Adaptive digital filters - page 30

New comment
 

Here is a graph of the frequency lag. Calculated through the FFT. I did it especially for you. So learn. Smart guy.


 

here's more - for a different period. zero frequency on the left, rising to the right.


 
And how was the lag measured? :-)
 

"Here's a graph of the lagged frequencies. Calculated by FFT."

a friend of mine once told me a funny story. he came to give a physics exam to a first-year student. one of the students gave him a problem. he was writing, writing, waving his hands, showing me sheets with integrals, counting and counting, sheet, two, three, and then my friend suddenly asked "what do you need to find"? that was the end of the exam. the student had to be kicked out of the class, he fell into a daze. you are exactly the same. I will not even ask about the lag WHERE you were counting and what you have on the axes in this chart. Nothing clever there anyway. Especially inspiring, of course, is the pearl "Calculated by FFT", and then the strange formula with some w. Do you know that you have a bunch of spectral components with known w, forming equidistant spectrum, with step, according to the window width in this FFT? What w you substituted into your Delay(T) - and why?

 

The calculation formula is delay(T) = Arctg(Im[S(jw)]/Re[S(jw)])*T.

S(jw) is the AFC of the SMA filter, T is the period. The first multiplier is the phase lag. When multiplied by period we get lag in units of time

Not even a hint of equal lag. So leave the idle speculation out of it.

 
mikfor:

"Here's a graph of the lagged frequencies. Calculated by FFT."

a friend of mine once told me a funny story. he came to give a physics exam to a first-year student. one of the students gave him a problem. he was writing, writing, waving his hands, showing me sheets with integrals, counting and counting, sheet, two, three, and then my friend suddenly asked "what do you need to find"? that was the end of the exam. the student had to be kicked out of the class, he fell into a daze. you are exactly the same. I won't even ask about the lag WHERE you were counting and what you have on the axes in this chart. Nothing clever there anyway.

Try to manually draw the sine waves and run it through the SMA. The method is monkey-like, but suitable for beginners.
 

alsu , you see,I am not a freshman, so I take responsibility for my words. SMA lags the same at all frequencies. And I'm too lazy to take apart your nonsense with the alleged FFT.

"bother to manually draw the sine waves and run them through SMA. The method is monkeyish, but suitable for beginners." - one such like you I so and proved exactly the same as I say now. And now - laziness.

 
By the way, Alsou, at least read on the same forum http://forextechnologies.ru/for/viewtopic.php?f=49&t=67
 
mikfor:

alsu , you see,I am not a freshman, so I take responsibility for my words. SMA lags the same at all frequencies. And I'm too lazy to take apart your nonsense with the alleged FFT.

"bother to manually draw the sine waves and run them through SMA. The method is monkeyish, but suitable for beginners." - one guy like you I so proved exactly what I'm saying now. And now - laziness.

You see, I'm not a freshman either, and I know enough about lagging of linear filters to tell you that you're wrong without any FFT. Remember (if you know) how the spectral transfer function of a filter with rectangular characteristic, i.e. SMA, looks like a curve sin(x)/x. And the corresponding MTF is not as straight line as you're imagining. And the pictures are just for your understanding.
 
mikfor:
By the way, Alsou, you should at least read http://forextechnologies.ru/for/viewtopic.php?f=49&t=67 on the same forum.
And what here contradicts my statements?
1...232425262728293031323334353637...41
New comment