Welcome to XImtool V1.1
XImtool is an image display server developed by the IRAF Project at the
National Optical Astronomy Observatories. To view images you need
client software (such as IRAF) to load images into the display, or it can
load images directly when run as a standalone task. XImtool is
interchangeable with older display servers such as SAOimage /
IMTOOL and with newer servers like SAOtng, but offers many new
features not available elsewhere.
More detailed help is available on the following topics:
- Basic Usage:
- Advanced Features:
Table of Contents:
Getting Started
GUI Overview
Mouse Operations
Keystroke Accelerators
Command-line Options
Client Connections
Frame Buffers
Markers
Panner Marker
Coords Box Marker
General Markers
Menu Options
Control Panel
View Controls
Enhancement Controls
Blink Controls
Options:
Autoscale
Antialiasing
Tile Frames
Warnings
Colormap Selection
Builtin Colormaps
User-defined Colormaps
Load Panel
Directory browsing
File Patterns
Direct File Load
Frame Selections
Save Panel
File Name
Format
Color
Print Panel
Postscript Options
Color Options
Processing Options
Printer selection
Info Panel
TclShell
As a display server, XImtool is started as a separate process from client
software such as IRAF. Once it is running it will accept
client connections simultaneously on fifo pipes, unix
domain sockets, or inet sockets. A display client like the IRAF DISPLAY
task makes a connection and sends the image across using an IIS protocol
(other/different protocols may be supported in the future). Once the image
is loaded in the display buffer it may be enhanced,
saved to a disk file in a number of different formats, or
printed as Encapsulated Postscript to a printer or disk file.
When run in standalone mode, images may be loaded on the
command line or by using the Load Panel.
This allows you to browse images and perform the same manipulations as if
they had been displayed by a client.
The GUI consists of a large image display window and a number of smaller
panels that control various specific functions such as image
Load, Save and Print
as well as a general purpose Control Panel. The main
window menubar has several menu buttons to the left: the Files menu
is used to load/save/print an image as well as quit the task. The View
menu let's you select the image orientation, zoom, colormap or frame. The
Options menu allows you to call up control panels, toggle markers
or blinking etc. Some of this functionality is duplicated elsewhere in
the GUI. The right side of the menubar contains command buttons to flip the
image as well as buttons for frame selection and the help button.
For more detailed information on the operation of the control panels please
see the on-line help (i.e. use the '?' button or Alt-h keystroke in the
main image window).
Clicking and dragging MB1 (mouse button 1) in the main image
window creates a rectangular region marker, used
to select a region of the image. If you do this accidentally and don't
want the marker, put the pointer in the marker and type DELETE or
BACKSPACE to delete the marker. With the pointer in the marker,
MB3 will call up a marker menu listing some things
you can do with the marker, like zoom the outlined region. MB1 can be used
to drag or resize the marker. See below for more
information on markers.
Clicking on MB2 in the main image window pans (one click) or zooms (two
clicks) the image. Further clicks cycle through the builtin zoom factors.
Moving the pointer to a new location and clicking moves the feature under
the pointer to the center of the display window. Holding down the Shift
key while clicking MB2 will cause a full-screen crosshair cursor to appear
until the button is released, this can be useful for fine positioning of the
cursor.
MB3 is used to adjust the contrast and brightness of the displayed image.
The position of the pointer within the display window determines the
contrast and brightness values. Click once to set the values corresponding
to the pointer location, or click and drag to continuously adjust the display.
The following keystrokes are currently defined in the GUI:
Ctrl-b Backward frame Alt-b Blink frames (toggle)
Ctrl-c Center frame? Alt-c Control panel
Ctrl-f Forward frame Alt-h Help
Ctrl-i Invert? Alt-i Info box popup
Ctrl-m Match LUTs Alt-l Load file popup
Ctrl-n Normalize Alt-p Print popup
Ctrl-p Print Alt-s Save popup
Ctrl-r Register Alt-t TclShell popup
Ctrl-t Tile frames toggle
Ctrl-u Unzoom (zoom=1)
Ctrl-x Flip X Ctrl-Alt-q Quit
Ctrl-y Flip Y Ctrl-Alt-f Fitframe
Ctrl-= Print
Ctrl-< Decrease blink rate Ctrl-+ Zoom in
Ctrl-> Increase blink rate Ctrl-- Zoom out
Alt-1 thru Alt-4 Set frame displayed
Ctrl-1 thru Ctrl-9 Set integer zoom factor
NOTE: These keystrokes only work with the cursor in the main image window,
not on the subwindows or in markers.
