'm not going to try to answer them all - I would suggest you read any of the good DSP text books for that. But I'll try to point...

Re: Fixed Point vs Floating Point

Reply posted 6 years ago (11/30/2018)
DSP signals come in all sizes, and the bit rate needs to be determined based on what you want to do with the signal.Linearly sampled speech signals sound noisy if...
If you have enough computational power to find an autocorrelation lag, then use Pisarenko Harmonic Decomposition (or any of several other eigenspace algorithms...

Re: PHY, L1, Baseband

Reply posted 6 years ago (10/26/2018)
The term "PHY" is frequently used in Ethernet to distinguish the physical layer of Ethernet from the MAC layer (since Ethernet straddles both layer 1 and layer 2...
One must be even more careful in using pseudorandom sources, as they are deterministic and not random at all. So, of course, it depends on the application.If the...
As mentioned by dsplib, you can generate a Gaussianly distributed random sequence using a uniform pseudo-random generator using Box-Muller or related methods. If...

Re: CMSIS CFFT bit reversal table -how to generate?

Reply posted 6 years ago (09/17/2018)
OK, I see. These are mixed-radix algorithms. Note that the powers of 8 (64, 512, 4K) are pure radix 8, while the powers of 2 that are not powers of 8 are mixed radix....

Re: CMSIS CFFT bit reversal table -how to generate?

Reply posted 6 years ago (09/16/2018)
I don't know what CMSIS is, but for a radix-4 FFT you need to reverse pairs of bits instead of bits. In other words, if for radix 2 abcd changed to dcba, for radix-4...
Many years ago I purchased a copy of "Bentley's Second Phrase Code".This book is a compilation of 5-letter "words" that represent partial or full sentences. Thus,...
From a quick look, it seems to be an in-place FIR.In a standard FIR you convolve on x and produce y. If you try do this in-place you overwrite a value (say x_i)...

Re: 1D MUSIC DOA.

Reply posted 6 years ago (07/12/2018)
It is not physics, it is simply geometry.The phase differences are completely determined by the distances from the emitter to the antennas.The squared distance is...

Re: 1D MUSIC DOA.

Reply posted 6 years ago (07/11/2018)
That came out more complex than it needed to be.Easier to say it this way - you will obtain a y coordinate that is the square root of the sum of the squares of the...

Re: 1D MUSIC DOA.

Reply posted 6 years ago (07/11/2018)
No, you will not obtain the projection onto the 2d plane, since the vertical distance adds an addition time component that contributes to the phase differences.The...

Re: 1D MUSIC DOA.

Reply posted 6 years ago (07/10/2018)
This ambiguity a well-known problem of all regular structures. Circular arrays have similar phase ambiguity, but there power measurement can help.Have a look at...

Re: extracting ultra accurate phase with FFT

Reply posted 6 years ago (06/13/2018)
It's very easy. Assuming that you want a PLL and not an FLL (frequency locked loop) you run an oscillator (called the NCO) of the correct frequency, deduce the phase...

Re: extracting ultra accurate phase with FFT

Reply posted 6 years ago (06/11/2018)
I meant a PLL in software (no additional circuitry).It is much less work than an FFT and much better at finding phase of a sine in noncorrelated noise (that is what...

Re: extracting ultra accurate phase with FFT

Reply posted 6 years ago (06/11/2018)
Why not use a PLL ? Y(J)S

Re: LPC Formant-Shifting

Reply posted 6 years ago (04/19/2018)
You are correct in what has to be done:1. Root the LPC denominator polynomial (using regula falsi or Bairstow or Halley). 2. Modify the pole angles and recompute...

Re: FFT complexity

Reply posted 7 years ago (12/13/2017)
I never heard the name Karatsuba, but it is one of those numeric tricks that old timers in numeric programming know about, but I have rarely seen mentioned in the...

Re: FFT complexity

Reply posted 7 years ago (12/13/2017)
The radix 2 DIT or DIF FFT has log_2(N) stages of N/2 butterflies each.Each butterfly contains one complex multiply with the W_N^nk term (the twiddle factor), thus...

Re: Correlation of FSK modulated PRBS - Confused!

Reply posted 7 years ago (12/05/2017)
I agree. If you are normalizing to 1 when the lag is zero, then the side lobes should not be 1/2. Now that I read your question more carefully, I understand that...

Re: Correlation of FSK modulated PRBS - Confused!

Reply posted 7 years ago (12/04/2017)
Your normalization is not clear.When you are shifted by an integer number of chips, then the autocorrelation would be zero if the PRBS had precisely the same number...

Re: fft/ifft scaling revisited

Reply posted 7 years ago (11/28/2017)
The requirement is for iDFT(DFT(x)) = x (without rescaling).Work this through and you will see that the product of the constants before the sums needs to be 1/N....

Re: reconstruct phase-shifted sine tone

Reply posted 7 years ago (11/21/2017)
Just to make sure I understand the question ...You are taking a real sinusoid sin( 2 pi 20/8000 n) and multiplying by a complex phase shifter e^-i pi/2. How are...

Re: Upsampling of data

Reply posted 7 years ago (11/21/2017)
For graphics purposes linear (or more generally polynomial or spline) interpolation is fine.However, it is a really BAD idea to use polynomial interpolation for...

Re: Audio speed changer without changing the pitch

Reply posted 7 years ago (11/19/2017)
These methods don't change the sampling rate, and thus don't change the pitch. One family of methods exploits the stationarity of speech over short time periods....

