Tutorial, how to get an image with XDim What you see: When XDim starts only a tool box will appear, most of these tools will have an effect only if an ACTIVE VIEWPORT is available. Only two icons are different, located at the bottom of the tool box. - An Icon looking like a rainbow => preference window will pop up - An Icon looking like cards with a grid drawn on top of each other: => A NEW VIEWPORT IS CREATED ************************** The mouse: XDim was designed to work with a three button mouse: 1. Button: start action 2. Button: force a normal redraw (for an explanation of a normal redraw start help in the tool box): 3. Button: test value, in most windows XDim allows you to get the coordinates and value of the data under the mouse pointer without starting an action. How to cook the first image: 0. If you want to try importing an gif/jpeg picture first, PLEASE load the 00picview.prf preference file from the examples subdirectory. Also note that XDim was NOT designed to be a picture viewer, XDim will ALLWAYS use a 3D representation of the picture => 1024x1024 image needs 2*1024*1024 triangles 3d-transformed AND every triangle may include more than one pixel (p.ex. 10 => 20 971 520 caculations in 3d Space). Keep this in mind when you compare XDim with image viewers like xv. Some hints to keep in mind with real life pictures: -TO ACTUALLY SEE THE IMAGE FORCE A NORMAL REDRAW WITH THE MIDDLE MOUSE BUTTON. -The ordering of color information to height uses the energy of the color, (similar to black and white) this is purely RANDOM, because there is NO natural order scheme for colors. -This energy ordering gives for some pictures a height relation that actually looks like a real 3d objekt. "Good" images are: monocolored background, point lighting source (so energy will fade with depth). This is the classic portrait situation. -For some images there will be practically no relationship between what XDim shows and its "real" 3d representation (think of a colored wallpaper). -For all images is true, START with NO rotation angle => look at it straight from above => should look like in a standard image viewer => try small rotation angles and REDUCE THE Z-SCALING to a small value to get good (=interpretable) results. -The 00picview.prf does the things suggested for importing pictures in Nr. 6. -Check out the effect of color interpolation on image presentation: import jpeg/testimg.gif force a normal redraw (middle mouse button) switch color interpolation on and off (Setup -> Preferences -> Preferences -> Misc -> Color Interpolation) the interpolated image will look strange/blurred load the examples/blacknwhite.prf preference file the image (after a normal redraw) should look like a black and white now switch color interpolation on and off again this time the interpolated image will look better zoom in (perhaps to the eye of the lady), now the effect of color interpolation becomes even more apparent. This happens, because there is NO natural order scheme for colors, while for black and white there is. To remove this effect, there will be a RGB separation option in the next (and probably the last release) ***PLEASE NOTE, images tend to have a few hundred pixels in x- and y-direction, therefore ALL LINE REDRAW MODES LIKE (LINES, HIDDEN, HIDDEN&FILLED ...) will be practically USELESS. They will only fill the image with a block of the line color (typ. BLACK). 1. Load a new preference file, the built in color palette was designed for systems with 16 colors, a good start is rainbow.prf. To see the color palette, pop up the preference window (Depending on your hardware, the internal representation of colors may differ from what you see, BUT exported images will always use the internal representation). To load a preference file use File/preferences from the tool box or File/Load in the preference window. Example preference files are in the examples subdirectory. 2. Create a viewport and load a file (p.ex. bfield.xd) from the examples subdirectory. To load a file use File/Load in the viewport window. After the file is loaded a GRID should appear on in the viewport showing the area of your data set. 3. Toy around with with the Icons in the Tool Box, they should rotate/move etc. the data set in the window. If screen redraw is very slow p.ex. if you use XDim over a network use Tool Box Setup/Fast Redraw and set it to lines or grid, this should fasten things up. To do a multiple move just LEAVE THE BUTTON PRESSED a timer is used to determine the next redraw request. 4. Use Button 2 (Middle Button) of the mouse over the drawing area of the viewport or Data/Redraw to force a normal redraw. A colored representation of the data will appear after some time, the look of this representation is changed in the Tool Box with Setup/Normal Redraw. 5. It is a good idea to use help in the different windows to get a hint of the possibilities of XDim p.ex. it is possible to change to height->color decoding in every viewport separately. 6. Try to import a picture p.ex. testimg.gif from the jpeg subdirectory. To make pictures like gif's etc. look good, you will probably have to turn off the color interpolation, which is in the preference window (rainbow icon), menu preferences->misc->interpolate colors. This is due to the effect of color interpolation. Exception are grayscale pictures that will actally look better with color interpolation turned on. 7. Try to import a data set from your favorite table calculation program or spread sheet. 8. Load the small.xd data set and interpolate it with Data/Size to a new size either linear or with spline. 9. Create a complex data field, interesting effects are produced by using testimg.gif as the imaginary part and small.xd as the real part (after interpolation to the same size as testimg.gif). Use the complex redraw modes from Setup/Normal Redraw. 10.Try the complex example bfx/bfy.xd. These two files are linked together as real/imaginary part. This enables you to test the complex redraw modes. 11.XDim was not designed to visualize 1-D data sets, but it is possible by setting width or height to 1 (NOT 0, see p.ex. line.xd). But in this case you will have to rotate to data set to see something usefull. ....