Timeline for DFT Polyphase Filter bank
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 4 at 23:00 | comment | added | Naveen | The page linked in the answer no longer exists. | |
| Jan 5, 2019 at 13:03 | comment | added | Narendra | I have one question in the derivation on the page you provided in your answer. Referring to Figure 2 on that page, if we are downconverting the signal to low pass filter it, we have to multiply with $e^{-j \dfrac{2 \pi kn}{M} }$ but it was given as $e^{j \dfrac{2 \pi kn}{M}}$. So, instead of doing DFT, I am ending up in doing IDFT. Can you tell me where I am doing wrong? | |
| Jul 10, 2013 at 6:54 | vote | accept | KharoBangdo | ||
| Jun 25, 2013 at 6:31 | comment | added | KharoBangdo | i just joined this site & my reputation 1 is not allowing me to post images of my signals. Just help me with the following query - I am building 4 channel filter bank. My freq band is 0 - 70 MHZ. My lpf cutoff is 8.75MHz ie fs/8. When i build the analysis filters ie Ek(n)=h(nM+k) i get 4 filters, E0 E1 E2 E3. Since i am decimating in time, signal expands in frequency. So Each one of E0 E1 E2 & E3 become filters from 0 to fs/2 ie 0 to 35 MHz. Now, when i pass my input signal(also polyphased by 4) Do i need to add relative shifts of fs to each E1 E2 E3 filters or not? | |
| Jun 24, 2013 at 12:53 | comment | added | Matt L. | You should add some more information to your original question, and formulate any other more detailed questions there. Try to be as clear as possible as to what you've done and how you've implemented everything. This will make it easier to help you. | |
| Jun 24, 2013 at 12:50 | history | edited | Matt L. | CC BY-SA 3.0 |
added 8 characters in body
|
| Jun 24, 2013 at 9:42 | comment | added | KharoBangdo | ykt is a 4x65567 matrix is the output of Analysis bank. W is a 4x4 DFT matrix. My output yktDFT=W X ykt . Is this the correct way to generate output? | |
| Jun 24, 2013 at 9:33 | comment | added | KharoBangdo | I have 4 signals in the freq band from -35MHz to 35Mhz. Using direct form I am able to split the signals into 4 sub-bands at 4 outputs of analysis filter. But when i try to implement this in Polyphase form, i am not able to split them. I am using 4 pt DFT matrix as given in the Wikipedia article. Is there a mistake i am making? | |
| Jun 24, 2013 at 9:14 | history | edited | Matt L. | CC BY-SA 3.0 |
added 3 characters in body
|
| Jun 24, 2013 at 9:14 | comment | added | Matt L. | Each DFT bin corresponds to 1 sub-band, so with an $M$-point DFT you have $M$ sub-bands. | |
| Jun 24, 2013 at 9:01 | comment | added | KharoBangdo | Thanks for this answer. Cleared some of my concepts. Also, when u say sub-band processing(separating i/p into different freq. bands) does the DFT block does sub-band processing or does it just facilitates it. I mean to say, does the output of DFT block give me M sub-band signals in frequency OR do i have to add some more blocks after DFT to break the signal into M freq sub bands? | |
| Jun 24, 2013 at 8:24 | history | answered | Matt L. | CC BY-SA 3.0 |