Datasqueeze Software

Technical Details

Home
Technical Details
Screen Shots
FAQ
Download
Purchase
Plans
About Us
Contact
Manual
Privacy Statement
Site Map
 System Requirements
General Description
What Datasqueeze will NOT do
Supported Data File Formats
File Input Options
Image Manipulation
Supported Graphics Output Formats
Calibration
Line Plots
Least-Squares Fits
Printing
Statistics
Batch/Scripting mode
Help/Documentation
Changes from Previous Versions

System Requirements

To install and run Datasqueeze you need the following:
  • A computer operating under the Windows, Linux, or Macintosh OS X (10.3 or above) operating systems. (The program also runs under Solaris, and perhaps other Unix, operating systems, but this has not been extensively tested.)
  • Java 2 version 1.4 or above. This is bundled with OS X and is available as a free download for Windows or Linux.
  • A CPU running at at least 300 MHz.
  • A color monitor display resolution of at least 1024 x 768
  • At least 60 MB of available memory
  • At least 15 MB of free disk space.
Image of ENIAC--history!

General Description

Datasqueeze is a graphical interface for analyzing data from 2D x-ray detectors (image plate, CCD, wire). The data are represented as a false color image, with substantial control over the colors displayed. The program provides facilities for changing the color scale of the false color image, recentering the image, correcting for detector tilt relative to the incident beam, changing the q-scale of the entire image (i.e., calibrating the angular range), producing x-y plots of intensity versus Q (the momentum transfer), 2-theta, Qx, Qy, chi, as a Porod or Guinier plot, or along an arbitrary line across the image, saving the image in multiple graphics formats, saving the x-y data as an ascii file, and adding or subtracting multiple data files. Multiple open windows allow visual comparison of different data sets. The x-y data extracted from the 2D images can be least-squares fits to a variety of functions.

Datasqueeze will run on multiple platforms, including Windows, Linux, and Macintosh OS X. This means that the scientist is not tied to the computer that produced the data, or to a central "analysis computer," but can carry the data off to his or her office to be examined at leisure.

Datasqueeze is particularly useful for the analysis of powder diffraction data, diffuse scattering from polymers or liquid crystals, or small-angle scattering ("SAXS") from colloids, polymers, gels, or solutions.

What Datasqueeze will not do

Datasqueeze is not well suited for extracting intensities from many sharp Bragg peaks in a single-crystal diffraction type experiment (although it has been used to extract information about pixel statistics in such images). Software is available from other vendors that will do a much better job of crystal structure determination from single crystal diffraction.

Datasqueeze is totally unsuitable for the analysis of radiographic or tomographic images obtained from the transmission of x-rays through a patient or sample.

Supported Data File Formats

Currently supported formats are the Bruker standard (wire or CCD Detectors, *.raw, *.unw, *.sfrm, SMART, GADDS, BIS/PROTEUM), Bruker-Nonius KCD format (for ccd detectors, *.kcd) ADSC CCD detectors (*.img), Brandeis CCD detectors, ESRF format (multiple detector types), MAR CCD detectors (*.img), MAR Image plate detectors (300, 345, and pck formats--*.img, *.image, *.pck), Rigaku image plates (Mercury, Raxis-II ,Raxis-IV, Raxis-IV++ formats--*.osc), Nonius DIP-2000 (image plates--*.ipf), Molecular Metrology detectors (*.mpa), Princeton/Roper CCD (*.spe), Fuji image plates (*.img/*.inf), Ditabis Image Plates (*.ipc), Gatan DM3 format (*.dm3), Black/White Tiff (*.tif, *.tiff), and XDR (*.xdr), and "Binary-little-endian", "Binary-big-endian". We anticipate adding more formats in the near future; contact us if there is a particular data format you would like to see added.
It is also possible to read in Grayscale graphics images--this might be a way of reading in an image that is a Fourier transform of some real-space data, for example produced by a microscope or AFM.
Graphics images are not the preferred way of reading in data from x-ray detectors--they are intended only as a way of reading in non-xray data, for example produced as a Fourier transform of some optical image. A graphics file is not the same as the original data file. A typical data file is 1024x1024 pixels, and has a depth of at least 16 bits (depending on the detector type). Graphics images are generally much smaller (so spatial resolution has been lost) and have much less dynamic range. Additionally, note that Datasqueeze expects grayscale images. It converts the image color to an intensity by adding the red, blue, and green components. If you start with a color image, this is almost certainly not how the data were encoded.
Right now only PNG, GIF, and JPG grayscale images are supported--expansion of this feature will be driven by customer demand.

File Input Options

Up to 13 files can be read in and added with arbitrary multiplicative and additive constants. So, for example, several patterns collected under the same nominal conditions could be added to improve statistics, a background file could be subtracted from a data file, or an arbitrary number of counts could be subtracted from each pixel.

The detector parameters (position of beam zero, angular range, etc.) can optionally be preserved upon reading in new data, so that these parameters can be set once using a calibration file and preserved thereafter.

