If you are using
Navigator 4.x
or
Internet Explorer 4.x
or
Omni Web 4.x
, this site will not render
correctly!
Vis > image_printing > GFDL Visualization Guide: BIT
GFDL Visualization Guide: BIT
NAME
bit - Bitmap Image Touchup program for SGI Workstations
SYNOPSIS
bit [options] [files] bit [ -vfabiIdenhCRrq -c file -s sec -p path -x f
-y f -m level ] [files]
DESCRIPTION
bit is an interactive full color image viewer and editor based on Silicon
Graphics GL.
As an image viewer, bit allows list of images to be viewed in their
original type (24bits or color index) in any order and in any of the many
built-in styles of display. With command line option -IN, bit simulates
and combines the functionalities of ipaste(1G), gifpaste(1L) and
jpegpaste(1L). Further, with the pan and zoom features, large images can
be viewed in full without being limited by the window or screen size.
Slideshow is also supported.
As an image editor, bit performs a large number of image editing and
processing tasks accurately and efficiently. It keeps information loss at
any stage of the editing process at an absolute minimum by performing
subpixel sampling automatically while conserving the input image type
whenever possible to avoid unnecessary promotion and/or quantization of
input images. The paint function features brushes, free hand drawing,
fill and many built-in patterns, all with 24bits color. In addition,
arbitrarily colored and sized text and simple geometric figures can be
rendered on top of an image for annotation with the option to make them
part of the bitmap or output separately to take advantage of the higher
printer resolutions.
Bit can also be used as a graphical user interface to an existing image
processing program by defining an external binding to it which in essence
turns a command line oriented program into a bit subfunctions and can be
accessed by a press of a button, and best of all, the processed image
will be read back and displayed immediately. Convolution using externally
defined matrices of arbitrary order can be performed dynamically giving
great flexibility in processing an image.
Upon startup, bit does the following and then waits for further command:
1. Reads the initial window position specification file WinPosition from
user directory /.bitrc unless the command line option -a (Auto),
-F (Full screen) -i (fit first Image) or -I (always fit Image) are
present. See Configuration for details.
2. Reads the configuration file Setup.opt from system directory
/usr/local/lib/bit and user directory /.bitrc in that order.
This step is skipped or modified by the command line option -n, -p or
-c.
3. Reads the command line filenames in order if present, and placing
these files on a list (the FileList) for future use.
3a.
Opens a tiny window on the upper-right corner of the screen with a
clock and a mailbox if flag -C is present.
4. Opens the main window, i.e., the window in which images will be
displayed, and clears the window to a pre-determined background
color. The window position and size are read from the file
WinPosition if -i, -I, -F and -a are not present.
5. Opens the file control panel (the FilePanel) on the lower-right
corner of the screen (or the position specified in WinPosition file
if flag -N is not present. The FilePanel is the most basic control
panel, consisting of command buttons to set program options and to
handle file IO.
At this point, if the FileList is not empty, the first file on the list
will be loaded and displayed in the main window.
To display next file on the list, click leftmouse in the main window or
press the SPACE key;
To display the previous file, click rightmouse in the main window or
press the p key;
And to display the current image (again), click middlemouse in the main
window or press the c key.
Besides adding files to the FileList via command line arguments, files
can also be added or removed from the FileList by the FilePanel button
labeled load, See Command Summaries for details.
To pause an indefinite interval at any point, press PAUSE key. To re-
activate, press PAUSE again.
To exit bit, click the Quit button on FilePanel or press <ESC> key
anywhere.
FEATURES
Bit performs all of its image manipulations interactively with a
continuous progress report, and once the processing is complete, the
processed image will be displayed. In most cases, transformation is
reversible and a simple undo will recover the un-processed image.
Current features include
o Rotates an image by an arbitrary angle with the option of anti-
aliasing and filling the "rotated out" region with an arbitrary color.
o Scales an image by arbitrary factors in x- and/or y-directions with
the option of subpixel sampling. It can also scale the image to any
requested size while keeping the aspect ratio by filling the empty
region with an arbitrary color.
o Renders text of arbitrary color, size or orientation on top of an
image with the option of rendering directly into the raster or saving
the text and raster separately, such as into a PostScript file, to
take advantage of a higher printer resolution. Besides normal text
strings, many predefined special symbols can be input conveniently as
in TeX(1), e.g., $\alpha$\small$\beta$ produces Greek alphabet a
followed by a slightly smaller Greek alphabet b Or you can input the
character code directly if the font encoding scheme is known, e.g.,
for PostScript Symbol font, $ã$ produces the copyright symbol
o Show and manipulate the histogram of an image, including histogram
equalization and direct histogram specification (not available in this
release).
o Performs arbitrary 1-to-1 pixel transformation. The transformation
function can be of arbitrary form and can be specified interactively
by mouse or by arbitrary math expressions. Further, the
transformation can be applied to RGB channels separately or
simultaneously and to the entire image or a portion of it.
o Modify the RGB values of an individual pixel or a group of pixels.
o Places simple geometric figures (SGFs), e.g., arrows, circles,
triangles etc., of arbitrary color, size and orientation into an image
at arbitrary locations addressed either in absolute pixels or in some
user definable coordinate system units, with the option of rendering
directly into the raster or save the sgfs and raster separately (e.g,
PostScript file). All SGFs can be scaled in x- and y-directions
independently or simultaneously (this is how to get an ellipse from a
circle).
o Full color paint with many built-in tools, which include brushes of
many different shapes, air brush with adjustable nozzle size and flow
rate, free-hand polygon, free-hand drawing, fill and many more.
o Crops a rectangular portion of an image, or add a solid frame to an
image. The crop region can be larger than the window/screen size.
o Merges and/or concatenates several images to form a single image, with
optional add/sub/diff/xor/mask operations on the fly.
o Cuts and pastes across images.
o Computes the pixel by pixel difference between two identically sized
images and displays the results as an image, whose histogram provides
quantitative information.
o Performs edge detection. In combination with image histograms and 1-
to-1 transformation, very accurate result can be obtained.
o Performs convolution with externally defined (square) matrices of
arbitrary order.
o Smoothes an image or a portion of it with thresholding.
o Sharpens an image or a portion of it.
o Median filtering an image or a portion of it.
o Magnifies any portion of an image by any amount.
o Displays a list of images in sequence with a user specifiable pause
interval between images. (SlideShow)
o Measures RGB intensities of a pixel at any location or all pixels
along an arbitrary line with the result presented as RGB intensities
or as an intensity vs. distance plot.
o Performs image type conversions.
o Converts image formats.
o Filters an image by external programs and read the filtered image back
and display it.
o Performs FFT on the entire image or a portion of it and display the
resulting power spectra.
