This was I get,
The derivative from the voigt is OK,
not the other one !!
===========================================================================
Patrick DUPRÉ | | email: pdupre@xxxxxxx
Laboratoire de Physico-Chimie de l'Atmosphère | |
Université du Littoral-Côte d'Opale | |
Tel. (33)-(0)3 28 23 76 12 | | Fax: 03 28 65 82 44
189A, avenue Maurice Schumann | | 59140 Dunkerque, France
===========================================================================
> Sent: Wednesday, December 14, 2016 at 8:55 PM
> From: "Antonio Olivares" <olivares14031@xxxxxxxxx>
> To: "Community support for Fedora users" <users@xxxxxxxxxxxxxxxxxxxxxxx>
> Subject: Re: Faddeeva
>
>
>
>
>
> On Wednesday, December 14, 2016 6:46 AM, Patrick Dupre <pdupre@xxxxxxx> wrote:
>
>
>
> Thank very good
>
> However, there is something that I do not understand
>
> I define
> W1=0.2
> z(x) = (x + W1*{0,1})
> w(x) = faddeeva (z (x))
> and
> Voi (x) = voigt (x, W1)
>
> Voi (x) and real (w(x)) are exactly the same
> Now,
> d_w (x) = - 2 * z (x) * w (z (x)) + 2*{0,1}/sqrt(pi) " which is the derivative
> D_voi(x,del) = (Voi (x+del) - Voi (x-del)) / 2 / del
>
> now
> plot [-4:4] D_voi(x,0.001)
> and
> plot [-4:4] real (d_w (x))
>
> are different except for W1=0
>
> What am I doing wrong?
> It seems that the issue is with the faddeeva, or when I my calls!
>
>
> Thank.
>
> ===========================================================================
> Patrick DUPRÉ | | email: pdupre@xxxxxxx
> Laboratoire de Physico-Chimie de l'Atmosphère | |
> Université du Littoral-Côte d'Opale | |
> Tel. (33)-(0)3 28 23 76 12 | | Fax: 03 28 65 82 44
> 189A, avenue Maurice Schumann | | 59140 Dunkerque, France
> ===========================================================================
> _______________________________________________
> users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx
> To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
>
>
>
> I am now a bit confused. {0,1} represents the number 0 + 1i correct?
> Then I would believe that Mr. Howell is correct in that the magnitude should be called upon since to get a real number function. His comment
> <quote>
> I am not a mathematician, but I have programmed extensively
> with complex numbers. W(x)=faddeeva(z(x)) will return a complex,
> right?
> The magnitude of a complex is its full value, so you would need
> the sqrt(real^2+imag^2) would you not?
>
> </quote>
> z(x) is the one defined with the z(x) = (x + W1*{0,1}) which would be z(x) = x + 0*i which is just x because the W is 0. If the function is defined correctly and I have a strong hunch that it is, then there is a problem with gnuplot and it has to be a bug somewhere. We can try plotting the function using plotshare.com which is a frontend to gnuplot.
>
>
> There is no output from the plots check in
>
>
> http://www.plotshare.com/index.ws/plot/432515199
> It actually plots one function only and the other one does not come up.
>
>
> Best Regards,
>
>
>
> Antonio A. Olivares
> _______________________________________________
> users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx
> To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
>Attachment:
Faddeeva.pdf
Description: Adobe PDF document
_______________________________________________ users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
