This is excellent sir!!! I am taking a filters course at the undergraduate level EE, and my professor is really good, but this is a nice supplement and summary of the differences between FIR and IIR.
Actually it is ok as written. Note that the sum in the denominator starts from k=0, and we define a_0 = 1. It would be equivalent to write "1 + sum" where the sum starts at k = 1.
Barry Van Veen ı have a question about fır filter can you help me about it? gokhan.arslan.763@gmail.com you can send mail to me and ı will send question to you on gmail. Thank you so much
This is a good video. I think that it would have been useful to say and show a diagram that in order for any filter to be able to recognize a band of frequencies, then the SHAPE of the impulse function of the filter must have a kernel containing the signal to be selected. * A Low pass filters would have an decaying impulse function which contains the SHAPE related to frequencies to be selected * A high pass filter would have an IMPULSE followed by the inverted shape of the impulse of a low pass filter. * A bandpass filter would have a decaying impulse that encloses an oscillatory signal related to the frequency to be selected. Then the FIR and the IIR Filters would have provided this shape which is shown on flat paper as a flat function but in reality, it is a three-dimensional rotating function going around a time axis After all the signal e^-jwt or e^jwT may be looked at as time locations or rotating vectors are in the form of cos ( wt) + j.sin( wt) which is a rotating vector that operates in the frequency and time domain. It would be a remarkable digital filter if it is shown in three dimensions where the pulses rotate and the delay and adders operate in three dimensions. It is also useful to look upon the Laplace and Convolution function as being three dimensional as we deal with e^jwt and not only the flat functions sine(wt) and cosine ( wt)
Never understood the difference between these two. Just 1 min into the video, my doubt is cleared and i realized i have already used them by name of moving average and software filter.
Thank You so much for this and for al people visiting this, do check out the link that shows up in the video..Really neat explanation of complex topics..Thank You from India Sir...
Had to deal with this in 1997, with limited internet and very few books left in the Uni library. Between this and working fuzzy logic out, I'm clearly smarter than I thought at the time. Since then, I'm clearly not. :D
Would've been nice if you briefly went over what all the different letters stand for in your equations. Not everyone learns the mathematical notation of filters at their university, you know :)
Interesting video. I had trouble making low pass filter (analog) and high pass filter, also analog. This may solve many troubles in electronic design. I need that for SSB (Single Side Band) generation utilizing 'Weaver method'. Yesterday tried to get that, and ended using three inductors for LPS, plus one for HPS, but the problem is that result is not that great. I got -60 dB attenuation at 3 kHz with LPS, but not that great for HPS for audio filtering before IQ mixers. Just... I am newbie in DSP, and don't know what is what in terms of varialbes (k=0?, M, n, and other things). Just wanted to implement that formula on my STM32, but this concept is not clear, or I am not yet familiar with that math. Do you have any link that may help solve this beginners nightmare? Thank you in advance.
I implemented a fir filter with code, and I understand summation notation usually, but this is hard for me to get. I'm only 5 minutes into the video tho.
Oops sorry i didn't notice that the limits for y[n] in the eq. above H(z) were from 1 to n, so it is 1+sum in deno. if k=1,2,3..N which is apparantly equal to sum if k=0,1,2,3...N
thanks for the video! I just have one question: at 1:50 you say that FIR filters have poles at z=0. I don't really understand why. Isn't a pole a value that makes the denominator 0? And since here the denominator is 1...
The system function (z-transform of impulse response) for a causal FIR filter is a polynomial in z^{-1}. For example, a simple one is H(z) = 0.5 - 0.5z^{-1}. This is equivalent to H(z) = (0.5z - 0.5)/z - in this form you clearly see the pole at z=0. Note that in the first form with z^{-1} when z = 0, the z^{-1} blows up.
Barry Van Veen So when do we use FIR and IIR? It seems that they both cannot be used for the same problem. When we say filters, I take it to mean band pass band stop notch filters, does it mean something more. How do I use this concept of FIR and IIR????