OPTIONS
The following command line options are interpreted by bit program and
override any program defaults, including options set in the start up
file.
-v Prints the current Version number of the bit program and exits. No
graphics required.
-f Prints an enumerated list of all supported image Formats. No
graphics required.
-a By default, bit searches file ~/.bitrc/WinPosition for initial
window positions. This flag sets initial location to Automatic,
which is about 80% of the screen size.
-i This flag forces the initial window size to be that of the first
image.
-I This flag forces the window size to be that of the current image.
-b prevents bit from running in Background so that compound command
like bit; other_command can be guaranteed to run in sequence.
-F starts bit with main window occupying the entire screen(FullScreen).
-N starts bit with No initial control panels. Once bit started, you can
get the FilePanel with m key.
-C starts bit with Clockmailbox window.
-e Brings up the Editing control panel automatically upon startup.
-n Instructs bit Not to read initialization file Setup.opt.
-c filename
Changes the initialization file name to filename. Default is
Setup.opt.
-h Prints a brief Help message summarizing all command line options and
what they do.
-d ds
Specifies the Display style. Valid styles are numbered from 0 to 16.
Style 100 is magic in that it will cycle through all styles.
-s sec
Enables SlideShow mode where images are displayed one by one with
sec being the pause time, in seconds, between each display. Due to
resolution limit of the system and various system delays, a value
less than 0.1 second is not terribly reliable (you can specify a
value less than 0.1 second, however). Note any keyboard or mouse
input takes precedence over the pause time specified, and in
particular, keyboard s turns off the SlideShow.
-R Suppress progress report.
-r Instructs the bit program not to Refresh screen before displaying an
image. The default is to refresh the screen.
-m level
Sets Message output level to level, with level loosely defined
between 0 and 4. Level 0 suppresses all messages except for error
messages; Level 1 will allow warning messages to be printed in
addition to error messages; level 2 may be used to print some
additional information, such as values of key parameters or
transparent and recoverable failures. Level 3 and 4 are for debug
and trace information. Note that level 3 and 4 may or may not be
available depending on the installation. Default message level is 1,
i.e., print error as well as warnings messages.
-q Quiet mode, same as -m0.
-x fx
Where fx specifies the zoom factor in the x direction. Only the
integral part will be used. A zero can be used to signal "AutoZoom"
in which the zoomed x size will fit the screen and in addition,
automatically sets the y zoom factor as to keep the aspect ratio.
-y fy
Vertical zoom factor.
-p dir
Replaces the default Path, ~/.bitrc, with dir for all user files,
e.g., configuration, external bindings, and symbol definition etc.
Note this flag does not prevent bit from searching the system
default path, /usr/local/lib/bit, for installation specific options.
Important Commands
NOTE Typical manuals would give detailed command summaries here. Due to
the length of the command summary and availability of the context
sensitive online help in addition to the LaTeX document for bit, the
command summary is presented in a later section so that more
important comments can be given for easy reference.
Online Help
In almost all editing modes, up-to-date help is available, typically
by selecting pop-up menu entry help if that particular mode is
controlled by the pop-up or by clicking the mouse on the empty
region of the control panel where no other objects such as buttons
etc exist that might interact with the mouse.
Online help is presented in a form of a browser. You can scroll
through the page by the mouse. In help mode, all control panels are
deactivated and no interaction can take place until you exit from
help mode by pressing the button on the title of the browser.
Special Handling
ClockMail interacts with the bit program through user signal 2. This
interaction is provided for emergencies (whatever that might be) or
when bit is not responsive, press <ESC> inside the ClockMail panel
kills all bit related processes cleanly. Pressing Q inside the
panel removes this panel.
Loading/Removing files
clicking mouse on the load button in the FilePanel will bring up a
browser filled with files in the current directory that match the
current selection pattern, with highlights on files that are
currently on the FileList. To add or remove files, click the mouse
on the filenames. You can walk the directories trees, either by
selecting the desired directory (marked with a D in front) in the
browser or by typing the directory name into the box labeled dir.
(you might need to activate the input field by clicking the mouse on
it first). Files in the browser will be updated to reflect the
directory change. If the files to be added or removed from the
FileList have common patterns, the pat input comes in handy: Simply
type into the pat box the common pattern (regular expressions) and
only filenames that match the pattern will be shown in the browser.
To select all files shown in the browser, click the mouse on the
LoadAll button and to de-select all files in the browser, click the
mouse on the UnloadAll button. Click Done button to exit file
selection. This will be enough to get you started viewing files in
any order.
Generating Output
To write current image to disk, clicking the Write button in the
FilePanel will bring up a panel listing all bit supported formats
and a browser filled with all files in the current directory
(subject to pattern matching with the current pattern). The default
filename is the current filename with a possible different
extension. Filename can be changed either by selecting it from the
browser or by typing it into the field labeled filename.
To choose the desired format to write, click the mouse on the button
that has the format name on it. Upon each new selection, several
things will take place: First, the filename is modified to reflect
the current selection by adding or modifying the extension, and if
it already exists, a highlight is added to the name in the browser
to warn about possible overwriting; Next, current settings for
writing, if any, are shown in the box labeled options. If there is
none, the option box shows the word "NONE". To change default
settings, press the mouse on the option box. Note the options
mentioned here are those that are specific to an image format, such
as the Q-value in jpeg or interlace in gif etc. Image type
conversion is transparent and taken care of automatically, e.g.,
requesting a jpeg with a gray image (either in colormap or rgb
representation) will produce a single channel jpeg file with
8bit/sample.
Click the OK button to initiate output. If successful, the write
control panel disappears.
SUPPORTED FORMATS
bit currently supports the following image formats
o IRIS RGB, native to SGI IRIS
o JPEG(JFIF) format
o MPEG Movie (ReadOnly)
o CompuServ GIF
o PNM(Portable Any Map), including PPM, PGM and PBM.
o XBM (X Window Bitmap)
o TIFF
o PostScript
o Sun Rasterfile
In addition, compressed version of aforementioned formats will be
automatically decompressed and loaded.
CONFIGURATION
bit is very configurable. You can change any or all of the following
flags either interactively or by modifying the Setup.opt file directly.
Flags are case insensitive and their values can be either string or
numerical. Use colon (:) to separate the flag and its value, e.g., both
DisplayStyle: 14
and
DisplayStyle: Sparkle
are valid and select the display style sparkle.
In the following listing, the numerical value is given in parenthesis.