It is also possible to read in ascii x-y data sets, either independently generated or previously produced by Datasqueeze and "massaged" by the user, for graphical analysis.

Image Manipulation

The data are represented on the screen as a 512 x 512 pixel2 image. The intensity values corresponding to least intense (black) and most intense (white) are under user control. Several options are provided for the intermediate color scheme. The intensity scale can be linear or logarithmic. The image can be demagnified by 50% or magnified by factors of 2, 4, or 8. Different choices for the color scale are available, including black-on-white and white-on-black. The data can also be represented by equal-intensity contour lines. The image can be "de-zinged" (individual very intense pixels, sometimes seen in CCD detectors, are set to the average of surrounding pixels) or smoothed. Multiple pixels can be binned together to create a smaller dataset with fewer pixels and more counts per pixel (useful for low-statistics data). The image can be corrected for detector tilt about an arbitrary axis (this amounts to compressing the image along the orthogonal direction). The image can be rotated in steps of 90o. The user can superimpose a cartesian or polar grid to help locate positions of features of interest. A Fourier-transformed image of the data can be displayed.

Supported Graphics Output Formats

The false color image can be saved to disk in the PNG or JPEG formats.

Calibration

The user can set the values of the wavelength, detector radius, sample-to-detector distance, angular range, or Q (momentum transfer) range. (Some of these values depend on each other). The pixel corresponding to the center of the image can also be set. These operations can be carried out using previously calculated numerical value, "by eye"with a point-and-click interface, btr using a Bruker calibration file (an easily created text file containing a list of d-spacings corresponding to a particular calibration material), or by auto-centering on a Bragg ring of known d-spacing.

Line Plot

The user can make a one-dimensional line plot using a variety of cuts through the data. The possible variables are Q (the magnitude of the momentum transfer), 2-theta (the scattering angle, in degrees), Chi (the azimuthal angle), Qx (the horizontal component of the momentum transfer), Qy (the vertical component of the momentum transfer), or the fraction of the distance along an arbitrarily chosen line segment. Each can be used as the plotting parameter or the parameter over which a sum or average is performed. For example, selecting Q/Chi causes the data to be plotted versus Q (a radial plot) while integrating over the azimuthal angle. The ranges of the dependent and integration parameters can be typed in or set graphically (click and drag on the corners of a control region). Plots can be made on linear, logarithmic, or semi-logarithmic scales. The user can drag a cursor over the line plot to get a readout of the integrated intensity and independent variable at each point. The data from the line plot can be saved in an ascii file for subsequent publication-quality plotting by a program specialized for this purpose (Origin, Excel, Kaleidagraph, Sigma-Plot, etc.)

Several preset configurations for line plots are provided. A Powder Plot plots the intensity versus Q=2 pi / d (with sensibly chosen limits and increment) and integrates over the azimuthal angle. A Pole Figure plots he intensity versus the azimuthal angle and integrates over Q. A Guinier Plot displays log10(intensity )versus Q2 (graphs of this type can be used to obtain the radius of gyration in a small-angle experiment). A Porod Plot displays Q4 I(Q) versus Q; it should approach a constant for large Q if the scattering arises from isolated objects. Two user-defined presets allow the user to save and re-use particular plot settings.

Least-Squares Fits

The x-y data extracted from the 2D images can be least-squares fits to a variety of functions, including Lorentzian, Gaussian, or Voigt lineshapes, polynomials, power law, sine wave, or the Rayleigh or Yarusso-Cooper functions (used for the analysis of approximately spherical dispersions). The user constructs a function as a sum of one to ten of the provided functions (so that, for example, up to ten different peaks could be analyzed at once), enters a starting value for each parameter, and selects which parameters are to be varied. In addition to minimizing the fitting parameters, the program calculates the one-sigma error bars for each parameter, both considered as completely independent parameters (one-parameter error bars) and including the correlations with all the other fitted parameters (multi-parameter error bars). The fitted function is plotted on top of the original data, and this plot can be printed or saved. The user can also save a text version of the fitted parameters.

line plot image

A graphical indexing tool assists in the determination of the lattice parameters of a powder diffraction pattern. The user selects a Bravais lattice, and then observes where the diffraction lines would fall for various choices of the lattice parameters.

Printing

The user has the choice of printing out a single summary page containing the false color image plus the line plot plus descriptive text, or the image only, or the plot only. (Fully supported on Macintosh and Windows only; results may vary on Unix/Linux platforms).

Statistics

The user can obtain basic information about maximum, minimum, and average intensity per pixel, and can also construct histograms of pixel intensity frequencies both for the data set as a whole and for the region selected for plotting.

Batch/Scripting Mode

Although the program is written as a graphical interface, it can also be run in a batch (or scripting) mode, in which commands are taken from a file that can be created by any text editor. This is useful for automated processing of multiple data files. There is also a "Process Multiple Files" option, in which multiple (in principle, hundreds or thousands) of files are all processed in exactly the same way.