Ximtool allows clients to connect in any of the following ways:
fifo pipes
The traditional approach. The default, global /dev/imt1[io] pipes may
be used, or a private set of fifos.
tcp/ip socket
Clients connect via a tcp/ip socket. There is a default port, or a
custom port may be specified. This permits connecting to the server over a
remote network connection anywhere on the Internet.
unix domain socket
Like a tcp/ip socket, but limited to a single host system. Usually
faster than a tcp/ip socket, and comparable to a fifo. By default each user
gets their own unix domain socket, so this option allows multiple users
to run ximtools on the same host without having to customize things.
By default ximtool listens simultaneously for client connctions on all three
types of ports. Clients communicate with XImtool using the IIS protocol,
other protocols may be supported in the future.
XImtool starts up using default frame buffer of 512x512 pixels. When loading
disk images the frame buffer configuration file will be searched for a
defined frame buffer that is the same size or larger than the current image,
when used as a display server the frame buffer configuration number is passed
in by the client. The default file used is /usr/local/lib/imtoolrc, this can
be overridden by defining a IMTOOLRC environment variable naming the
file to be used, or by creating a .imtoolrc file in your home
directory.
The format of the frame buffer configuration file is
configno nframes width height [extra fields]
e.g.
1 2 512 512
2 2 800 800
3 1 1024 1024 # comment
At most 128 frame buffer sizes may be defined.
The following command-line options are currently recognized:
-basePixel < num > Base colormap pixel number
-cmap1 < file > User cmap 1
-cmap2 < file > User cmap 2
-cmapDir1 < dir > User cmapDir 1
-cmapDir1 < dir > User cmapDir 2
-cmapInitialize < bool > Initialize colormap at startup
-cmapName < name > Private colormap name
-config < num > Initial config number
-defgui Print default GUI to stdout
-displayPanner < bool > Display panner box
-displayCoords < bool > Display wcs coords box
-fifo < pipe > Fifo pipe to use
-fifo_only Use fifo pipes only
-gui < file > GUI file to use
-help Print command-line summary
-imtoolrc < file > Frame buffer configuration file
-inet_only Use inet sockets only
-invert Invert colormap on startup?
-maxColors < num > Number of colors
-memModel < type > Memory model (fast,small,beNiceToServer)
-nframes < num > Number of frames at startup
-port < num > Inet port to use
-printConfig < file > Printer configuration file
-port_only Use inet sockets only
-tile Tile frames on startup?
-unix < name > Unix socket to use
-unix_only Use unix sockets only
< file > File to load on startup
The panner window always displays the full frame buffer. Try setting the
frame buffer configuration to a nonsquare frame buffer (e.g. imtcryo) and
then displaying a square image (e.g. dev$pix) and the panner will show you
exactly where the image has been loaded into the frame.
The panner window uses two markers, one for the window border and one to
mark the displayed region of the frame. Most of the usual marker keystrokes
mentioned below apply to these markers as well, e.g.
you can use MB1 to reposition on the panner window within the main image
display window, or to drag the region marker within the panner (pan the
image). Resizing the region marker zooms the image; this is a non-aspect
constrained zoom. The panner window itself can be resized by dragging a
corner with MB1. Typing delete or backspace anywhere in the panner window
deletes the panner.
A special case is MB2. Hitting MB2 anywhere in the panner window pans the
image to that point. This is analogous to typing MB2 in the main display
window to pan the image.
Ximtool provides a limited notion of world coordinates, allowing frame
buffer pixel coordinates and pixel values to be converted to some arbitrary
client defined coordinate system. The coords box feature is used to display
these world coordinates as the pointer is moved about in the image window.
The quantities displayed in the coords box are X, Y, and Z: the X,Y world
coordinates of the pointer, and Z, the world equivalent of the pixel value
under the pointer. All coordinate systems are linear. The precision of a
displayed quantity is limited by the range of values of the associated raw
frame buffer value. For example, if the display window is 512x512 only 512
coordinate values are possible in either axis (the positional precision can
be increased however by zooming the image). More seriously, at most about
200 pixel values can be displayed since this is the limit on the range of
pixel values loaded into the frame buffer. If a display pixel is saturated
a "+" will be displayed after the intensity value.
The coords box is a marker (text marker) and it can be moved and resized
with the pointer like any other marker.
Although ximtool doesn't do much with markers currently, they are a general
feature of the Gterm widget and are used more extensively in other programs
(e.g. the prototype IRAF science GUI applications). Ximtool uses markers
for the marker zoom feature discussed above, and also for the
panner and the coords box. All
markers share some of the same characteristics, so it is worthwhile learning
basic marker manipulation keystrokes.