The best way to set up a configuration is to change the settings
interactively and then write out the current settings to a file and name
that file to be the default start-up file.
DisplayStyle
Many different styles are available to display an image. The
simplest and fastest, also the default, is the Block style, which
simply displays the image in one scoop using lrectwrite(3G) or
rectwrite(3G) depending on the image type. Other styles are
enumerated as follows:
center tab(;);
r l.
Options; Effects
Block[0];display in one scoop
Down[1];from top to bottom
Up[2];from bottom to top
Left[3];from left to right
Right[4];from right to left
DownUp[5];from top and bottom toward center
UpDown[6];from center to top and bottom
LR[7];from left and right toward center
RL[8];from center to right and left
RInt[9];RowUniform interlace
CInt[10];Column uniform interlace
Gint[11];Interlace, GIF style
Rect1[12];rectangular, towards center
Rect2[13];rectangular, from center
Sparkle[14];sparkle fade
Fade[15];blending fade
Cycle[100];cycle through all styles
Note Fade is applicable to RGB images and is ignored if the image is
in CI.
DoubleBuffer
Turn (hardware) double buffering on and off. If hardware is not
capable of double buffering, this flag has no effect. Valid options
are Off[0] and On[1].
AlwaysClearScreen
The screen by default is cleared between successive images. To leave
the screen as it is between images, toggle this flag (useful for
animation). Valid options are No[0] and Yes[1].
AlwaysUseBorder
All windows, except for the main window and the initial FilePanel,
do not have borders by default. To force a border, set this flag to
Yes. Note this flag does not take effect immediately, i.e., it does
not affect panels already on screen. Turning a panel off and then on
again will reflect the settings of this flag.
SlideShow
Enable or disable the SlideShow mode. Numerical values given should
be in seconds. 0 disables slideshow.
ProgressReport
How frequent to report progress in percentages. Built-in percentages
are 2, 5, 10 and 20. A value of 0 turns the report off. Note
however, cursor rotating can not be turned off.
PSResolution
Number of pixels to use in rasterizing PostScript files via
Ghostscript interpretor gs(1) or psrender. Default is 96DPI.
AutoPan
Controls if to pan an image automatically if it is larger than that
can be shown. Valid options are Never[0], SlidesOnly[1], and
Always[2]. SlidesOnly means to pan automatically if the current
display mode is SlideShow.
ShowCutBuffer
Controls if to show current cut buffer while looking for pasting
locations.
PreserveWMColors
Controls if to preserve the colors window manager uses to minimize
colormap flashing. bit does this by re-ordering the colormap and in
certain cases, where the color index has specific meanings, this
might not be desirable. Valid settings are No[0] and Yes[1].
AlwaysReadScrn
Cropping is normally done via internal memory copy to avoid reading
a dithered framebuffer (unless there are text and SGFs on screen and
in that case, screen read is performed and after that, the text and
SGFs are deleted). Set this flag to Yes to override the default for
special effects or for speed on better equipped machines.
AutoCropFill
Meaningful when the crop region is larger than image. For regions
outside the image, you have the option to read the screen as it is,
fill the region with a solid color, or interactively choosing one of
the two. The valid options are No[0], Ask[1] and Yes[2].
AutoRemoveIcon
Controls if to delete image browser thumbnail images when original
images no longer exist. The actual deletion could happen either at
the time original image is deleted via Delete button on ControlPanel
or when a directory is browsed.
QuantizationMethod
Three methods are implemented. The default is the Q256(1), which is
very effective for reducing the image colors down to 256 colors (or
very close to but less than 256 colors). General purpose but slower
quantization method is the MedianCut[2]. Octree[3] is for
demonstration purposes only. If the setting is Ask[0], the method
will be prompted interactively. Note in this pre-release, only Q256
is available. However, general quantization can be accomplished by
using external bindings to djpeg(1) or ppmquant(1)
QuantDither
Whether allow a dithering step after quantization. Valid options are
No[0] and Yes[1]. Default is Yes.
HalftoneMethod
Controls the method to use in converting images to B&W. Valid
options are Ask[0], Ordered8[1] and FS[2]. If this flag is set to
Ask, the method will be prompted interactively.
SimuDblBuffer
Turn simulated double buffering on and off. If the hardware is
capable of double buffering and DoubleBuffer is on, this flag has no
effect. Valid options are Off[0] and On[1].
SimuOPDblBuf
Turn simulated buffering in overlay/pupdraw on and off. This is only
used in Marking. If the hardware has no overlay bitplanes, the
simulation might not be correct in pupdraw.
AutoChangeExt
Change or append an extension to filename according to the requested
format when writing. Valid options are No[0] and Yes[1].
AutoRemoveFromList
If for some reason, a file can not be loaded, by default, the file
is removed from the FileList upon confirmation. This flag modifies
the default. The valid options are No[0], Ask[1] and Yes[2].
ReportMouseLocation
Controls whether to report mouse location in certain situations and
if yes, select origin to which the distance is measured. The valid
options are No[0], Window[1], Image[2]. Option Window sets the
origin at the lower-left corner of the main window while option
Image sets the origin at the lower-left corner of the image.
MessageVerbosity
Controls the amount of messages to print. Valid options are
Error[0], Warning[1], Info[2] and possibly Debug[3] and Trace[4]
depending on the compiler option when bit was built.
KeepTextAndSgfs
The duration of text and SGFs in deferred mode is by default limited
to a particular image or up to when you do a cropping. Setting this
flag to Yes overrides the default so that text and SGFs can be
carried to the next image.
DelFromDiskConfirm
When deleting files from disk, a confirmation will be prompted for
by default. To turn confirmation off, set this flag to No[0].
ShowEditAuto
Whether bring up the editing panel automatically upon startup.
Valid options are No[0] and Yes[1].
SmoothLines
Turn on and off anti-aliased line drawing in marking and paint. If
hardware is not capable, this flag has not effect. Valid options are
No[0] and Yes[1].
Initial Window Positions
Initial window position and size are read from file
/WinPosition unless command line option -a, -i, -I or -F is
present. In file WinPosition, 8 numbers should be specified. The
First 4 numbers are the main window parameters which specify the
coordinates of the lower left corner of the main window and its
width and height. The rest 4 numbers are the InfoPanel and
ControlPanel locations. Note the size of the InfoPanel and
ControlPanel can not be changed. The following is a valid
specification (# is the comment character )
10 10 800 600 # main window size 800X600 with ll corner at (10,10)
820 10 # info window location
820 180 # edit panel location
You can interactively shape and move the windows and then write out
the position via SetUp button on InfoPanel.
