Details of Auto-Phase

March, 2003

The autophase routines (QA, QP and AP) were re-written to give faster and better results.  In the course of testing algorithms, several parameters for the AP command were created.  While in most cases, users have no need of these extended commands, they are available, and are described here for completeness.

Each spectrometer generates characteristic distortions of phase and baseline, so finding an algorithm that works in all cases is difficult.  The user should experiment with commands and parameters to see what works best for his/her data.

Part of the process involves a decision as to which data points are baseline and which are peaks.  One option is to ignore points at each end of the spectrum, which is useful for data that is digitally filtered, because these spectra commonly have badly distorted points at each end (see example).  

In the descriptions below, "A" phase refers to zero-order and "B" phase refers to first-order phase correction terms.

QA adjusts only the zero-order phase parameter.  QP 

The AP and QP commands can take arguments, when executed in the non-2-letter command. mode.  They are described here for AP, but can be applied in the same manner with QP.

QP is a subset of AP and therefore somewhat faster and slightly less accurate.  

AP ? 
will display possible arguments.

AP log
will do the AP process and log the steps the AP routine takes, written to a file in the current data directory.

AP stepA #
sets the initial step size for the zero order phase.

AP stepB #
sets the initial step size for the first order phase.

AP A
will phase only the zero order phase.

AP look 
will exchange the current data for the automatically generated reduced data point data buffer used by the AP routine.

AP enhanced #
will set the AP routine to use advanced baseline detection when # is one. When # is zero the AP process will not use advanced baseline detection. This feature works better with spectra which having peak-free baseline at each end.

AP IgnorePts #
will set the number of points on each end to ignore in the phasing process to #.  By default (IgnorePts = -1) the AP process ignores 5% of the data points on each end of the spectrum. When the data set to be phased has peaks close to the ends it is better to set this number to the minimum allowed by the data's baseline.

AP RMS #
will set the number of RMS noise multiples to use in the advanced baseline peak detection mode. Default is 100 times.

AP Pts2Ave #
will set the number of points to average when making the reduced data set for the AP process. The default value is zero which allows the AP process to automatically determine how much to reduce the data based on initial data sizes. QP reduces the data more than AP. If this argument is set to one, then the data is not reduced and the process is slightly better, but a lot slower.

AP quality
does not do an automatic phase process but returns the value used by the AP process for the quality of the current spectrum's phase.

AP Simplex
does the AP process using the Simplex method and creates a log to the current data directory.

AP map
sets the first order phase to -360 degrees and does a zero-order optimization.  It then increases the B phase in 10 degree steps and optimizes the zero order phase at each step until the first order phase reaches positive 360 degrees. It then corrects the current data set to the best phase values it found in this mapping process. This process also creates a log to the current data directory.

Last updated: 4/22/03