- MB1 anywhere inside a marker may be used to drag the marker.
- MB1 near a marker corner or edge, depending on the type of marker,
resizes the marker.
- Shift-MB1 on the corner of most markers will rotate the marker.
- Markers stack, if you have several markers and you put one on top
of the other. The active marker is highlighted to tell you which of the
stacked markers is active. If the markers overlap, this will be marker
"on top" in the stacking order.
- MB2 in the body of a marker "lowers" the marker, i.e. moves it to
the bottom of the stacking order.
- Delete or backspace in a marker deletes it.
- Markers have their own translation resources and so the default
keystroke commands will not be recognized when the
cursor is in a marker.
For example, try placing the pointer anywhere in the coords box, then press
MB1 and hold it down, and drag the coords box marker somewhere else on the
screen. You can also resize the coords box by dragging a corner, or delete
it with the delete or backspace key. (The Initialize button will get the
original coords box back if you delete it).
- MB3 (mouse button 3) calls up the marker menu (by default).
- Zoom does an equal aspect zoom of the region outlined by the marker.
In this way you can mark a region of the image and zoom it up.
- Fill exactly zooms the area outlined by the marker, making it fill
the display window. Since the marker is not likely to be exactly square,
the aspect ratio of the resultant image will not be unitary.
- Print prints the region outlined by the marker to the printer or
file currently configured by the Print Panel.
- Save saves the region outlined by the marker to the file currently
configured by the Save Panel.
- Info prints a description of the marked region. The text is
printed in the Info Panel.
- Unrotate unrotates a rotated marker.
- Color is a menu of possible marker colors.
- Type is a menu of possible marker types. This is still a little
buggy and it isn't very useful, but you can use it to play with different
types of markers.
- Destroy destroys the marker. You can also hit the delete or
backspace key in a marker to destroy the marker.
The Frame box will list only the frame buffers you currently have
defined. Currently, the only way to destroy a frame buffer is to change the
frame buffer configuration, new frame buffers (up to 4) will be created
automatically if requested by the client.
The text display window gives the field X,Y center, X,Y scale
factors, and the X,Y zoom factors. The scale factor and the zoom factor
will be the same unless autoscale is enabled. The scale is in units of
display pixels per frame buffer pixel, and is an absolute measure (it doesn't
matter whether or not autoscale is enabled). Zoom is
relative to the autoscale factor, which is 1.0 if autoscaling is disabled.
This information is also presented in the Info panel.
The numbers in the Zoom box are zoom factors. Blue numbers zoom,
red numbers dezoom. Zoom In and Zoom Out may be used to go to
larger or smaller zoom factors, e.g. "Ctrl-5" followed by "Zoom In" will get you
to zoom factor 10. Specific zoom factors may also be accessed directly as
Control keystrokes, e.g. Ctrl-5 will set zoom factor 5.
Center centers the field. Toggle Zoom toggles between the
current zoom/center values, and the unzoomed image.
Aspect recomputes the view so that the aspect ratio is 1.0.
Aspect also integerizes the zoom factor (use the version in the View menu
if you don't want integerization).
Fit Frame makes the display window the same size as the frame
buffer. Note that autoscale has much the same effect,
and allows you to resize the display window to any size you want, or view
images to large to fit on the screen.
At the top is a scrolled list of all the available
colormaps. Click on the one you want to load it. You can add your own
colormaps to this list.
The two sliders adjust the contrast (upper slider) and
brightness (lower slider) of the display. The Invert button
inverts the colormap (multiples the contrast by -1.0). Note that due to the
use of the private colormap the sliders are a bit sluggish when dragged to
window the display. If this is annoying, using MB3 in the display window is
faster.
The Normalize button (on the bottom of the control panel) will
normalize the enhancement, i.e. set the contrast and brightness to the default
one-to-one values (1.0, 0.5). This is the preferred setting for many of the
pseudocolor colortables and for private colormaps loaded from disk images.
- Blink frames is the list of frames to be blinked. When blink
mode is in effect ximtool just cycles through these frames endlessly, pausing
"blink rate" seconds between each frame. The same frame can be entered in
the list more than once. To program an arbitrary list of blink frames, hit
the Reset button and click on each blink frame button until it is set
to the desired frame number.
- The Blink Rate can be adjusted as slow or as fast as you want
using the arrow buttons. If you set the blink rate small enough it will go
to zero, enabling single step mode (see below).
- The Register button registers all the blink frames with the current
display frame. Frames not in the blink list are not affected.
- The Match LUTs button sets the enhancement of all blink frames to
the same values as the display frame. Frames not in the blink list are not
affected.