Re: Digital Modulation Methods

Reply posted 7 years ago (10/22/2017)
Your are asking for the constellation diagram. There are standards, but the important issue is to use a Gray code in order for small errors to only flip one bit....
It must it be possible to go back and forth between the original waveform and the representation, otherwise the question is meaningless. If there is no requirement...

Re: Floating Point to Fixed Point Conversion

Reply posted 7 years ago (10/12/2017)
Understood.You must decide whether you want to go for variable scaling (i.e., at each butterfly check if rescaling is required) or constant scaling (which can be...

Re: Floating Point to Fixed Point Conversion

Reply posted 7 years ago (10/06/2017)
You are right that FFT/iFFT need special care in order to avoid overflow without ruining SNR. This topic has been discussed here before - try to find the previous...

Re: Gapped-signals

Reply posted 7 years ago (10/04/2017)
Your comment puzzles me.All of the methods estimate the spectral components in one way or another, however, the eigenspace methods directly find the frequencies,...

Re: Gapped-signals

Reply posted 7 years ago (10/01/2017)
Interpolation and extrapolation of signals are two of the 3 reasons for which linear filters were invented (see the seminal work by Wiener: "Extrapolation, Interpolation...

Re: Generating Random Numbers Through Audio

Reply posted 7 years ago (09/24/2017)
...

Re: Generating Random Numbers Through Audio

Reply posted 7 years ago (09/24/2017)
Nelson, Off hand, all the physical sources of noise that I can think of (the ones already mentioned, and radioactive decay, avalanche diode noise, etc.) are either...

Re: Generating Random Numbers Through Audio

Reply posted 7 years ago (09/20/2017)
Nelson, I see that quite a few people already answered you, but here are my 2 cents.First, if your project is to produce random noise, don't allow people to convince...

Re: Generating Random Numbers Through Audio

Reply posted 7 years ago (09/19/2017)
The audio you mention will probably not have sufficient randomness to be useful for random number generation. As you said it will probably have tones and various...

Re: Fixed-point FFT scaling

Reply posted 7 years ago (09/18/2017)
Although adding two fixed point numbers requires one more bit in the worst case, this does not usually happen at every stage. A worst case would be the DC signal...
There are many tricks for finding the reciprocal, the sqrt of sum of squares, and the reciprocal of the sqrt sum of squares.Search for Goldschmidt's algorithm and...
I headed such a team for about 10 years. It was indeed great fun, but that all ended about 5 years ago.The opportunity window on a lot of our products closed, and...
Sumit, I've never seen the book you referenced, but you are quite correct that the subject deserves a full sized book. Having worked over 30 years in large commercial...

Re: Cross Correlation with Increasing Finite functions

Reply posted 7 years ago (07/06/2017)
One obvious approach already proposed would be to reduce the noise level by FIR filtering and then finding where a threshold was crossed by each.A related approach...

Re: Cross Correlation with Increasing Finite functions

Reply posted 7 years ago (07/06/2017)
Sorry, but cross correlation is the wrong approach here.When you say "follow a logarithmic growth pattern" I assume that you mean that x(t) = a log(kt), and by "time...
Not all OFDM systems use head and tail times in addition to the cyclic prefix.Those systems that perform windowed overlap-and-add can benefit from lower out-of-band...

Re: The adoption of the frequency measure "Hertz"

Reply posted 8 years ago (03/30/2017)
The SI convention regarding units that are proper names is that when spelled out they are NOT capitalized (thus ampere, volt, coulomb) unless used as modifiers...
Yes, acoustically exciting MEMS devices can induce undesired output signals, especially when in the resonance range of the device.Many analog attacks have been demonstrated...

Re: Why do we want an amplifier to be linear?

Reply posted 8 years ago (01/31/2017)
The most important property of linear time invariant systems is that they obey              Y(w) = H(w) X(w).In the presence of nonlinearities you...

Re: How to create a vector graphic of a block model

Reply posted 8 years ago (01/31/2017)
When writing my DSP textbook I created a language that I called flow, which compiled to TeX. It only handled drawing DSP flow graphs, but vastly simplified that...

Re: Interactive digital filter design

Reply posted 8 years ago (01/31/2017)
How would that delete a particular zero or pole (rather than whatever happens to be listed as the "last" one) ?

Re: Interactive digital filter design

Reply posted 8 years ago (01/31/2017)
Really nice.How can I delete a zero or a pole ?Y(J)S

Re: LTI system

Reply posted 8 years ago (11/03/2016)
Sure, the concept is easily extendible (multiple dimensional signal processing is a very well established field). All I was saying is that you can no longer say...

Re: LTI system

Reply posted 8 years ago (11/02/2016)
We normally speak of LTI regarding a single input (x) - single output (y) system, while 2 of your examples have 2 inputs (a and b). A system with 2 inputs can be...

Re: PLL in presence of noise

Reply posted 8 years ago (07/26/2016)
ChessI am not sure we understand the problem.Differentiating always increases noise because it is a highpass filter. By taking the mean I assume that you mean a...

Re: "System Identification" Ideas

Reply posted 8 years ago (07/26/2016)
Fred, It doesn't matter that you think of the switch position as an input - if it changes the observed behavior, you can consider it as separating the two possible...

Re: "System Identification" Ideas

Reply posted 8 years ago (07/25/2016)
Fred3, It seems to me that it would be more useful to consider this as a classification problem rather than a system identification one. Even when nonlinear, system...

Use this form to contact Y(J)S

Before you can contact a member of the *Related Sites:

  • You must be logged in (register here)
  • You must confirm you email address