| Electronics > Projects, Designs, and Technical Stuff |
| Low frequencies Vector Network Analyzer, arduino based. |
| << < (10/20) > >> |
| MasterT:
--- Quote from: Dave on September 23, 2018, 10:31:56 am --- --- Quote from: MasterT on September 22, 2018, 10:04:40 pm ---FFT here is the most efficient BPF, that allows to cut 244 Hz out of 250 kHz . Now, I 'd repeat last time, there is NO alternative to fft, stop crying about bins that goes to garbage can, fft is doing one specific function: it's sharply cut 244 Hz out of 250 kHz. Anything else, single bin DFT, Goertzel, FIR etc - don't do this simple task. Easy they are not able to shrink bandwidth, or they are extremely inefficient on a clock cycle scale. --- End quote --- You seem to lack the very basic understanding of what FFT is and how it works. Perhaps this video can clear up some of your misconceptions. --- End quote --- I'm not commenting on FFT, there is another thread. https://www.eevblog.com/forum/microcontrollers/fft-processing-using-ucpu/ Know it better than me, good, be happy. Stop talking, show your code. My FFT library is only 10% of on timing compare to Keil's, what is yours? |
| Dave:
--- Quote from: MasterT on September 22, 2018, 10:04:40 pm ---FFT here is the most efficient BPF, that allows to cut 244 Hz out of 250 kHz . Now, I 'd repeat last time, there is NO alternative to fft, stop crying about bins that goes to garbage can, fft is doing one specific function: it's sharply cut 244 Hz out of 250 kHz. Anything else, single bin DFT, Goertzel, FIR etc - don't do this simple task. Easy they are not able to shrink bandwidth, or they are extremely inefficient on a clock cycle scale. --- End quote --- --- Quote from: MasterT on September 23, 2018, 01:17:08 pm ---I'm not commenting on FFT, there is another thread. --- End quote --- :palm: Hmm... what could possibly make me think otherwise? --- Quote from: MasterT on September 23, 2018, 01:17:08 pm ---Stop talking, show your code. My FFT library is only 10% of on timing compare to Keil's, what is yours? --- End quote --- Boy, it sure is difficult to acknowledge you were wrong, isn't it? It doesn't matter how efficient your FFT implementation is, it's never going to be more efficient than a single bin DFT. You can convince yourself that you're right as much as you want, but you're arguing against the laws of mathematics here. |
| MasterT:
--- Quote from: Dave on September 23, 2018, 02:54:59 pm --- --- Quote from: MasterT on September 23, 2018, 01:17:08 pm ---Stop talking, show your code. My FFT library is only 10% of on timing compare to Keil's, what is yours? --- End quote --- Boy, it sure is difficult to acknowledge you were wrong, isn't it? It doesn't matter how efficient your FFT implementation is, it's never going to be more efficient than a single bin DFT. You can convince yourself that you're right as much as you want, but you're arguing against the laws of mathematics here. --- End quote --- Nothing of the kind. Is it engineering forum? I don't like to be involved in any philosophical / political / or abstract mathematics debates. And I 'm not worry if there is a life on a Venus. The point is, engineers should talk in different terms, leaving abstract "better", "less memory" " faster". Tell me in dB, uVolts, microseconds, or in this particular case, tell me the Bandwidth in Hz for single bin DFT. There is no issue with pure math, consider ratio: "Bandwidth / Clock_cycles_uCPU". I'm not argue, that single bin (6-8 math operation code) is faster than 2048 bins, it's all about bandwidth (second page in a row). For single bin Bandwidth is 250 kHz. |
| radioactive:
The Goertzel algorithm is even more efficient than a single-bin DFT. https://www.dsprelated.com/showarticle/495.php |
| ogden:
--- Quote from: MasterT on September 23, 2018, 03:32:18 pm ---The point is, engineers should talk in different terms, leaving abstract "better", "less memory" " faster". Tell me in dB, uVolts, microseconds, or in this particular case, tell me the Bandwidth in Hz for single bin DFT. There is no issue with pure math, consider ratio: "Bandwidth / Clock_cycles_uCPU". --- End quote --- :-DD LOL it is already said here in this thread. He not only ignores my posts but most likely ignore posts of everybody but him: --- Quote from: xaxaxa on September 22, 2018, 01:47:56 am ---If you multiply your input signal by sine and cosine, then do integrate and dump, the result you get is mathematically equivalent to one frequency bin of your FFT. You can see this just from the formula of the DFT. FFT takes O(n log n) to compute whereas the single frequency DFT takes O(n) even without taking into account memory locality effects, so the integrate and dump is a better choice for this application. --- End quote --- [edit] Please somebody poke him with xaxaxa post |
| Navigation |
| Message Index |
| Next page |
| Previous page |