- The Blink button turns blink on and off. When the blink rate is
set to zero the Blink button will single step through the blink frames, one
frame per button press.
NOTE: you can blink no matter what ximtool options are in effect, but many
of these will slow blink down. To get the fastest blink you may want to
turn off the panner and coords box, and match the LUTs of all the blink
frames. All the ximtool controls are fully active during blink mode, plus
you can load frames etc.
- Autoscale
- If autoscale is enabled then at zoom=1, the frame buffer will be
automatically scaled to fit within the display window. With autoscale
disabled (the default), the image scale is more predictable, but the image
may be clipped by the display window, or may not fill the display window.
- Antialiasing
- When dezooming an image, i.e., displaying a large image in a smaller
display window, antialiasing causes all the data to be used to compute the
displayed image. If antialiasing is disabled then image is subsampled to
compute the displayed image. Antialiasing can prevent subsampling from
omitting image features that don't fall in the sample grid, but it is
significantly slower than dezooming via subsampling. The default is no
antialising.
- Tile Frames
- The default display mode is to view one frame at a time. In tile frames
mode, 2 or 4 frames may be viewed simultaneously in the display window. All
the usual operations (zoom and pan, colortable enhancement, cursor readback,
etc.) still work for each frame even when in tile frames mode.
- Warnings
- The warnings options toggles whether you see warning dialog boxes in
situations like overwriting an existing file, clearing the frame buffer, etc.
By default XImtool will display images using either a grayscale colormap
if loaded by a client, or a private colormap when loading an image from
disk that contains a colormap. Each frame defines its own colormap so
you can define different colormaps or enhancements for each frame, they
will change automatically as you cycle through the frames.
Once loaded, the colormap may either be changed using the builtin colormap
menu under the View menu button on the main window, or from the
Enhancement box on the control panel. Ximtool has about a dozen colormap
options builtin, other user-defined colormaps may
optionally be loaded.
The cmap[12] and cmapDir[12] resources (or command line
arguments are used to tell ximtool which specific colormaps to make
available or where to look for colortables respectively. The colortables
are loaded when ximtool starts up, or when it is reinitialized (e.g. by
pressing the Initialize button in the control
panel). Ximtool will ignore any files in the colormap directory
which do not look like colortables. New colortables will also be added
for each images loaded from disk.
The format of a user lookup table is very simple: each row defines one
colortable entry, and consists of three columns defining the red, green,
and blue values scaled to the range 0.0 (off) to 1.0 (full intensity).
R G B
R G B
(etc.)
Blank lines and comment lines (# ...) are ignored.
Usually 256 rows are provided, but the number may actually be anything in
the range 1 to 256. Ximtool will interpolate the table as necessary to
compute the colortable values used in Ximtool. Ximtool uses at most 201
colors to render pixel data, so it is usually necessary to interpolate the
table when it is loaded.
The name of the colortable as it will appear in the Ximtool control panel
is the root name of the file, e.g., if the file is "rainbow.lut" the
colortable name will be "rainbow". Lower case names are suggested to avoid
name collisions with the builtin colortables. Private colormaps for disk
images will be have the same name as the image loaded. If the same colortable
file appears in multiple user colortable directories, the first one will be
used.
The directory "luts" in the ximtool source directory contains a sample set
of colortable files. This can be installed as /usr/local/lib/imtoolcmap
when ximtool is installed.
The Load Panel allows you load images from disk directly to the frame buffer,
this is analogous to loading an image on the command line except that
browsing is possible. At present recognized formats include IRAF OIF format
(i.e. .imh extension), simple FITS files, GIF, and Sun rasterfiles. The
task will automatically sense the format of the image and load it
appropriately. Images with private colormaps (such as GIF) will be loaded
using the private colormap by default (meaning that changing the
brightness/contrast enhancements will render a random-colored image). If
the Grayscale button is enabled the image will be converted to
grayscale and loaded with the standard grayscale colormap.
When loading new images the frame buffer configuration table
(imtoolrc) will be searched for a frame buffer that is the same size
or larger than the new image size, if no frame buffer can be found a custom
buffer exactly the size of the image will be created. This means that the
image may not fill the display window when loaded, or you may see a subsection
of the image in the main display window. Setting the
autoscale option will scale the entire image to fit
the main display window.
Images with more colors than can be displayed will automatically be quantized
to the number of available colors before display. Formats which allow more
than 8-bit pixels will be sampled to determine an optimal range in the data
to be used to compute the transformation to the number of display colors.
This is the same transformation used by the IRAF DISPLAY task.