ENVIRONMENT VARIABLES
BITHELP The directory name where online help, icon and system wide
configuration and control files are located. Default is
/usr/local/lib/bit. All configuration and control files in this
directory are subject to overrides by those in BITDIR.
BITDIR The directory name where the personalized configuration file and
external binding files are located. The default is /.bitrc.
For any needed files, except for online help, bit will search
BITHELP first, followed by current directory or BITDIR. If
command line option -p is present, that directory or file will be
searched or read instead of BITDIR. However, current directory
will always be searched
TMPDIR The directory where temporary files reside. Default is /usr/tmp.
COMMAND SUMMARIES
Some Definitions
A mouse click, unless explicitly specified, means the press and
release of either the leftmouse button or middlemouse button.
Menubutton always refers to the rightmouse button.
Dragging the mouse means moving the mouse while one of the mouse
button is pressed.
A rubber band as used in bit refers to a simple geometric figure,
typically a rectangle, whose size, location and possibly shape
can be changed by mouse clicking or dragging.
A control panel (also known as dialog box) is a window with many
built-in objects whose functions are to modify the attributes of
an object, such as text color, or to modify task parameters, such
as the scaling factor. Some of the control panels have no borders
by default, to force a border, turn on the AlwaysUseBorder flag
either through Setup button or ~/.bitrc/Setup.opt.
Unless explicitly specified, a keyboard input means the specific
key, not the ASCII code, thus keyboard command is case
insensitive.
Editing Panels
The edit control panel, EditPanel, can be toggled on and off by
clicking the mouse on the edit button on the FilePanel. Command
line option -e can be specified to bring up the control panel
automatically upon startup.
To activate a specific editing mode, click the mouse on the
button labeled by the function.
Type Conversions
bit internally works either in RGB mode or in colormap mode
depending on the input image type. Typically, there is no need to
convert type explicitly because bit converts image types
transparently if it needs to. For unknown reasons, I decided to
have these goodies available, maybe someone might find them
useful or interesting to play with. Note that term RGB for image
type does not mean that the image is in color. Rather, it refers
to the current pixel representation. All type conversions are
initiated by clicking the mouse on the desired types shown in
EditPanel. The following is a summary of all types explicitly
supported:
Color(RGB)
Convert current image to RGB image. This is meaningful only if
current image is in colormap mode. The converted type will be in
color (as far as bit is concerned. Note that prior to version
0.74, the converted type is in Gray(RGB) if the input image is in
grayscale). Conversion from Gray(RGB) to Color(RGB) is a simple
operation that does not involve actual pixel manipulations.
However, the conversion could manifest itself when generating
output, e.g., writing a Gray(RGB) image into JPEG is very
different from writing a Color(RGB) image even if the images are
identical (except for the type).
Color(Map)
If the current image is in RGB, this conversion will initiate
quantization, whose parameters can be set interactively. If
quantization is performed for the sole purpose of saving disk
space, don't do it. JPEG is much better at that than quantization
and JPEG compression retains much more information as well.
Gray(RGB)
Convert current image to RGB gray regardless what the original
type is.
Gray(Map)
Convert current image to colormapped grayscale. Quantization will
be invoked if necessary.
B&W Convert current image to black and white. Two methods are
available, one is the ordered-dithering of order 8 (16 X 16
matrix), and the other is the Floyd-Steinberg error diffusion
method which typically gives better looking results. B&W image
can also be converted from other types by the 1-to-1
transformation with a step function.
Misc. Command
In the main EditPanel, two buttons labeled Convolve and Filters
are handles to external bindings (i.e, external programs or
matrices). To execute a binding, press the mouse on the convolve
or filter button and if there are indeed any bindings defined, a
panel contains all the definition will be shown, click on the
bindings you want to execute, and bit will try to pipe the
current image through the filter, and if successful, the results
will be read back and displayed.
The definition files must be named BIT_filters and BIT_convolve
respectively and they are searched in ~/.bitrc or whatever the
environment variable BITDIR points to, and concatenated with
possible system wide definition found in /usr/local/lib/bit or
HELPDIR with repeats removed.
Filter A filter is defined as an external program that reads at least
one of bit supported formats from file specified by argv[1] (not
stdin in order to guarantee fseekability) and outputs the
possibly transformed image to stdout. bit will intercept, load
and display the output.
All bindings must start with at (@) character and any line that
does not is considered to be comments and is silently ignored.
The format for external bindings is as follows:
@label; program ; options; format1; [format2;[format3;] ..]
where label is the string that will label the action button; and
program is the name of the program to run (popen(3S) will be
called to run it). options are command line options to the
program. If the option field contains a question mark (?) alone,
the option will be prompted before the program is actually
executed. Format[0-4] specifies the formats the program is
capable of reading, such as gif, ppm etc. Note that it is very
important to give the formats in decreasing generality, that is,
ppm; pgm; pbm; never pbm;pgm;ppm.
For example, the following are some sample (valid) bindings:
Shrink the current image by 20%
@ Scale0.8; pnmscale -xscale 0.8 -yscale 0.8; ppm;pgm;pbm;
Quantize an image
@ Quantize256; ppmquant; -fs 256 ; ppm; pgm;
@ Jquant256; djpeg; -GIF; jpeg
@ Jquant; djpeg -colors ; ?; jpeg
The format name can be gotten by running bit with the -f flag.
The format specifier is enclosed in parenthesis in the bit -f
output.
Convolve
The convolution matrix must be defined using the following
format:
@name; order; n1, n2, n3, ......;
where name is the button label and order is the rank of the
convolution matrix, and n is an integer representing the matrix
elements in column major, i.e.,n1 is the (0,0) element, n2 is the
(0,1) element and so on. The matrix need not be normalized.
For example, you can define the following matrix to sharpen an
image
@Sharpen; 3; 0,-1, 0, -1,10, -1, 0,-1, 0;
Magnify Magnify a rectangular region of an image by any amount, limited
only by the window size. In this mode, a moving rectangle defines
the region of interest. Press leftmouse or middlemouse to magnify
this region to the maximum window size. If left control key is
down, leftmouse click increases the zoom factor by decreasing the
size of the region of interest while middlemouse decreases the
zoom factor by increasing the size of the region of interest.
menubutton controls the pop-up menu, which can be used to set
some common zoom factors.
The above method changes the magnification factor uniformly in
the X- and Y-directions. To adjust separately the magnification
factors, use the following keyboard command:
X: Increase x magnification by decreasing region size.
x: Decrease x magnification by increasing region size.
Y: Increase x magnification by decreasing region size.
y: Decrease x magnification by increasing region size.
