 ### Omer Sayli (@omersayli)

B.S. and M.S. in Electrical and Electronics Engineering, Ph.D. in Biomedical Engineering. Biomedical signal analysis, machine learning and image processing are the main interests.

pretty good

## Re: Comparing Utterances

Hi,If you have no signal processing background, these tasks would be very hard for you. Alternatively, If you have deep learning knowledge, you may use Neural Networks...

## Re: extract analysis and synthesis weighting filters

I guess it is this;https://dsp.stackexchange.com/questions/83632/extr...

## Re: Python (numpy) to Matlab/Octave conversion problem

In NumPy, np.newaxis is used to add 'dimension' to the vector or matrix. The * is used for one-to-one multiplication of elements (not matrix product). I guess...

## Re: Python (numpy) to Matlab/Octave conversion problem

Hi,p2 += p1[:, np.newaxis] * self.coeff[k]means p2 = p2 + p1[:, np.newaxis] * self.coeff[k]

## Re: Viva Ukraine!

I hope that Ukrainian people get through this invasion without much pain and loss

:)

## Re: Possible to get audio signatures?

Hi,These are some quick search results you may be interested;Classification of Vehicles Based on Audio Signals using Quadratic Discriminant Analysis and High Energy...

## Re: Phase brake in tonal signal and frequency estimation

You introduce discontinuity -a spike like effect- at the instant where s2 starts just after s1. It is inevitable that spectral effects will occur, depending on the...

## Re: How to estimate the real channel response from its PSD?

Hi,I finally got it, as you said, when using centered frequency result, we have to use ifftshift for the correct result instead of h = ifft(fftshift(sqrt(psd)));that...

## Re: How to estimate the real channel response from its PSD?

Hi, The imaginary part of h is very small compared to the real part, here their maximum values; >> max(real(h)), max(imag(h)) ans =1.4761ans = 0.0099 Hence,...

## Re: Proving Time Invariance of a system

Hi,For time invariance, you have to check whether if the shifted input signal also leads to the same output shifted by the same amount of time delay. i.e.y(t) =...

...

## Re: Nyquist frequency & odd even FFT

Hi,X value is complex exponential, x = exp(j*2*pi*1/2*(0:1000-1)); it will have a single frequency peak in the frequency domain.Even for a real x, if x is at the...

## Re: Help with polynomial zeros

Actually we should normalize 'b' by b in the loop (and after the loop, not by sum(b)), then we get the same result suggested without for loop;without for (added...

## Re: Help with polynomial zeros

Trick for ff1 without loop solution is nice.  Yes, the roots are too close, stopband bandwidth (2xFb/Fs ~ 0.002) is also too low compared to sampling rate.

## Re: Help with polynomial zeros

Hi,even after b/sum(b), coefficients of b remain high;   1.0e+15 *  Columns 1 through 8    0.0183   -0.2008    0.9831   -2.7972    4.9864   -5.3791 ...

## Re: Sensor data

1 column would mean one random variable.  There won't be matrix for covariance for a single column data.  I commented in the code, you have missed them. 3 columns...

## Re: Sensor data

I coded in Octave actually, I tested this in Matlabclear all; close all; clcN=100; % number of snapshots.M = 3; % Number of dimensions/variablesRzz = zeros(M, M);for...

## Re: Sensor data

Hi, Could you check this code; clear all; close all; clcN=100; % number of snapshots.M = 3 % Number of dimensions/variablesRzz = zeros(M, M)for j=1:N  # z has...

...

## Re: Combining two wave generation formulas not working as expected (MATLAB/OCTAVE)

Yes, this is what you want I think  (I plot f0=5 case to show clearly that it works); function saw = saw_wave_trig(fs, f0, s, m, w) % fs = sample rate % f0...

## Re: Combining two wave generation formulas not working as expected (MATLAB/OCTAVE)

It works, (In your code, I noticed the wrong use of "t" in your calculations. I tried to change your code, you may check below )function saw = saw_wave_trig(fs,...