- Directory browsing
-
The load panel contains a list of files in the current directory that may
be selected for loading by selecting with left mouse button. If the file
is a directory the contents of the new directory will be loaded, if it's
a plain file an attempt will be made to load it as an image. Directories
in the list are identified with a trailing '/' character, you will always
see any directories available even if a filter is
specified.
The Root button will reset the current directory to the system root
directory. The Home button will reset the current directory to the
user's login directory, the Up button moves up one directory level,
and Rescan reloads the file list by rescanning the directory. The
current working directory is given below the file selection window.
- File Patterns
- By default all files and directories will be listed. You may specify a
filter to e.g. select only those files with a given extension like "*.fits"
to list only files with a ".fits" extension. Directories will always be seen
in the list and are identified with a trailing '/' character. Any valid
unix pattern matching string will be recognized.
- Direct File Load
- If you know exactly which file you wish to load, you may enter its name
in the Load File text box and either hit or the Load button to
load it. An absolute or relative path name may be given, if a simple filename
is specified it will be searched for in the current working directory.
- Frame Selections
- By default images will be loaded into frame number 1, you may select a
different frame using the Frame menu button.
The Save Panel lets you save the current contents of the main display window
to a disk file (including the Panner/Coords markers, any general graphics
markers, or overlay graphics displayed by the client program). Presently,
only the contents of the main display window may be saved, there is no
facility for saving the undisplayed contents of the entire frame buffer
other than to enable the autoscale feature. A limited
number of formats are currently available, others will be added in future
versions.
- File Name
- The File Name text box allows you to enter the file name of the
saved file. A "%d" anywhere in the name will be replaced by a sequence number
allowing multiple frames to be saved with unique names.
- Format
- The Format box allows you to choose the format of the image to be
created. Not all formats are currently implemented.
- Color
- The Color box lets you choose the color type of the image to be
created. The options will change depending on the format, e.g. FITS doesn't
allow color so no color options will be allowed. Formats which allow 24-bit
images will be written using the current colormap after converting to a 24-bit
image, pseudocolor images will be written with the current colormap.
The Print Panel allows you dump the contents of the main display window as
Enacpsulated Postscript to either a named printer device or to a disk file.
The Print To selects the type of output, the Print Command
box will adjust accordingly, either as a Unix printer command or as a file
name. A "%d" anywhere in the name for disk output will be replaced by a
sequence number allowing multiple frames to be saved with unique names.
Selecting printers from the installed list will
automatically change the command to be used to generate the output. This
command does not necessarily need to be a printer command, the
printer configuration file lets you define any command
string to process the image.
The Color box lets you choose the color type of the image to be created.
PseudoColor or 24-bit postscript will be created using the current colormap.
- Orientation
- Set the page orientation.
- Paper Size
- Select the paper size to be used.
- Image Scale
- Set the scale factor used to compute the final image size.
- Auto Scale
- The auto scale toggles whether or not the image is automatically scaled
to fit the page. If not enabled, the image scale will be used to
dtermine the output image size.
- Auto Rotate
- Auto rotate determines whether or not the image will be rotated to fit
on the page. When set, an image larger than the current orientation will be
rotated and possibly scaled to fit the page.
- Max Aspect
- Max Aspect takes images smaller than the page and automatically increases
the scale so the image fills the page in the current orientation.
- Annotate
- The annotate option toggles whether or not the final file includes
annotation such as the image title, a colorbar, and axis labels.
The printer selection list lets choose the printer to be used. The printer
configuration file is /usr/local/lib/ximprint.cfg by default or may be reset
using the printConfig resource. The format of the file is simply
name < tab > command
The name value is what appears in the selection list and may be more
than a single word, the command can be any command that accepts EPS
input from a pipe, the two fields must be separated by a tab character.
Normally the command will be
a simple 'lpr -Pfoo' or some such, but can also include converters or
previewers. At most 128 printer commands may be used.
The information panel is underused at present but is meant to provide
basic information about the frame being displayed. It is updated to be
current while changing enhancements, pan/zoom regions, or frame selection.
In cases where the image title string is truncated in the main display window,
the user can always pop up the info window to see the full title.
The TclShell is mostly used as a development or debugging
tool for the GUI. It allows the user to type commands directly to the
TCL interpreter letting you send messages to the object manager or execute
specific procedures in the TCL code that makes up the GUI. Most users will
never need this, but for an example of what it does, bring it up and type a
command such as
send helpButton set background red
Cool, huh.
XImtool was developed by the IRAF Group at the National Optical
Astronomy Observatories in Tucson, AZ.
This document was last updated 11/6/96.