Release Notes


Fixed scaling bug for Nortek Vector instruments.


Fixed display bug when a boundary definition file is used on an existing dataset.


Decimal separator in CSV files can now be specified through the configuration panel. This gets around an issue where the CSV file does not use the same decimal separator as is standard for the country set on the system the VSA is being run on.


Fixed locale issue when importing Vectrino II Profiler .dat files (files converted with a comma as the decimal separator were not imported correctly).


Added file extension column to import dialog table, and formatted the "Last Modified" time.


Added option to scale wavelet transforms by the instantaneous power before normalising by the total energy (see v1.5.60 of the User Guide).


Improved scaling for FWT contour plots. At each time, the power is scaled to match the power of the original time-series at that time, and then normalised by the mean power.


Improved scaling for CWT contour plots. At each time, the power is scaled to match the power of the original time-series at that time, and then normalised by the mean power.


Tidied up probe axis transformation from the file name and rotation correction. The configuration "Invert axes" check-boxes should be configured such that the data are in a left-handed co-ordinate system.


Fixed rotation correction calculation bug.


Slight modifications to multi-run calculations to improve speed.


Fixed multi-run bug where mean values for properties such as turbulence intensity were not calculated correctly for the mean run dataset.


Changed TI graph to show true TI rather than scaled by the overall mean U velocity.


Changed power spectral density to remove the mean before calculating, so it now only operates on the fluctuating part of the signal. The integral (mean power) is therefore the standard deviation of the input signal.


Default to the last used configuration file on start-up.


Scale wavelet transforms so the total energy (integral over frequency and time of the wavelet transform output) is the same as the time-series.


Improved estimation of pseudo-frequency for DWT (uses FFT approach, or falls back to "number of crossings of zero" approach).


Power spectral density is now scaled such that the integral of the PSD over frequency equals the mean power of the time-series.


Added option to use Welch's method to calculate the power spectral density.

Corrected calculation of -5/3 line for power spectral density.

Corrected duplicate of ".csv" file extension, which prevented .csv file import.


Added Matlab connection output for scatter (XY) graphs.


Fix locale issues when importing converted Nortek files (.dat).


Added display of "-5/3" line to power spectrum display.


Added functionality to allow the order of the spike replacement polynomial to be set by the user.


Fixed bug in calculation of Q/A (introduced when dataset level configuration was added).


Increased range of values allowed for bank positions and water depth.


Some of the configuration which was global (default directories, CSV file separator) is now configurable on individual datasets.

The format of the CSV file can now be specfied in the configuration panel, using 'x', 'y' and 'z' to represent the velocity in the x, y and z directions respectively, 'xs', 'ys' and 'zs' to represent the SNR in the three directions, and 'xc', 'yc' and 'zc' to represent the correlation in the three dimensions. The field codes are separated by a comma, e.g. x,y,z,xc,yc,zc would be used for a CSV file holding the three velocity components followed by the three correlation values, x,xs,y,ys,z,zs would be used for a CSV file holding the velocity in the x direction followed by the SNR for the x velocity, followed by the velocity in the y direction, etc..


Added functionality to allow configuration files to be copied so that different configuration files can be created and used for different projects.


Fix bug introduced in v1.5.35 where initialisation failed if no configuration file was available.


Configuration file can now be selected at start-up via drop-down on licence box.


This release includes a number of development changes which should not affect existing users. In addition, it contains: Fix for a configuration bug ("seconds to ignore" was not being saved correctly). Multi-run functionality now does the filtering at the run level, rather than on the run-mean time-series.


Modified the velocity correlation calculation (previously this was a stub which just output data to the console). Fixed bug in which right-click functionality on the main velocity graphs had been broken.


Continuous Wavelet Transform (CWT) added.


Cosmetic changes only.


Added auto-trim functionality.


Added wavelet analysis functionality.


Minor bug fixes: Remove spaces from variable names when writing datapoint details to Matlab file. Correct enabling of configuration "Apply" button when changing bank positions.


Fixed bug retrieving probe setup details for saved datasets (bug introduced in v1.5.23).


Added functionality to allow sending of datapoint details direct to Matlab.

Corrected errors in text.


Dataset specific configuration GUI now uses the same layout as for the configuration dialog. This does not affect functionality.


Fixed freeze when using direct Matlab connection.


Corrected bug with enabling of "Export as Matlab Script" functionality (the old export functionality). This should work as before. ***Please Note*** When using the new direct export to Matlab functionality introduced in v1.5.20, the VSA has been seen to freeze. I am working on a fix.