For global magnification, see Section Configuration and -x|y
flag.
Hide Hide the EditPanel. It can also be hidden by the Edit button on
the FilePanel.
SysMap Install the default system color map.
Info Hide the FilePanel. Useful when doing a very large crop (with
AlwaysReadScrn on). In case both the FilePanel and EditPanel are
hidden, press M key inside the main window will bring up the
FilePanel.
Select a color
To select a particular color from a colormap (if one exists) or
by choosing RGB separately. Note that this is not a separate
command with direct applications, but is indirectly used in many
mode, e.g., text, rotate, scale and others.
In RGB mode, the three sliders representing the value of Red,
Green and Blue intensities are shown and each slider is
individually adjustable. The resulting color is shown on the OK
button on the right-hand side of the panel. Some predefined
named colors, mainly from /usr/lib/X11/rgb.txt, are available by
pressing the PreDefCol button.
In colormap mode, a panel containing the current colormap with 64
entries per screen is shown and each individual entry is
represented by a colored square inside the panel. Two large
arrows on the sides of the panel can be used to scroll forward
and backward through the color map. A particular color is
selected by pressing the mouse on the square having the desired
color. Current selection is marked by a cross and is also shown
on the bottom-most button cabled Selected. The five small boxes
on the bottom of the panel show, respectively, the index into the
color map, red, blue and green intensities and luminosity.
In both RGB and colormap mode, leftmouse and rightmouse can be
used to select any pixels on screen inside the main window as the
current selected color by clicking on the pixel while holding
down the CNTRL key. The selected pixels color will be shown in
the appropriate boxes and/or buttons. The pixel picking mode is
indicated by a cross cursor. This feature can also be used to
get the intensity of a particular pixel by turning on the mouse
position reporting from the Setup menu.
Crop In crop mode, a copy of the currently displayed image is made and
all changes that occur are made to this copy until you either
save the change or exit from the crop mode. At that point, the
copy replaces the original image and all changes become
permanent. Cropping a cropped image can be accomplished by using
the save feature without exiting crop mode. The actual cropping
can be of either internal (memory copy) or external (screen
reads) operations. Screen reads are performed if there are text
and SGF's on top of the image, otherwise internal cropping is
done unless global option AlwaysReadScrn is on. In general,
screen reads should be avoided on machines with less than 24
bitplanes for normal framebuffer. Note also that a screen read
reads everything within the region of interest, as defined by the
cropping rectangle, which may be undesirable and in some cases,
you might need to hide all the control panels.
Upon entering crop mode, a rectangle with two small crosses on
the opposite corners is shown. Clicking and dragging the cross
move the corner of the rectangle and hence both the size and
location of the crop region can be changed this way. For precise
placement, the arrow keys and the keyboard key hjkl may be used
(as in VI(1) ). The corner to which the key applies is the one
closest to the mouse. This behavior can be overridden by the
shiftkey, i.e., if leftshift is down, hjkl always applies to the
upper-left corner of the rectangle regardless which corner is
last touched by the mouse. Similarly rightshift key will force
the hjkl key to apply to the upper-right corner. The number of
pixels each hjkl moves can be changed by key [1-9]. If CONTROL
key is down, the aforementioned command will move the rectangle
without changing its size. The process can be repeated until a
satisfactory crop region is found.
The size of the rubber band can also be symmetrically increased
and decreased by the keyboard key xXyX.
The size and placement of the crop region are reported
continuously on the FilePanel.
Pressing the menubutton is the main window will bring up a pop-up
menu having the following entries
DoIt discard the image outside the current rectangle and redraw the
screen. If current region is larger than the image on any sides,
a fill color will be prompted for (See Select a color). This fill
feature can be used in conjunction with frameN command to add a
border to the image. See also undo and save.
Help Detailed crop command summary, similar to the ones given here,
but might contain more up-to-date info.
RBcolor Cycle through all available colors of the rectangle. Exact number
of colors available depends on the hardware.
RBobject
Change the bounding objects. Currently only circles are
available, but cropping along the circle is yet to be
implemented.
Frame1 Enlarge the rectangle by 1 pixel on all sides. Convenient to add
a frame to the current image.
FrameN Enlarge the rectangle by n pixels on all sides. N will be
prompted for.
Auto Search the image for solid regions (regions with same color),
starting from all sides and the findings are indicated by a
properly placed and sized rectangle.
Save Save all the changes made so far and mark the current image as
the starting point for future undo.
Undo Undo cropping. Note that undo is of unlimited depth. Executing
this command will undo everything since last save.
Done Return control to main.
To enlarge an image without filling the outside region, turn off the
CropFill flag through configuration file or Setup menu. This can be very
useful in combination with cut & paste to import cut buffers or to
concatenate images.
Rotate There is no copies made in this mode, as soon as an image is
rotated, the rotated image replaces the original image as the
current image. Thus you can rotate a rotated image.
The rotation options are controlled by the rotation control
panel, which consists of a counter to change the rotation angles
(between -180 and +180 degrees measured counter-clockwise from
the horizontal direction), and other buttons for convenience or
other options. To change the rotation angle, simply click the
mouse on either the single or double arrows in the counter.
Leftarrow decrease the angle and right arrow increase the
rotation angle. Rotation of multiple of 90 degrees or flipping
about x- or y- axis are provided as shortcut buttons and can be
set by clicking the mouse on appropriate buttons. After you have
set the angle you can live with, click the mouse on the OK
button.
For rotations other than multiple of 90 degrees, a fill color can
be selected to fill the "rotated out" portions of the bounding
rectangular region. To do so, click the mouse on FillColor
button. In addition, for rotations other than multiple of 90
degrees, you can activate or deactivate subpixel sampling for
anti-aliasing. Anti-aliasing always gives better results. The
only situation where you do not want subpixel sampling is to keep
the exact colors as the unrotated image. To activate or
deactivate subpixel sampling, click the mouse on round subpixel
button. If the subpixel is turned off, the rotated image will
have the same type as the unrotated image.
For online help, click the mouse anywhere inside the rotating
control panel where there is no other objects.
Cut&Paste
The cut & paste region is defined by a rubber band whose size and
location can be modified interactively by the mouse and/or
keyboard. Use leftmouse to cut (or C key) and middlemouse (or P
key) to paste. Keyboard key uU undoes last pasting.
To change the size of the cut region, keyboard key xXyY can be
used to decrease/increase the width and height of the cut region,
key [1-9] controls the step size for each key press. The size can
also be changed by clicking the menubutton, which will bring up a
static rubber band with two crosses on it. You can change the
size and location by dragging these two crosses. Keyboard key
hjkl and the arrow keys can also be used to manipulate the rubber
band. To resume cut & paste, click away from the crosses.