## Re: Covariance Matrix diagonal or not?

Think about calculating standard variance, you divide by the number of samples - 1 (for unbiased sample variance). Here N is your data sample number in the covariance...

...

## Re: Covariance Matrix diagonal or not?

Hi, No, there is no absolute value operation in covariance definition, but transposition and conjugation.

## Re: Covariance Matrix diagonal or not?

Hi,Since you use 5 samples for the 3 frequencies in complex exponential functions, you should use below for calculating covariance function (Resembling Matlab's...

## Re: Difference between harmonic (vocalized) and non (consonant) sounds

Hi,In a speech (lets say a word utterance), simply vowels are vocalized (harmonic) parts, consonants are plosives, fricatives, nasals, liquids, and semivowels. You...

## Re: PLL fun : some personnal work that I'd like to share

This is nice indeed, could be made much better by showing model block diagram?

## Re: Noise estimation and thresholding methods in Discrete Wavelet Transform

Hİ, Even if it is fixed, you apply in on sub-bands (there could be coefficients exceeding threshold or not). Coefficients are found for each level of tree, hence...

## Re: What are the wavelet packet functions and how are they different from the wavelet and scaling functions?

Hi,You know in Continuous wavelet Transform, you scale wavelets. In decimated discrete wavelet transforms, after each filtering you downsample the signal by 2 factor...

## Re: Moving Average Filter's Magnitude Response

I'm glad that it helped.. Different libraries in a computing environment may provide same functionality sometimes, and you may also derive simple coefficients yourself...

## Re: Moving Average Filter's Magnitude Response

Also you seem to get frequency response for output signal (fvtool(movAvgOut)), not filter response  (MovAvg)

## Re: Moving Average Filter's Magnitude Response

Hi,This is simple filter with window length of 5 (coefs 0.2). I got this using Python;  import numpy as npfrom scipy import signalimport matplotlib.pyplot as pltb...

## Re: How well can discrete wavelet packet transform reduce noises that are similar to the input signal in the same frequency band?

Hi,I noticed that you use extremely high sampling frequency of 937kHz, while the EEG signal frequency content is below 100 Hz. Is it 937 Hz, is there a typo? Practically,...

## Re: Generate a signal with linearly increasing Bandwidth

Can you check this

## Re: Generate a signal with linearly increasing Bandwidth

or addding segments of gaussian white noises filtered with changing band pass filters could be tried also. You may check this.

## Re: Generate a signal with linearly increasing Bandwidth

Hi, A simple way is to use multiple chirp generators. YOu can control the bandwidth with the "boundary" chirps and also can add chirps in between these two. YOu...

## Re: Auto-Correlation based Carrier Frequency Recovery

Hi, Maybe filtering the signal for the strong tones before your algorithm could work ?

## Re: Matlab Loop Construction ?

Hi RichardI am happy that I was helpful, although a bit late :) Omer

## Re: Matlab Loop Construction ?

Hi,Actually I have always used Wavelet analysis for "off-line" analysis. Also the formulation of wavelet analysis somewhat requires whole data, think about the scaled...

## Re: Is there any example to show how to use the “OFDM Baseband Modulator” block in the newer versions of Simulink?

THis one?http://ecelabs.njit.edu/ece489v2/lab6.pdf

...

## Re: ADC and Complex Mixer Nyquist question

Hi Rick,Sorry, of course your Xc(m) is negative shift by fc (I stated sinosoids but missed total signal xc(t)-Xc(m)) My point was that as you have stated X(m)...

## Re: ADC and Complex Mixer Nyquist question

Hi sachinwannabe,Assuming communications signal (?), I want to remind , a baseband signal of interest with a much low frequency content (Bandwidth) -compared to...

## Re: ADC and Complex Mixer Nyquist question

Hi Rick,You say "Sequence xc(n) can have independent spectral components anywhere in the one GHz range of -0.5 Ghz -to- +0.5 Ghz.". Since we shift the spectrum to...

## Re: Extract pitch of irregular signal ?

If you look at the tunelabguy's plots above, he confirms indeed fundamental frequency around 190 Hz and its 9 harmonics. YOur plot is also in this way, confirming...