Help/Documentation

Onboard documentation is provided in the form of a help window with a detailed description of each component and a short tutorial. A complete manual in pdf format is supplied when the application is downloaded and is also available as a standalone file. Most graphical components in the user interface can be right-clicked (control-clicked on Macintosh) for a short functional description. Additional email support will be provided for new users.

Changes from previous versions

  • Since 2.0
    • The installer and the application have been been brought into full compliance with Windows Vista.
    • A new tool allows the user to view a Fourier-transformed image of the data.
    • A family of "mask" features allows the user to filter out bad or unwanted pixels from plots and calculations.
    • A new "Movie" feature has been added in the Image window. Each time you read in a new file, you can save the false color image, and then rapidly play them all back in sequence. This may be helpful in picking out small changes between different files (for example, by alternating rapidly between 2 files) or in following gradual changes as a function of temperature, etc.
    • The maximum number of possible fitting submodels has been increased to 22.
    • The previous "Image" window was split into "Image" and "Manipulate" windows. The Manipulate window includes new data folding features, which allow the user to symmetrize the data about the vertical axis, horizontal axis, or center point (inversion symmetry).
    • Datasqueeze now supports multiple images in Princeton/Roper (*.spe) file formats.
    • Datasqueeze now supports file in a pure binary file (square 2-byte array, either big- or little-endian).
    • If Datasqueeze cannot open a file with the format you chose, it now tries all the others before giving up.
    • An optional "chi offset" feature has been added which simplifies analysis of fiber or other patterns that are rotated relative to the detector.
    • "Contextual Menus" have been added, so that you can right-click (control-click on Macintosh) on most components in the user interface to get more information on what that component does.
    • The newest Bruker data file format (version 11, BIS/PROTEUM compression, used by APEX and other CCD's) is now supported.
    • A new method of calibrating data files has been introduced, in which the beam center and detector range are calculated via least squares fit to a Bragg peak of known d-spacing. The previous method (visual comparison of a series of calculated Bragg rings to the data) is still available.
    • Rendering of 1D plots from the 2D data is now substantially faster in many cases.
    • It is now possible to import one dimensional (x-y) data. This means, for example, that you could open a data file, save your x-y plot in a text file, modify the file (for example by removing "bad points " or merging two data sets), re-import the data, and continue least-squares or other analysis.
    • When you save the 1D data from a plot, the calculated uncertainty (error) is now written in the third column.
    • If you accidentally start to exit the program (for example by clicking the "Close Window" box or ^Q, you now have the option of "Cancelling" this request and continuing with your analysis.
    • A new graphical indexing feature was added, which allows the user to compare the peaks in a powder diffraction pattern to selected Bravais lattices with variable lattice parameters.
    • Bessel functions and squared Bessel Functions were added to the list of least-squares fitting models.
    • Support was added for B/W Tiff formats, such as those used for CCD detectors at CHESS.
    • A new "elliptical" plot option was added, in which the azimuthal (polar) intensity is averaged along an elliptical rather than a circular path. This may be useful for the analysis of asymmetric patterns due to sample strain, etc.
    • In the Plot window, the plot type selection was changed from radio buttons to a drop-down menu.
    • The behavior of the false color image with small floating point formats was improved.
    • The Ditabis IPC, Gatan DM3, and XDR formats formats were added.
    • The RaxisIV++ Format was added to reflect changes in the Rigaku Raxis data format.
    • A bug was corrected in the Windows version; the bug had prevented Datasqueeze from "remembering" the working directory.
    • The user now has the option of superimposing a grid (rectangular or polar) on top of the false color image to better locate the positions of interesting features in the data.
    • When writing the false color image as a graphics file or copying it on to the clipboard, the user now has the option of keeping any "overlays" (a grid, the x at the center, or the region selected for plotting).
    • The user now has the option of writing the file name directly on the false color image--useful for keeping track of images when printed or pasted to another application.
    • A variety of small bugs were corrected.
  • Since 1.8
    • Performance: memory management was improved and there were a variety of performance enhancements.
    • Image: the interface was made easier to use, more options for viewing the data were added.
    • Interface: tabbing was improved in conformity with normal Windows and Macintosh usage; the screen appearance, particularly on Windows XP, was improved. Cut-and-paste support was enhanced.
    • Calculations: Plotting options were expanded, fitting functions were added.
    • Operating system interface: It is now possible to start Datasqueeze and open data files by double-clicking (Windows and Macintosh) or to provide Datasqueeze with a list of files to open upon initialization (Linux).
    • Documentation: A Manual, in pdf form, is now provided with the distribution package.
    • System requirements: This version of Datasqueeze requires Java 1.4. If you are using 1.3 or below you will have to upgrade, which is now possible on all Java-enabled platforms (including Macintosh). Instructions for doing this are on the download page.
    • Data file formats: Support for Fuji Image Plate files was enhanced. Support was added for the ESRF (European Synchrotron Radiation Facility), Molecular Metrology, and Bruker-Nonius Kappa-CD (.kcd) data file formats. It is now possible to read in PNG, JPG, or GIF grayscale images as data files.
    • Bugs: A variety of small bugs were fixed.

Last updated December 30, 2007

Send Us Your Comments