jan
  • Member for 8 years, 6 months
  • Last seen more than 2 years ago
Can we break the Shannon capacity?
Accepted answer
22 votes

Most certainly not. While there has been some claims to break Shannon here and there, it usually turned out that the Shannon theorem was just applied in the wrong way. I've yet to see any such claim ...

View answer
How can a signal have no DC component
Accepted answer
11 votes

A signal having a mean-value or DC component of zero is commonly referred to as mean-free or as having no DC component. It does not mean that it cannot be averaged, just that the average comes out as ...

View answer
Why does pulse-width modulation work?
Accepted answer
11 votes

Depends on what you are controlling. For DC-motors it is the inertia of the device that acts as a low-pass filter of the PWM modulated signal resulting in a continuous motion. For most LEDs it is the ...

View answer
Why is X(0) the DC component
Accepted answer
8 votes

Follows from the DFT definition. It's defined as \begin{equation} X(k) = \sum_{n=0}^{N-1} x(n) e^{-j2\pi \frac{kn}{N}} \end{equation} So $X(0)$ is \begin{equation} X(0) = \sum_{n=0}^{N-1} x(n) e^{-...

View answer
Why is PSD estimated, and not simply computed?
Accepted answer
6 votes

Well, because in a lot of real world problems this If I have a deterministic signal with a fixed number of samples, shouldn't I be able to directly determine its spectral information? is just not ...

View answer
QPSK soft decoding
Accepted answer
6 votes

Unfortunately you did not mention what your channel is so I assume that merely AWGN is present. Furthermore I assume symbol-by-symbol detection is desired since a decoder for the Hammingcode is ...

View answer
What does the one-tap equalizer used with OFDM accomplish?
6 votes

The OFDM signal as a whole is affected by frequency selective filtering. It is usually designed such that the subcarrier bandwidth is smaller than the channel coherence bandwidth. This yields you ...

View answer
Can two nonzero signals $x[n]$ and $y[n]$ give a zero convolution
Accepted answer
5 votes

Yes, for example let $$x[k]=1$$ for all $k$ and $$y[k] = \begin{cases}1 & k=0\\-1 & k=1\\0 & otherwise \end{cases}$$ It is easy to see that in case of a convolution, the result will be ...

View answer
Conceptually, how does real time convolution reverb work?
5 votes

It comes down to latency vs. complexity. If your filter is 10 seconds long, you need to store the audio data of the last ten seconds and then you are able to calculate the current output audio sample ...

View answer
Situations where complexity is too big to exceed linearity and gaussianity
5 votes

In addition to Peter's answer, if you have a nonlinear system that is well-behaved in a sense of being only mildly nonlinear or at least exhibiting no discontinuities, special variants of the Kalman ...

View answer
Dedicated DSP for delayed replay
Accepted answer
4 votes

Hard to believe, that you actually looked. You probably don't even need a DSP for that, a normal microcontroller should be sufficient. For example, the STM32F4-Discovery board comes with an ARM ...

View answer
understanding time constant meaning in signal processing
Accepted answer
4 votes

There is no general formula to calculate time constants for all signals. Time constants are defined for some signals. For your exponential signal \begin{equation} e^{-a \cdot t} = e^{-t / c} \end{...

View answer
Reason for >8 channels in an audio codec
4 votes

Since it will most likely not hurt to have the possibility for more streams, this might for one be just to future-proof the system. Nothing would be more stupid than to specify for 8 channels and ...

View answer
How to model the noise in Kalman Filter?
Accepted answer
4 votes

You are most likely not having a control vector $u$. Maybe you can somehow model it into the filter, but it is not required in your case and uncommon. Furthermore $\overset{.}{x}$ is not your gyro ...

View answer
Given an experimental FFT plot, how to find the best match from a set of generated FFT plots?
Accepted answer
3 votes

Regarding the problem with high correlation values, this happens when both signals are not zero-mean. Assume two signals $x[k]$ and $y[k]$ of length $N$ each. Each signal can be decomposed into the ...

View answer
Continuous time-domain electrical signal that only has a finite number of possible values
2 votes

Well, a DC signal for sure. Other than that, there is probably more context from the book required (especially about the assumptions).

View answer
Skew map to match aerial image
Accepted answer
2 votes

Well you could try to estimate a transformation matrix that maps between the coordinates in the two images. The idea is that \begin{equation} \begin{bmatrix} x_a\\ y_a \end{bmatrix} = \begin{bmatrix} ...

View answer
Carrier frequency offset (CFO) estimation using cyclic prefix
2 votes

In addition to Deve's very good answer above, it should be noted that the phase difference $\arg(r_n^* r_{n+N})=2\pi\beta$ is circular and you cannot tell the difference between $0, 2\pi, 4\pi, ...$. ...

View answer
So is there such a thing as a DSP engineer nowadays?
2 votes

Seems to me that the DSP position goes the same way as most jobs in large corporations. Join a big enough company and you will specialize yourself because for the other things there is a guy who ...

View answer
Normalise data which has peaks
2 votes

You could normalize the average amplitude, i.e. YData = YData / mean(abs(YData)); Or you could normalize the signal power to one, i.e. YData = YData / sqrt(mean(abs(YData).^2)); If just the peaks ...

View answer
About sampling and wireless networks
2 votes

What you describe is a direct-sampling architecture. Those are practically available for radio amateurs using frequencies up to 30MHz. They are also actively researched to open up their usage for much ...

View answer
About designing digital filters
Accepted answer
2 votes

Filtering can be done in the frequency domain which is actually a very efficient technique (and it can very well be, that Matlab does this internally). However, for very long signals it's not as ...

View answer
Circular Convolution in Scilab
2 votes

I can't tell about SciLab but if it is not implemented, you can always get the circular convolution by IDFT(DFT(x1) * DFT(x2)) where x1 and x2 are your signal vectors and the multiplication is ...

View answer
What is the optimal adaptive grid for calculating a DFT using a fixed number of sampling points?
Accepted answer
2 votes

This can be split up in two questions for the sampling frequency and the window. 1) The required sampling frequency is given by the Nyquist theorem to be $f_s>2*f_{max}$ with $f_{max}$ being the ...

View answer
deciphering the frequency response from the input data
Accepted answer
1 votes

Since the signal is most certainly real-valued, you can ignore the FFT values after $f_s/2$ (or sample 300 in your plot) since they will mirror the first ones. With a sampling frequency of 2500Hz, the ...

View answer
Simulate a pulse of specific power
Accepted answer
1 votes

Well, for a signal $x[k]$, the energy is: \begin{equation} E=dt*\sum_{k=-\infty}^{\infty}x^2[k] \end{equation} You want the energy to be one by dividing $x[k]$ through a value $a$, hence: \begin{...

View answer
comparison between frequency offset estimators
1 votes

Well, since you simulate you could calculate the residual frequency offset which is basically the remaining CFO after correction or the difference between the true CFO and the estimated CFO. Another ...

View answer
Head Phone Transfer Function MATLAB
1 votes

It's actually a band-pass characteristic, since very low frequencies will be inaudible on headphones. Firstly, searching for "headphone frequency response" brings up a lot of results, especially on an ...

View answer
Power of summation of two signals
1 votes

No, that is not true. It is \begin{equation} |a+b| \leq |a| + |b| \end{equation} Edit: This is not true for the power which was asked for by the OP (see the comments). I over-read the re-definition ...

View answer
Normalized cross-correlation in detail
1 votes

The inner product or dot product or scalar product will not output a vector, it will output a scalar which is (with exception of the $1/n$ in front of the sum) the same as the first equation that you ...

View answer