Bruker spectra analysed by WINISI? Log Out | Topics | Search
Moderators | Register | Edit Profile

NIR Discussion Forum » Bruce Campbell's List » I need help » Bruker spectra analysed by WINISI? « Previous Next »

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

Fernando Morgado (fmorgado)
Senior Member
Username: fmorgado

Post Number: 32
Registered: 12-2005
Posted on Friday, March 15, 2013 - 8:26 am:   

Dear Kelley :

I have a program for move *.SPC ( jdx) to *.nir or *.nir *.cal to JDX. Will be necesary create again *.cal for run the models in winisi.
But if you have jdx a lot of quemometrics program work with this.
U-cal from Unity Scientific is a nice program
Regards
Fernando
Top of pagePrevious messageNext messageBottom of page Link to this message

Vaithilingam (nvlingam)
New member
Username: nvlingam

Post Number: 2
Registered: 1-2013
Posted on Friday, March 15, 2013 - 5:04 am:   

Hi Friends,

Could you please send me more information on this subject as I need few calibration to be converted from OPUS to WINISI 4.
Please suggest the detail procedure.
N V Lingam
Top of pagePrevious messageNext messageBottom of page Link to this message

Pierre Dardenne (dardenne)
Advanced Member
Username: dardenne

Post Number: 24
Registered: 3-2002
Posted on Tuesday, May 15, 2007 - 7:34 am:   

Hi Liam,

We are performing continuous transfers of spectra from OPUS to WINISI and WINISI to OPUS. I could give more details if you email me directly. Dozens of data bases have been already successfully transferred.

I think that all the software (using PLS) will give very similar results on the same data sets. It can be slight differences from the pretreatments, the derivatives, etc..

If the results are significantly different from one instrument to another, do consider carefully the sample presentation. If the Foss 6500 gives better results, it can be due of the use of the visible part and than the comparison is biased. I would like also a better definition of your independent test set: is it a (random) selection among the total available set or a real test set, meaning new samples collected and measured in a different period?

I would be very curious to play with these data sets.

Pierre
Top of pagePrevious messageNext messageBottom of page Link to this message

Liam Kelley (reflex99)
New member
Username: reflex99

Post Number: 2
Registered: 12-2006
Posted on Tuesday, May 15, 2007 - 2:44 am:   

Hi All,

Thanks so much for your great suggestions (and speedy response). I should add some more detail...we had Bruker Applications Specialists produce their best calibration with OPUS using Bruker spectra and it was significantly worse (R2 SE, independent validation set) than a calibration produced using WINISI with spectra collected on a Foss 6500 for exactly the same dataset.

So is this a consequence of the different software or spectral differences?

I will try the range of suggestions - Unscrambler, Cesar's conversion option and will contact ISI directly. I'll keep the board informed.

Finally this is an incredibly supportive and helpful scientific community - you should be proud of what you have!!
Top of pagePrevious messageNext messageBottom of page Link to this message

Cesar Guerrero (cesar)
Member
Username: cesar

Post Number: 12
Registered: 3-2006
Posted on Saturday, May 12, 2007 - 1:59 pm:   

Hi Liam!
I usually work with a Bruker MPA. My spectrometer is a Bruker MPA FT-NIR, and the spectrum is not collected every 2 nm. It is collected at constant wavenumber (for example, I work with a resolution of 8 cm-1). I suppose that you have the same apparatus.
With this FT-NIR, the spectrum have more than 2000 points (between 12000 and 3800 cm-1). With OPUS (see in Manipulate), you can transform automatically the spectrum from wavenumber to wavelenght (for example to a new spectrum of 1000 points between 833 to 2632 nm).
I observed always better (slightly) calibrations when they were made with spectra transformed to wavelengths. Probably its a consequence of a smoothing, but I'm not sure at 100%.
In OPUS you can modify many parameters, such the maximum range of the model (10 is the default number). I don't know if you need many PLS-vectors in your application, but you can try with an increase in the limitation of the maximum range.
Anyway (as other option), you can access to the spectrum data making double click in the History of the spectrum. You can copy and paste this matrix with the wavelength (or the wavenumber) and the absorbance from OPUS to Excel.