The cut buffer so generated (by leftmouse) may come in one of two
ways. It may come as a result of a direct screen read or it may
come as an internal memory copy depending on the settings of the
ReadScr/Internal button. When ReadScr is active, everything that
falls within the cut region will be read into the cut buffer
while when Internal is active, the cut buffer always comes from
the image in core. Screen reads may be useful in certain
situations but in general it should be avoided on machines with
less than 24 bitplanes.
A cut buffer may also be obtained by importing an external image
in any of bit supported formats. To do so, click the Import
button. The imported image behaves exactly like a normal cut
buffer except that it can not be named, thus a new cut or new
import overwrites the current imported image.
To save a cut buffer for later pasting, click Save button. The
saved buffer will be available during the entire session of bit.
There are two different kinds of pasting, i.e., Overwrite and
mask. In Overwrite mode, the cut buffer replaces completely the
targeted image area, both on screen and in core while in mask
mode, a mask operation is performed before replacing the pixels
in the region. The mask is defined as NewPix = (PixInCutBuffer !=
MagicPix ) ? PixInCutBuffer: OriginalPix for every pixel in the
region (current MagicPix is 0). An obvious application of this
feature is to import external images as logos or signatures and
paste it into the current image.
Bit further has a Preview mode in Cut&Paste where all pasted
buffers only alter the pixels in the framebuffer and leave alone
the image in core. To alter in the framebuffer as well as
altering the image in core, toggle on the Commit flag.
Note that in Commit mode or Preview with Masking turned on, both
the cut and paste regions are not allowed outside the image area,
thus if an image larger than the current image is imported in
Commit or Preview with masking turned on, a lock up of the rubber
band might occur, therefore, it is generally not a good idea to
import an image larger than the current image, do the reverse
instead.
For a simple cut&paste, see Paint function.
Scale Scale is controlled by the scaling control panel. By default, the
aspect ratio is kept for scaling, thus changing one scale factor
will change the other. If the purpose of scaling is to correct
aspect ratio, click mouse on the KeepASP button to disable it,
and x and y scaling factors will become independent and can be
adjusted separately. The scales can be shown either as a ratio
(current size/original size) or as the resulting size in pixels
by toggling the ShowSize or ShowScale button. Subpixel sampling
can be turned on or off by clicking the mouse on Subpix button.
In certain situations, it might be desirable to make the image
having a particular size without changing the aspect ratio. To
achieve this, toggle on the Exact button, and the scale factors
become independent, adjust each until you've got the desired
size, bit will scale the image to the requested size, using the
smaller of the two factors, and filling the empty region with a
fill color, which can be set by clicking on the FillColor button.
To start scaling, click the mouse on OK button. To exit scale
mode, click the mouse on Done button.
Merge Load an image and concatenate with the currently displayed image.
If the separation between the two images are negative one image
is merged into the other.
To load the image to be merged, click on the Image2M button.
There are total of 12 alignments between two images (3 on each
side, e.g., if we want to place the new new image on the left of
the current image, within this constraint, there are center, top
and bottom. If the new image has the same size as the current
image, the three placements degenerate into a single location)
and each is shown as a small arrow. The meaning of each arrow
becomes clear if you consider the manual button as the current
image, and all other buttons are the possible new image
locations. The counter below the alignment button controls the
separation between the two images, thus a negative separation
merges two images. In performing merging, there are two ways to
handle the overlapping part, the default is for the new image to
overwrite the old one, i.e., new image on "top". To change the
default, click on the Top/Bottom button.
In addition to overwrite, other operations are possible. Current
implemented operations are defined as follows (assuming A is a
pixel in the current displayed image and B is the corresponding
pixel in the image to be merged):
center tab(;);
r l.
Name ; Definition
OverWrite; B
Mask; B ? B:A
Add; Min(B+A, 255)
Sub; Max(B-A, 0)
Diff; Abs(B-A)
XOR; B^A
Switch B&A if the top/bottom setting is bottom.
The placement buttons and the separation counter provide adequate
alignment between two images in most cases, nonetheless, a manual
button is provided in case you want to control precisely where
the image should be placed. To activate manual placement, click
on the manual button. Then you can use the arrow keys, hjkl[1-9]
to move the the rubber band, or you can use the mouse to move the
rubber band.
Upon a satisfactory location is found, click on the Merge button
to initiating merging/concatenation, and a new image is formed
with all empty regions filled with the fill color, changeable by
clicking on the FillCol button.
After merging/concatenating, although the image displayed and in
core have the merged size, further alignment requests are still
against the pre-merged image, not the merged one. To change this,
click on the Update button and subsequent alignment requests will
be against the current image. Depending on the situation and
relative image sizes involved, Update button may prove to be very
useful, e.g., to concatenate four identically sized images
together, the only way to do it without manual placement is to
concatenate the first two, and update the alignment reference
size, and concatenate the rest without update.
To exit merging mode, click on the Done button.
PixTran Arbitrary 1-to-1 pixel (color) transformation, i.e., all pixels
having value I will be replaced by F(I) in output pixel, where
F(I) is the transformation functions. The transformation can be
applied to the entire image or a rectangular region of it on RGB
separately or simultaneously.
Upon entering PixTran mode, shown is a control panel with three
XY graphs representing the RGB transformation function with the
horizontal axis being the input pixel value and the vertical axis
being the output pixel value. The initial transformation is an
identity transformation, i.e., F(I) = I. To change the
transformation function, click the mouse on the control points
(shown as little squares) on the graphs and the cursor will
change to a small cross. Keep mouse pressed and drag the control
points to change the transformation function. The number of
control points can be changed by the counter labeled
ControlPoints.
Initial size is the entire image, as indicated by rubber band.
To adjust the editing size, drag the small crosses on the rubber
band or use keyboard key hjkl (or arrow keys) and xXyY.
On top-right of the control panel, three rounded buttons labeled
Red, Green, and Blue are shown. Immediately below, there is a
large button labeled RGB. These four buttons control to which
component, or simultaneously all three components, the current
transformation applies to. These four settings are of course
mutually exclusive and setting one will unset the other three.
