| T O P I C R E V I E W |
| villettecbm |
Posted - 03/09/2007 : 04:18:11 AM
Origin Version (Select Help-->About Origin): Pro 7.5
Operating System: Win2000
I want to fit FTIR spectra of close samples, in order to compare them in a semi-quantitative way. that means there are many peaks (even if i "cut" the data in sub-spectra i have at least 5 peaks), they should be fitted by voigt function, and i need parameters sharing since the position and width of the peaks should be the same in the different samples and only the amplitude should change, reflecting the slight modification of the compound.
- the peak fitting module doesn't work, because it will fit only one spectrum at a time, with no possibility to fix the peak position (or i didn't find it)
- the NLSF doesn't accept replicas of voigt functions
how can i do that???
i would appreciate any ideas, thanks
|
| 5 L A T E S T R E P L I E S (Newest First) |
| rlewis |
Posted - 03/13/2007 : 7:44:12 PM
I used the PFM to fit 3 Voigt peaks to the curve posted ..
 |
| villettecbm |
Posted - 03/13/2007 : 10:06:06 AM
1- here is a hand-made picture of what happens when picking peaks:
(i hope it does show an image for you...)
2- using the refresh button doesn't improve things... :(
in fact if i close the window and open it again just after it does keep the limits for position but not for width...
|
| Mike Buess |
Posted - 03/13/2007 : 09:42:24 AM
quote:
if i ask it to find three peaks in my spectrum, they will all be located near the maximum of my principal peak, no way it will find the other, whereas they are really easy to locate. i wonder if it can be due to my data, while i cannot see why, they have nothing really strange...
I can't reproduce that problem. PFM has done pretty well at picking my peaks, even when they're overlapping. You can also adjust manually.
quote:
i cannot fix the upper limits for the peak width, these limits are lost as soon as i try a fit
You need to click the Refresh button before closing the Peak Parameters Display window.
Mike Buess
Origin WebRing Member |
| villettecbm |
Posted - 03/13/2007 : 08:13:00 AM
i have troubles with the pfm:
- if i ask it to find three peaks in my spectrum, they will all be located near the maximum of my principal peak, no way it will find the other, whereas they are really easy to locate. i wonder if it can be due to my data, while i cannot see why, they have nothing really strange...
- i cannot fix the upper limits for the peak width, these limits are lost as soon as i try a fit
this last problem makes the fit impossible to run correctly for my data. Is it a known bug?
really no way to get the NLSF to accept replicas of voigt (that would be perfect for me in fact!)??
Thanks anyway
|
| Mike Buess |
Posted - 03/09/2007 : 12:47:12 PM
quote:
the peak fitting module doesn't work, because it will fit only one spectrum at a time, with no possibility to fix the peak position (or i didn't find it)
PFM is best for a spectrum with many peaks. According to PFM help if you need to fix the peak positions in several datasets you can define them in an ASCII file. Here is the latest version of PFM help (see pg 58 about peak files)...
http://www.originlab.com/pdfs/PFM_Manual.PDF
Mike Buess
Origin WebRing Member |
|
|
