Igor 6.22 has been released

Igor 6.22 has been released.

This is primarily a bug-fix release, though there are a few new features:

  • Added log colors to trace f(z) color, image plots, and contour line colors
  • Date/time axes in graphs can now be reversed


As always, choose "Updates for Igor" from Igor's help menu to find the best updater for the version you are running, or use the direct links below.

IGOR Pro 6.22 is available as a free update to licensed users of the English version IGOR Pro 6.0 and later.

• Macintosh users of IGOR Pro 6.0 or 6.1 should download the IGOR Pro 6.22 installer:

http://www.wavemetrics.net/Downloads/Mac/Igor6.dmg

• Windows users of IGOR Pro 6.1 or 6.2 may download the IGOR Pro 6.1 to 6.21 updater:

http://www.wavemetrics.net/Updaters/WinUpdateIgor61ToLatest.exe

• Windows users of IGOR Pro 6.0 should run the IGOR Pro 6.22 installer after uninstalling IGOR Pro 6.0:

http://www.wavemetrics.net/Downloads/Win/setupIgor6.exe

• Windows users who wish to run the 64-bit beta version of Igor (6.23B01) should see this page for instructions and the download link:

http://www.wavemetrics.net/Updaters/ReadMe%20(64-bit).html

For a detailed list of bug fixes and changes, see:

http://www.wavemetrics.net/Updaters/WhatsChangedSince6.1.html
Today I tried the new feature of Multipeak Fit 2: "The ExpModGauss peak shape has been modified to accept a negative shape factor". Meanwhile I use the new version 6.22A of course (on Win XP Prof).
I can't get this to work. When I set a negative tau I get a extreme 'cliff' towards negative values instead of a peak with the tail to the left. How does this work?
Ok, I got it to work. Didn't know that the GaussWidth has to be negative too when tau is negative. Still, the initial guess was distracting because there the GaussWidth is calculated as a positive value at first, which looks broken. I have attached a test experiment if someone wants to see how it looks.
It certainly isn't expected that both GaussWidth and ExpTau must be negative, so I downloaded your experiment file and took a look at it. Using the already open Multipeak Fit panel in your experiment file, I tried modifying GaussWidth by removing the minus sign. Indeed, things went quite wrong!

So I tried doing the fit from scratch: I clicked the Auto-Locate Peaks Now button, changed to ExpModGauss peak type, and did the fit. The result was exactly the same fit but with positive Gauss Width. If I now change GaussWidth to negative, it looks just like your case where positive GaussWidth goes bad. With negative GaussWidth all the derived parameters like FWHM, etc., are garbage.

Ah- I see bugs... I'm working on it.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com
OK- I'm not sure how it happened, but your list got out of sync with reality. To update after editing in the list, the code tried to evaluate the old coefficients and subtract the old peak from the graph curves. Then it would evaluate the new coefficients and add it to the curves. I think I was trying to be very clever and make it update quickly, but it was too clever.

I have changed the code so that it updates the coefficients and simply completely re-computes the graph curves. This is more robust in that if something goes wrong, it will self-correct.

The new version of the procedure file is attached. It is compressed- unzip it first. Find your Igor Pro folder (Help->Show Igor Pro Folder) and then navigate down to WaveMetrics Procedures:Analysis:Peak Fitting:. Drag the new, unpacked file into Peak Fitting to replace the old one. The new file will be included in the next minor release.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com