In principle, the interaction mentioned above with the
transformation function is capable of any transformations
imaginable, but in practice, it is not always easy to set the
transformation to some known functions, nor to change all the
control points by some given amount, such as gamma corrections
and brightness adjustment. To solve these problems, a math
expression parser and several common functions are provided as
shortcuts. The parser is rather lenient in the syntax and error,
but demands you name the variable x. Photo inversion, step
functions, gamma correction are among the shortcuts. In addition,
the intensity can be shifted either downward or upward. For large
number of control points with upward shifts, the final
transformation can be very interesting because of ignored
overflows. The overall brightness of the image can be adjusted in
two ways: the brightness can be increased by an absolute amount
(F(I) = I + Delta or by a percentage (F(I) = (1 + Delta * I. The
selection can be changed by clicking the mouse on text
Brightness.
Once a satisfactory transformation is found, click the mouse on
the button labeled Apply. The current image will be transformed
according to the transformation and the settings of which
component or components the transformation applies to. The
transformed image will then replace the original image as the
current image. To undo the transformation, press the mouse on
the button labeled Undo. Note, however, unlike in crop, only one
undo is supported, that is, if you did two transformations in
succession, only the last transformation can be undone.
To exit PixTran mode, press mouse on the Done button.
By using combination of the 1-to-1 transformations, some very
pleasing effects can be achieved. For example, to simulate the
effect of looking the image through a bluish glass, shift red and
green intensity down by two bits, and shift the blue intensity
down by one or zero bit, do the transformation. The result is
very nice. Similar tricks can be applied for reddish glass or
greenish glass. Be creative, when you find something
interesting, please let me know. I might add a shortcut button
for it.
Note when I said the interaction mentioned above is capable of
any transformation imaginable, I lied. It can't simply because
the transformation is three one-dimensional graphs instead of the
correct one three-dimensional one. For this reason, two
additional functions are supplied to swap two arbitrary pixels
values (e.g., change black to white and white to black) or simply
replace all pixels with a particular color with another color
(e.g., change all dark gray pixels to bright red).
Text A panel with many objects is shown in text mode. Text color,
size, and orientation can be set by clicking the mouse on
appropriate objects inside the text control panel, and any
changes in attribute take effect immediately.
Initial text is the current filename and can be changed by typing
into the control panel field labeled Text. The new text is shown
as soon as you type, both inside the field and in the main
window. To move the location of the text string rendered, click
the mouse inside the main window, the string will move to the
location of the mouse. You can also drag the text string by
keeping the mouse button down. For precise placement, use the
arrow keys or hjkl.
Sometimes, it is desirable to render the text onto a fixed
(uniform) background. To this end, turn on the Bkgrnd button, and
a uniform background will appear beneath the text. The background
color can be changed by clicking the mouse on the BkColor button.
Note that if the background is not enabled, neither is the
background color. Similarly, if you like to have a bounding frame
around the text, turn on the frame feature by clicking the mouse
the Frame button. In conjunction with PixTran's region feature, a
special background can be created that gives the image as appears
through semi-transparent glass.
Some special effects are available, currently, Shadow, engraved
and emboss are implemented. For any of these to look nice, you
need to play with the text and background color a bit. In
general, for shadow to look nice, a strong text and background
are needed and for engraved to look nice, a mild text and
background should be used and text should be stronger than the
background, and further, the background and text should be close.
The same is true for emboss effects.
Note there are two modes of operation, one is the immediate and
the other is the deferred method. In immediate mode, text are
rendered into the raster while in deferred mode, text is saved
separately from raster and may be output separately to get better
looking hardcopies. The mode of operation can be toggled by
pressing on the appropriate buttons. If text are saved in
deferred mode and later on, you decide to render the text into
the raster, you can do a crop.
If the purpose of adding text is to get an annotated hardcopy,
justification is extremely important. Keep in mind that screen
has a much lower resolution than a printer. The same string
displayed left justified and center justified (with right shift
of half of the string length) might look the same on the screen,
but could be very different on hardcopy. Therefore, always use
proper justification if hardcopy is desired. To preview what the
test might look like on hardcopies, choose the PrinterFont by
toggling the ScreenFont. This will switch the fonts used to
printer matched ones which give better indication how long the
text would be or how different lines of text would align on
paper.
If special symbols are desired, you can input them just like in
TeX(1), where special symbols are surrounded by dollar signs ($)
to indicate font switching. Unlike in TeX, superscript and
subscript do not work. Minor size changes on the same line can be
achieved by using command iny \small,
ormal, \large and
\huge. For symbols not defined by bit, you can still get them by
inputting their character code directly in octal preceded by
backslash (\), provided you know the encoding scheme of font you
are using. For text via PostScript, ISOLatin1Encoding is the
default (unless printer is Level 1 only and in that case standard
PS encoding is used).
Once everything, including the location, is satisfactory click
the mouse on the OK button to place the the current text into the
image (not necessarily into the raster, depending on
immediate/deferred settings). To exit text mode, click the mouse
on the Done button.
Note: in immediate mode, strings that are outside the image can
not be rendered. To render a string outside the image, use
deferred mode and do a crop following the text mode and fill the
empty region with whatever color you might choose.
Marking Mark locations on image with simple geometric figures, (arrows,
circles, rectangles etc). This function is controlled by a
control panel showing all available basic SGFs. Location, size,
color, orientation and line thickness can be individually
adjusted by clicking on appropriate buttons. By default, the
origin of the coordinate system is at the lower left corner of
the image and all distances are measured in pixels from origin.
Sometimes, the underline image may have intrinsic units
associated with it, and working in that units might be easier,
such as a map with one pixel-pixel distance equal to 0.83 km
(kilometers). To switch to this unit, toggle on the User
coordinate button and enter the conversion between pixels and
this unit, defined as the ratio between 1 unit in the pixel
coordinate system and the value it represents in the new unit
system, and in this example, you would enter 1.20482 (which is
1/0.83), meaning 1.20482 km is represented by 1 pixel-pixel
distance. Once in user coordinates, all distances and sizes
reported and requested should be in the new units, for example,
you can request to draw a circle of radius 120 km etc. The
origin of the coordinate can be changed by pressing the mouse on
the SetOrigin button, and the current location of the mark will
be the new zero of the coordinate system.
To move the location of the sgfs, you can either click inside the
main window where you want the sgf to be, or you can input the
coordinates of the locations you desire, and the SGF will move to
the new location. The input feature is extremely useful if user
coordinate system is in force.
For speed reasons, only the x- and y- components of the current
location is reported in the control panel by default. To get the
distance from the origin, turn on the Dreport option.
To place a mark, press the mouse on the OK button, current object
together with all of its attributes, such as color, location,
etc, are saved. To exit mark mode, press done button.
