A quick fascinating approach to the mellin transform of sin(x)

Поделиться
HTML-код
  • Опубликовано: 12 ноя 2024

Комментарии • 35

  • @treint6751
    @treint6751 9 месяцев назад +24

    5:56 didn't expect that 🤨

  • @anasharere
    @anasharere 9 месяцев назад +14

    sinx = imaginary part of exp(ix)

    • @idjles
      @idjles 9 месяцев назад

      that's what i thought he'd do

  • @MrFtriana
    @MrFtriana 8 месяцев назад +1

    A bit "wrong" way to do it is remeber that sin(x) = im(exp(ix)) and do a change of variables: ix = -at. With this, only remains calculate the Laplace transform and take the imaginary part.

  • @MrWael1970
    @MrWael1970 9 месяцев назад +1

    Very interesting approach for this integral. Note that for the substitution you should put du instead of dt. Thanks alot.

  • @dominicellis1867
    @dominicellis1867 9 месяцев назад

    I used tablature integration to derive the finite series representation of the Mellin transform. I then evaluated the series at k = s-1 because the other terms 0 out. The answer becomes sin(pi*k/2)*Gamma(s)/Gamma(s-k+1) evaluated at k = s so Gamma(s)sin(pi*k/2) is the final answer

  • @manstuckinabox3679
    @manstuckinabox3679 9 месяцев назад

    3:16 I miss referencing gamelin! But I'll leave the reference as an exercise for the reader and instead just say awesome vid mah dude!
    I think we can also create a left hand inequality to say that the imaginary part of ((-i)^-s*Gamma(s)) = Gamma(s) sin(pis/2) of which we perform the adequate procedures or rigor so that we don't make Gamelin angry.

  • @josephlorizzo8997
    @josephlorizzo8997 7 месяцев назад

    couldn't It be done by im(e^ix) for sinx and a substitution to directly use the gamma function from the definition?

  • @CM63_France
    @CM63_France 9 месяцев назад

    Hi,
    Nice. What can we do with this transformation?

  • @Calcprof
    @Calcprof 9 месяцев назад

    My favorite transform

  • @IrDdn07071
    @IrDdn07071 9 месяцев назад

    Beautiful! And Fun!

  • @ahmedamir3516
    @ahmedamir3516 9 месяцев назад

    Well done buddy!

  • @joelchristophr3741
    @joelchristophr3741 9 месяцев назад +1

    Master can you solve my integration doubt please 🙏
    Int 0 to 1 [ e^x / (x+1) ] dx

    • @ambiguousheadline8263
      @ambiguousheadline8263 9 месяцев назад +2

      multiply the integrand by e/e, you then get 1/e int 0 to 1 e^(x+1) / x+1 which is simply the exponential integral evaluated from 1 to 2, so your answer is (Ei(2)-Ei(1))/e

  • @MohamedachrafKadim-jm5yr
    @MohamedachrafKadim-jm5yr 9 месяцев назад

    Nice bro❤

  • @yaozhang1687
    @yaozhang1687 9 месяцев назад +1

    When s=0, the result is incorrect. Gamma(0) is not defined.

    • @Noam_.Menashe
      @Noam_.Menashe 9 месяцев назад +1

      The limit Gamma(s)*sin(pi*s) exists

    • @maths_505
      @maths_505  9 месяцев назад +1

      Yes and using the famous sin(x)/x limit and the recursion formula for the gamma function, we get pi/2 as guaranteed by the Dirichlet integral.

  • @DavideCosmaro
    @DavideCosmaro 7 месяцев назад

    The Merlin transform 🪄

  • @paulmukendi1826
    @paulmukendi1826 9 месяцев назад

    The condition must be 0

  • @shivanshnigam4015
    @shivanshnigam4015 9 месяцев назад

    hey, please help me on this one
    int 0 to 1 (ln(x + sqrt(x^2+1))/x

    • @bjornfeuerbacher5514
      @bjornfeuerbacher5514 9 месяцев назад

      Just some ideas... First I noticed that ln(x + sqrt(x^2+1)) is equal to arsinh(x). Hence substituting u = arsinh(x) transforms the integral into one from 0 to arsinh(1) over u times tanh(u) du. Then I tried integration by parts; now I'm left with an integral over ln(cosh(u)). I think solving that integral probably works similarly to the integral over ln(cos(x)), the latter one can be found in many places...
      Hope that helps. Probably there is an easier way, but that's the first ideas which came to my mind.

    • @maths_505
      @maths_505  9 месяцев назад +1

      I'll give it a shot bro

    • @shivanshnigam4015
      @shivanshnigam4015 9 месяцев назад

      @@bjornfeuerbacher5514 I'm only in class 12th so I don't have the best knowledge about improper integral
      And I don't even know if this is doable, but I did see its graph and it seems to be convergent
      Thank you

    • @Yoshinoyo1
      @Yoshinoyo1 9 месяцев назад

      ​@@bjornfeuerbacher5514 tanh(u) = (exp(2u)-1)/(exp(2u)+1).= exp(-2u) (exp(2u)-1)/(1+exp(-2u)). Then you could use the serie expansion of 1/(1+exp(-2u)) in power of exp(-2u), I guess ? You would be left with integrating a bunch of u*exp(-k u) and exp(-k u) factors I think.

    • @ambiguousheadline8263
      @ambiguousheadline8263 9 месяцев назад

      @@shivanshnigam4015your integral can be shown to be equal to int 0 to arcsinh(1) of -ln(sinh(x)) which can be evaluated by using the exponential definition of sinh(x), separating the ln(2), factoring out an e^x term from ln(e^x-e^-x), then invoking the series expansion of the log, leaving you with an integral with an antideriative -(1/2)x^2+xln(2)-Li_2(e^-2x)/2, which when evaluated from 0 to arcsinh(1) gives you the answer (1/12)pi^2 - (1/2)arcsinh^2(1) + ln(2)arcsinh(1) -(1/2)Li_2(3-2sqrt(2))

  • @kevinscheengsbier6130
    @kevinscheengsbier6130 9 месяцев назад

    Wow

  • @tamasbarath7501
    @tamasbarath7501 9 месяцев назад +1

    Why are likes hidden?

    • @maths_505
      @maths_505  9 месяцев назад +4

      I honestly have no idea and I've noticed this in the last video too

    • @tamasbarath7501
      @tamasbarath7501 9 месяцев назад +1

      Oh, anyways great video keep up the good work

  • @AronHardemanGoogle
    @AronHardemanGoogle 9 месяцев назад

    I think "Melin" is misspelled in the title :)

    • @maths_505
      @maths_505  9 месяцев назад +1

      Thank you
      Fixed it