Voigt funcion in mpfit

Discussion:

(too old to reply)

abc

2017-10-25 11:15:57 UTC

HI, I have set of data of an absorption spectra and I want to fit the voigt profile using mpfit. The function I am using is http://www.heliodocs.com/php/xdoc_print.php?file=$SSW/yohkoh/ucon/idl/metcalf/voigt.pro
But I am getting error as
GDL> res=mpfit('voigt',p0,functargs=fa)
% VOIGT: function VOIGT takes 2 params: 'Result = VOIGT(A,U)'
% Execution halted at: MPFIT_CALL_FUNC_EXTRA 1396 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% MPFIT_CALL 1439 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% MPFIT 3185 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% $MAIN$
Kindly help

Markus Schmassmann

2017-10-25 13:37:18 UTC

Permalink

Post by abc
HI, I have set of data of an absorption spectra and I want to fit the voigt profile using mpfit. The function I am using is http://www.heliodocs.com/php/xdoc_print.php?file=$SSW/yohkoh/ucon/idl/metcalf/voigt.pro
But I am getting error as
GDL> res=mpfit('voigt',p0,functargs=fa)
% VOIGT: function VOIGT takes 2 params: 'Result = VOIGT(A,U)'
% Execution halted at: MPFIT_CALL_FUNC_EXTRA 1396 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% MPFIT_CALL 1439 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% MPFIT 3185 /home/abc/Documents/gdl/pro/pro/mpfit.pro
% $MAIN$
Kindly help

I recommend looking at the documentation in MPFITFUN.pro & VOIGT.pro &
https://en.wikipedia.org/wiki/Voigt_profile#Relation_to_Voigt_profile

I'll probably end up with something like this, but check the derivatives
before using AUTODERIVATIVE=0 or PARINFO[i].MPSIDE = 3.
You can do this manually or using
; PARINFO[i].MPSIDE = 3 ; Enable explicit derivatives
; PARINFO[i].MPDERIV_DEBUG = 1 ; Enable derivative debugging mode
; PARINFO[i].MPDERIV_RELTOL = ?? ; Relative tolerance for comparison
; PARINFO[i].MPDERIV_ABSTOL = ?? ; Absolute tolerance for comparison
for more details see the documentation in MPFIT.pro

I hope this works, beyond compiling I have not debuged it.

Good luck, Markus

function voigt_wraper, x, p, dp
voigt, p[1]/(sqrt(2)*p[0]),x/(sqrt(2)*p[0]),h,f
if n_params() gt 2 then begin
requested=dp
dp = make_array(n_elements(x), n_elements(p), value=x[0]*0)
if requested[0] ne 0 then $
dp[*,0]=-p[2]*(p[1]*f-x*h)/(sqrt(2*!dpi)*p[0]^3)
if requested[1] ne 0 then $
dp[*,1]=-p[2]*( (2*p[1]-h)/(sqrt(2*!dpi)*p[0]^2) $
-(p[1]^2*h+2*x*f*p[0]-x^2*h)/p[0]^3 )
if requested[2] then ne 0 dp[*,2]=-1/(sqrt(2*!dpi)*p[0])*h
if requested[3] then ne 0 dp[*,3]=1d
endif
return, p[3] -p[2]/(sqrt(2*!dpi)*p[0])*h
end

p0=[sigma0,gamma0,norm,cont]
p_out = MPFITFUN('voigt_wraper', x, y, err, p0)

wlandsman

2017-10-25 15:01:12 UTC

Permalink

It looks the program is using the intrinisc IDL Voigt function which has 2 parameters:

http://www.harrisgeospatial.com/docs/VOIGT.html

--Wayne

Craig Markwardt

2017-10-25 19:28:30 UTC

Permalink

First of all, the VOIGT you linked to is a procedure, not a function. MPFIT only works with functions, not procedures. It tries to call VOIGT() as a function, and it gets the system version. Is that what you want?

Second of all, VOIGT has two inputs. How are those meant to be passed? You don't show it in your example.

Third, where is the data? What you showed is just a bare call to VOIGT. Don't you want to compare the VOIGT function results to some data? MPFIT() by itself wants you to compute the residuals. Or, if you don't want to do that, you can use MPFITFUN instead, but you still need data.

Fourth, what are the parameters? Are you trying to solve for A and V?

Craig