Good luck!
Top of pagePrevious messageNext messageBottom of page Link to this message

Gavriel Levin (levin)
Senior Member
Username: levin

Post Number: 34
Registered: 1-2006
Posted on Saturday, May 12, 2007 - 12:31 pm:   

Hi Guys, some time ago there was a sequel about translation from FT wave numebrs to nm so spectra can be used across techniques, so it coulld possibly help.

If everything fails, my recommendation get a package which is instrument independent, such as Unscrambler or equivalent and you would like the results.

Thanks,

Gabi Levin
Brimrose
Top of pagePrevious messageNext messageBottom of page Link to this message

David W. Hopkins (dhopkins)
Senior Member
Username: dhopkins

Post Number: 109
Registered: 10-2002
Posted on Saturday, May 12, 2007 - 9:34 am:   

Hi Liam,

I think the issue with the Foss ISI software is that it written for spectra with either 700 or 1050 points per spectrum. I have never used the Bruker Opus software, so I am only guessing, but I think the FTNIR spectra have considerably more points per scan. Therefore, you cannot do a direct importation. You will lose some information if you attempt to shorten your data to fit.

A more direct problem is you. How long did it take you to learn to use ISI effectively? I think you should recognize that you need to devote more time to using Opus effectively. Again, I have never used Opus myself, but my understanding is that it is very powerful and effective. I recommend you get help from Bruker to get over the initial hurdles.

There is another option. You could learn to use another chemometrics program to work with your data. These programs are designed to work with data of various lengths, and have the power you need. The list includes Grams, Matlab with chemometrics tool-boxes from various sources, Pirouette, Simca-P, and Unscrambler, to name those that come to mind. And, you will have to invest time to become an effective user of any of those.

My advice is hard, but you need to recognize that you will need to learn various software packages over your career. You probably should make Opus the next one.

Best wishes,
Dave
Top of pagePrevious messageNext messageBottom of page Link to this message

Howard Mark (hlmark)
Senior Member
Username: hlmark

Post Number: 95
Registered: 9-2001
Posted on Saturday, May 12, 2007 - 5:57 am:   

Liam - I think your best bet would be to contact ISI directly.

Your second-best approach would be to see if the Bruker software can export the data to, and the WINISI can import the data from, a non-proprietary format such as JCAMP-DX.

There's a bigger difference in conversion between FTIR and grating-type data than the spacing difference, however. As you say, FTIR data has constant spacing between wavenumbers. The equivalent data in nanometers does not have a constant spacing. There are two ways to deal with this:

1) use software that "knows" what it's dealing with and can correctly interpolate between the measured wavenumbers.

2) "Cheat": if you're working with a JCAMP-DX file you can leave the data alone, but modify the file so that it looks like the data is in nanometers instead of wavenumbers, before importing it into WINISI. In this case the burden is on you to keep track of what's what, but the computer won't know the difference and will treat the data as what it represents itself to be.

The remaining problem would be whether the WINISI software can handle the amount of data that the Bruker files contain. I don't know the answer to that, but here again, your best information will come from ISI.

Howard

\o/
/_\
Top of pagePrevious messageNext messageBottom of page Link to this message

Liam Kelley (reflex99)
New member
Username: reflex99

Post Number: 1
Registered: 12-2006
Posted on Saturday, May 12, 2007 - 1:01 am:   

Hi Folks,

I have a number of spectral datasets collected on a Bruker MPA instrument. I am not that happy with the calibrations obtained with the Bruker software and would like to use my old WINISI software to develop calibrations using the Bruker-collected spectra. The reason is that I understand the WINISI software but the Bruker software is much more black box.

Does anyone have experience with converting data between these two formats? The problem is that (to my understanding) the Bruker data is collected every 2 nm and the normal FOSS data for which WINISI is designed is every 10nm. Secondly, the Bruker data is in wave numbers and not wavelengths Therefore exporting directly as a JCAMP won't work. But others who are cleverer than me might have some simple solutions.

Alternatively does anyone from FOSS reading this know if there are upgrades of the WINISI software to enable it to cope with a greater data resolution.

Add Your Message Here
Posting is currently disabled in this topic. Contact your discussion moderator for more information.