## Re: Extract pitch of irregular signal ?

Your plot also reflects peaks at harmonics of ~185 Hz.  380 Hz is one of them. (Fs=44100 Hz, good enough for resolution - not a reflection case here)

## Re: Extract pitch of irregular signal ?

This is also a very nice approach, eliminating noise. I also found (also in this thread a sent) 380 Hz but there is a previous peak at 185-190 Hz. That is the desired...

## Re: Extract pitch of irregular signal ?

I used autocorrelation (in Pyton, used librosa library) with the restriction that 120-240 Hz is the range for pitch find range.  import numpy, scipy, matplotlib.pyplot...

## Re: Extract pitch of irregular signal ?

For human sound analysis, there are a lot of algorithms based on autocorrelation or frequency spectrum for "pitch estimation". spectrum.jpeg Looking at spectrum,...

:))

## Re: A question regarding MATLAB's 'invfreqz()' command

You are right Rick ,   b = [1 0 0 0 -1]  a = [1 -0.9]  freqz_points = 512  [h, w] = freqz(b, a, freqz_points);  [imp,t] = impz(b,a);  h(1) = sum(imp)  b_Order...

## Re: A question regarding MATLAB's 'invfreqz()' command

Hi Mr Lyons,First value of h is NaN as I have checked. Maybe that is the reason?Edit: (As noted by anamariatome also, that is the problem indeed, pole at z=1(f=0) causing...

## Re: The first algorithm for computing the inverse chirp z-transform (ICZT) in O(n log n) time

Yes, it was big news indeed. Engineers solve 50-year-old puzzle in signal processingan interesting comment".. The authors are applying for an international patent...

...

## Re: Advice in Noise Suppression in Voice Processing Context

Hi,you may check this book;Noise Reduction in Speech Applicationshttps://www.cin.ufpe.br/~hama/Arquivos/Noise%20Reduction%20in%20Speech%20Applications.pdfa python...

## Re: Correlation between a signal segment and its stretched replica

I think this is more like a 'pattern search-matching. Once match is found, delay is to be computed.Edit: I meant of course stretching-shortening signal of interest,...

## Re: Correlation between a signal segment and its stretched replica

I also thought that but the signal to be searched has to satisfy wavelet criterion to be used in a wavelet analysis as a 'wavelet'. If he can find a similar wavelet...

## Re: Correlation between a signal segment and its stretched replica

very wise solution

## Re: Average Impulse Response from multiple measurements

Hi,If I were to do that, I would first check in time domain the similarities-differences of the impulse responses. Then in frequency domain, I would check differences...

## Re: Fixed Point Library for Python

Hi,I use Python actively (forgetting even Matlab:) and found numpy/scipy libraries very decent and sufficient. Some info is here especially for DSP; https://docs.scipy.org/doc/scipy/reference/signal....Your...

## Re: Estimating SNR in the Frequency Domain

Hi,Your updated version seems to calculate SNRs, I tried different f0 values. What was particular combination of parameters?

## Re: Estimating SNR in the Frequency Domain

Hi, actually taking square values of the spectrum (sig_r_spectrum  = (sig_r_dft(1:NFFT/2)).^2)  is needed for finding powers, and multiplication by two is for...

## Re: Auto Correlation in DSP processor

Hi,If you really need a real-time operation, you may try DSP processor based solution. But unless you need that, I suggest to use Octave/Matlab, Python based analysis/solutions,...

...

## Re: Filter a rectified waveform?

HiStill you can apply rectify-Low pass filtering-ThresholdingYou may need to collect data to determine the threshold value. Also low pass filter cut off frequency...

## Re: Matlab Loop Construction ?

Hi,Actually as fas as I have understood from your code, you are trying wavelet denoising, hence your question is related to the applicability of wavelet analysis...

## Re: Filter a rectified waveform?

Hi,Actually it depends on noise type but first rectification and then applying low pass filter is better for thresholding purposes, as I have understood from your...

Use this form to contact omersayli

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

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