Improved Matlab integration - figures can now be exported directly to Matlab without cutting and pasting of a script. Added version checking so any new updates should be shown in the about box.


Fixed scaling bug in power spectrum estimation.


Added functionality to allow all velocity details (velocities as measured by probe, translated velocities and filtered velocities, plus correlation and SNR values) to be exported to an ASCII file.


Fix Vector import bugs - probe id and sampling rate were read incorrectly from .vec files.


Added option to filter by percentage of mean correlation and SNR rather than an absolute value.


Fix lock bug.


Fix converted Vector file import bug, introduced when the SNR and Correlation filtering was added.


Made the number of windows used for the Bartlett's method power spectrum calculation configurable.


Fix binary file read bug which was putting the y-velocities into the z-velocities component.


Added pre-filtering functionality. Exclude Level or Correlation and SNR filtering may now be applied as a pre-filter before another filter is applied, such as Phase-Space Thresholding.


Changed power spectrum estimation to use Bartlett's Method.


Fixed About box sizing bug.

Corrected spectral distribution labels.


Fix export bug when there are negative y-coordinates.


Fix initialisation bug when there is no configuration file (i.e. on first use of the software).


Fix bug when importing multi-probe, multi-run CSV files (dialog not opening).


Minor bug fixes.


Minor bug fixes.


Addition of option to store the data in binary format, considerably reducing the file size. Please note that once a dataset has been created and had data-points added to it, the binary/xml option should *not* be changed.


Multi-run synchronisation index bug fix for when reloading a saved dataset.


More bug fixing following update of external libraries.

Fixed issue for multi-run, large file datasets.

Bug fixing following update of external libraries.


Updated external libraries - now using JFreeChart v1.0.14, JCommon v1.0.17 and JAMA version 1.0.3.


Added "wDiff" filtering for instruments which give two z-axis (w) velocities. Filtering is done by specifying a maximum absolute difference (|w1-w2|) as a percentage of the RMS value of the w1 time-series.

Added graphical display of signal correlation and SNR data, and wDiff data.


Added SNR filtering (Vectrino, PolySync, Vector and Vectrino II files).


***Bug fix for thread lock when calculating data point parameters.***

Added configurable channel boundary functionality.


Added save to Matlab file functionality.


Multi-run bug fixes.

Performance improvements.


Added probe details display.


Added Vectrino II support for NDV (.dat) files and Matlab format (.mat) files. Vertical profiles are reconstructed using the cell size and number of cells data held in the .hdr (for .dat) or .mat files.


Added trim functionality - data point time-series may be trimmed to the required sub-series.

v1.3.16 & v1.3.17

Minor bug fixes.


Fixed problem importing some Vector .vec files.

Changed reading of NDV .hdr files so that filename is no longer used to identify the start of the data definitions (this fixes issues if the files are renamed after the conversion to .hdr/.dat files).


Fixed import bugs due to probe setup configuration changes.

Multi-run data point details do not show the data for the individual runs unless the "-a" flag is set


Turbulent Flow Instrumentation Cobra probes are now supported (unconverted .thX files)

Probe setup configuration has been improved

Negative coordinates are now allowed (e.g. -10--20.vno for the point at y=-10, z=-20)


Vector .vec files are now supported, along with converted Vector files (.dat)

Vector .vec burst mode files are now automatically split (using the same method as outlined below for large files)

Sampling rate is automatically set based on the value in the imported file (Vectrino, Vector and PolySync files only, both converted and unconverted)

Import of converted Vectrino files has been made more robust

There is now an option to split large files to make them more manageable ("large" means greater than approximately 200000 measurements but it's all down to user experience - for large files it is slow to show time-series graphs, etc). For example, if splitting is activated then a file "0-10.vno" will be imported as a number of data points: "0-10", "X-10", "2X-10",... where X is the number of measurements per split file divided by the sampling rate (i.e. it is the time spanned by the data in each split file). The y-coordinate thus becomes the start time for the data in that file relative to the start of recording. For burst mode, X is just the burst interval.

Please Note

This software is released in the hope that it will be useful to other users of Nortek ADVs or other devices with compatible outputs. It is released under the terms of the GNU General Public License, version 3. Please follow the link for formal licencing information, but note that this software comes with no warranty (having said that, it is designed to only read from the original data files, not write to them, so your data should be safe!).

Test version - please do not use this version unless you have been asked to!

For more information about the software, please contact:
Feedback is welcomed - please let me know if you find the software useful or have any problems with it.