What is a Discrete Fourier Transform (DFT) and an FFT?

Поделиться
HTML-код
  • Опубликовано: 3 янв 2025

Комментарии • 24

  • @문태원-b2o
    @문태원-b2o 23 дня назад

    Thank you so much!! you are the best teacher i've ever met.

    • @iain_explains
      @iain_explains  23 дня назад

      That's great to hear. I'm glad you are finding the videos helpful.

  • @YilinWang-v3j
    @YilinWang-v3j 8 месяцев назад +5

    Thank you so much sir, it helps my final week!

    • @iain_explains
      @iain_explains  8 месяцев назад

      I'm glad to hear it. Good luck - it sounds like you've got exams coming up.

  • @fishgrey6441
    @fishgrey6441 8 месяцев назад +1

    Thank you very much for your explaination about the DFT. Back in my college year, my DSP teacher just told us periodical in time domain is discontinuous in frequency domain and so on. Your video builds a link between those abstract concept and is really helpful to me. By the way, would you please consider making a video about the windowing, such as the Taylor window, Chebyshev window...

    • @iain_explains
      @iain_explains  8 месяцев назад +1

      Thanks for your comment. You've read my mind - I was actually planning to make a video on windowing, but realised I needed to first have a video explaining the DFT. So you can expect to see a video on windowing in the next couple of weeks. 😁

  • @ozonewagle
    @ozonewagle 4 месяца назад +3

    5:43 Nearly slipped convolution from mouth 😂😂😂😂.
    Btw I am in race to finish this playlist.
    Completed my engineering, but still I have been puzzlled about it.
    It still feels like a magic to me.

    • @iain_explains
      @iain_explains  4 месяца назад

      I'm glad my videos are helping you.

  • @sam2938
    @sam2938 Месяц назад

    I understand that you plotted only the magnitudes of the FT. What about the phases? Do they exhibit any property when the signal is discrete or when it repeats?

    • @iain_explains
      @iain_explains  Месяц назад +1

      The phase plots are generally less easy to visualise, and often we're mostly interested in knowing which frequencies "exist" in a waveform (eg. to find the bandwidth, etc), so I didn't try to plot the phases. But yes, the phases are important, and they repeat, just like the amplitudes. For more insights on the phase in Fourier transforms, see: "Is Phase important in the Fourier Transform?" ruclips.net/video/WyFO6yBQ0Cg/видео.html and " What Does "Linear Phase" Mean?" ruclips.net/video/aQ__XatMxJo/видео.html

    • @sam2938
      @sam2938 Месяц назад +1

      @ makes sense. Thank you very much

  • @noNameGamerAbsolutely
    @noNameGamerAbsolutely 4 месяца назад

    I know what is DFT (very briefly), your videos are usually great, but I didn't quite understand what you explained below. What am I missing? what else I should revise before I watch this video again to understand?

  • @weicai1911
    @weicai1911 3 месяца назад

    You are the best teacher I ever met, I su subscribed your channel immediately. I have a question about this video at the ending part, if we sample the signal with time T, then the highest frequency we can get is 1/2T, not 1/T, isn't it?

    • @iain_explains
      @iain_explains  3 месяца назад

      It's not exactly clear what you mean by "the highest frequency we can get". In the frequency domain, the function repeats itself every 1/T.

  • @bahramhedayati423
    @bahramhedayati423 7 месяцев назад

    Is the FT of sin function drawn correctly?

    • @iain_explains
      @iain_explains  7 месяцев назад +1

      Yes, but I've only drawn the magnitude. I haven't shown the phase shift.

  • @iahmad1857
    @iahmad1857 8 месяцев назад +1

    Thanks for the content.
    However, I watched the video only to learn what's FFT, but you didn't explain it.

    • @iain_explains
      @iain_explains  8 месяцев назад +3

      An FFT does exactly the same thing as a DFT. It’s just an efficient implementation of the algorithm. The details of the algorithm’s internal implementation are tedious and not really conducive to a video. Anyway, this video is about what it does, not how it’s implemented.

    • @AJ-fo3hp
      @AJ-fo3hp 8 месяцев назад +1

      ​@@iain_explains With kind, I would like to correct the phrase in reply what you have provided
      " It is just an efficient implementation of algorithm"
      Correct phrase is as below
      "FFT is an efficient algorithm for implementing DFT".
      FFT effectively reduces time and memory taken to compute the DFT (each frequency component (N point) of Discrete signal)
      DFT is formula for computing each frequency component (N point) of Discrete signal.
      FFT is algorithm for DFT.

    • @iain_explains
      @iain_explains  8 месяцев назад +3

      You haven't "corrected" me. You've simply re-stated what I said!

    • @AJ-fo3hp
      @AJ-fo3hp 8 месяцев назад

      @@iain_explains
      you written ".. an efficient implementation of algorithm"
      FFT is not "implementation of algorithm"
      FFT itself algorithm for DFT.
      If somebody finds better algorithm for DFT than FFT, in terms of better computing in terms of "time and memory" or in any other way,then we call by different name(other than FFT) for that algorithm for efficient implementation of DFT.
      Fast Fourier Transform is an algorithm.
      The FAST name comes its redcution in time took to compute the DFT.

    • @olaguibel27
      @olaguibel27 6 месяцев назад +1

      @@AJ-fo3hp There are other algorithms you can make to calculate the DFT, the FFT is 'simply' one that does it in an efficient manner. You haven't corrected anything, just restated what was already implied...

  • @JosephLewis-i7h
    @JosephLewis-i7h 3 месяца назад

    Walker Lisa White Elizabeth Young Carol