That is the SpecView2008 engine with GA for profile fitting
and FFT embedded for intellegent comparison of simulation and experimental spectrum
- Specview2008
Use toolbar to display current partition function and normalize spectra intensities! -
Custom increments for dynamically changed constants! -
in a menu choose Constants -> Increment -> Custom
you might use increment by value or percentage
You may also find useful a couple of new interactive features:
1. Instanteniously measure a distance between two lines - position a mouse pointer to the first line, press '1' key, then position the pointer arrow on the next line and press '2'. A small dialog box will show the accurate number for frequency distance in current units.
2. Mark the lines under interest with "cubes", "circles", "pies" and other labels. For that just simply position a pointer close to the line that you want to label and press keyboard key '3' to '9' for various labels. To remove all labels press '0'.
Table of currently available Labels:
| Shape '3' | Shape '4' | Shape '5' | Shape '6' | Shape '7' | Shape '8' | Shape '9' |
|---|---|---|---|---|---|---|
| Cube | Circle | Pie | Pie inv. | Pie horiz | Pie inv. horiz | Half circle |
Also:
The most recent developments in Profile fitting! Use rapidly converging to obtain the best set of constants without even making line by line assignment!
Some limitations apply, go to the home page for the most recent advances in GA Profile Fit algorithms
Technical details:
The Genetic Algorithm tune-up parameters are selfexplanatory, you can find the meaning on one of the many GA web pages (here is the example I grabbed from Google: GA demo), but here is the set of parameters that never failed me:
Poolsize - number of genes in the population - 80-100
Selectivity - 4-5, may be larger if evaluation function surface is very plain
# of cycles - genes born and die, number of bio revolutions - usually 20-30 is enough, limited with the calculation time. The total number of spectrum recalculations is # of cycles X poolsize.
Reincarnation rate - search for other minimuma - keep it at default value of 15%
Crossover - two parents get two offsprings, good for converging stability - 45%
Mutation - improving the minimum - 45%
Surviving - genes getting to the next cycle - not active since this percentage is defined by the following formula: 100% - reincarnation - mutation - crossover
Mutation degree - the default 50% does a great job!
For explanations on Profile Fit dialog box refer to the Profile Fit Help Page
Download
SpecView2008.exe (GA with FFT-based comparison) now
--To use upgrades from this page you have to have the original package SpecVi38a.exe installed for proper files configuration.--
You are very welcome to leave comments or express feelings you experienced using SpecView. Your feedback is important for development of the best spectrum analysis tools.
Last revised on December 31, 1969, 7:00 pm