The marks so "rendered" may or may not be rendered into the image
depending on the settings of the Immediate/Deferred mode. In
Immediate mode, sgf is rendered into the raster and bit has no
record of it. In deferred mode, full vector information of the
sgf is retained and the sgf is drawn on top of the image, but not
rendered into the raster. The deferred mode is useful for
generating hardcopies via PostScript where the vector info is
used to scan convert the sgf on a printer which typically has a
better resolution (>300dpi) than the computer screen (<100dpi).
If sgfs are rendered in deferred mode, and later on there is a
need to render them into the raster, do a crop.
Note in immediate mode, all sgfs that are outside the image are
clipped. To render a mark outside the image, enlarge the image
first via crop.
For convenience, the following keyboard keys can be used: xXyY
have the usual meaning, that is xy decreases the width and height
respectively and XY increases the the width and height. Arrow
keys (or hjkl) can be used for precise placement of the sgfs. rR
rotates the sgf in discrete steps which can be set by [1-9].
If the hardware is capable of drawing anti-aliased lines, bit
will try to draw all lines anti-aliased. However, for colormap
images or lines with width that is not 1, anti-aliasing will
not be performed.
EditMap Only meaningful if the current image is in colormap. Upon
entering this mode, current colormap, 64 entries per page, and a
color mixer are shown. To select a specific entries in the
colormap, click on the little square inside the colormap. The
current selection will be marked by a small cross and current
selection's RGB intensity will be shown in the color mixer. You
can change the color of current entry by dragging the specific
color channels in the mixer and corresponding color in the map as
well as all pixels in the image that match the current entry will
reflect current settings. On the bottom of the colormap, there
are five small boxes, showing respectively, the index into the
map, the red, green, blue and luminosity of the entry. Entries
in the map can be scrolled by the large arrows on the sides of
the colormap.
To exit colormap editing, click the mouse on the Done button on
the colormap panel.
Enhance This is really histogram equalization generalized to color
images. Instead of equalizing the RGB histograms separately with
their own transformation function, a function based on the
luminosity will be applied to the RGB components. There is no
interaction with this function. A transformed image will be
displayed and the original image will be replaced by the
transformed one. In cases where the image of interest is
superimposed on a strong (either black or white) background,
histogram equalization gives poor results in the background
region, maybe a thresholding before equalization is called for.
Smooth Apply a special convolution where each pixel is replaced by a
weighted sum of itself and its nearest (3x3) and next nearest
(5x5) neighbors.
Upon entering this mode, a copy of the current image is made and
all changes that occur are made to this copy until return and at
that point, the changes becomes permanent. To force the changes
to be permanent before return, click on the Save button. Click on
Undo to undo all changes since last Save.
By default, the smooth is performed on the entire image. To
smooth a particular region of the image, drag the small crosses
on the rubber band. See Crop for more details on the interactive
sizing of the rubber band.
To preserve edges, a threshold may be used according to the
following definition:
If(Abs(SmoothedPixel- OldPixel) >= threshold)
NewPixel = SmoothedPixel
else
NewPixel = OldPixel
Different thresholds for R,G,B can be used and are set by
pressing on the counters.
To start smooth, click on the Smooth button. To finish up, click
on the Done button.
Sharpen Apply a special convolution where each pixel is replaced by a
weight sum of itself and its nearest neighbor. A rubber band
defines the region to which the sharpen operation applies.
MedianFilter
Perform a 3x3 median filtering. Again, A rubber band defines the
region of interests.
Histogram
Clicking on this button will bring up a window with four panels,
showing red, green, blue and white graphs. These graphs are the
separated histograms of the RGB components of the current image.
The horizontal axis is the intensity and vertical axis is the
population count. Note that this is not the true histogram
because correlation is ignored, nevertheless, some insight can be
gained by examining these graphs.
The graphs are static for efficiency reasons and therefore, even
when image has been changed, the histogram is not updated
automatically. To show the histogram of current (changed) image,
clicking the mouse on the Histogram button again.
Click on the title bar brings up the pop-up menu, whose entries
include re-scaling, change of plotting styles etc. Select Done
to finish.
Edge
Uses a variant of orthogonal Sobel gradient operators. If the input image
is in color, the output edge map will also be in color. The colormap used
depends somewhat on the image itself. By looking at the histogram of the
edge map, a good threshold value can be chosen and then a two-level step
transformation will produce a bi-level image. No interaction.
Image Difference
Computes the differences between two identically sized images pixel by
pixel and displays the results as an image. More information can be
gained by showing the histogram of the difference image.
Image Panning
Image panning makes viewing images larger than the screen possible. The
panning control panel consists of nine buttons labeled by the directions
of the movement. The center dots will center the image both vertically
and horizontally. The slider immediate below controls the pan step, in
units of pixels. In SlideShow mode, image panning is automatically
performed in both directions unless disabled by the initialization file.
Measuring Pixels Intensities Along a Line
The length and angle of the defining line can be changed by dragging the
crosses at the two ends of the line. The origin of the line is marked by
a circle around the cross. The intensity is presented as four graphs with
horizontal axis being the distance, in pixels, along the line from the
origin. The vertical axes are the Red, Blue, Green intensities and
luminosity (or color index depending on the image type). To activate or
update the reporting, click mouse inside the main window. Menubutton
again controls the pop-up menu through which the line color can be
changed or online help can be obtained.
Edit Individual Pixels
Upon initiating, a rectangle attached to cursor is shown and the region
within the rectangle is magnified and shown in the PixEdit window. Once
you have found the region you intend to edit, click leftmouse or
rightmouse within the main window, the rectangle will cease to move with
the cursor and the area within the rectangle will be the region you can
edit.
To select the pixel to edit, click leftmouse on the pixel in the PixEdit
window. Its value can be changed in the same way as in Select a color,
i.e., by dragging the slider and/or changing the colormap entry.
To set a pixel to the current color (as shown in the mixer/colormap),
click leftmouse on the pixel while holding down the ALT key or. by
clicking middlemouse. Dragging will set all pixels along the path.
To change all pixels that have a particular value to a different one,
edit one first, then use the pop-up entry Repeat. For example, if one
wants to change all bluish pixels to red, select one bluish pixel and
change it to red, then do a Repeat thru popup (To changing/swap pixels
for the entire image or a large region of it, See PixTran function).
The pop-up menu has the following entries Help, MarkCol, GridCol, Repeat,
Cancel and Done. MarkColand GridCol cycles the the colors of the pixel
mark and the grid; Repeat Do the last editing on all pixels in the
PixEdit window. Cancel discard all changes and return. and Done saves
all changes and return.
The following is not editing functions per se, but its interaction is
complex enough to warrant an explanation:
Image Browser
Click on the IBrowse button to activate the image browser. If the
thumbnail images exist (created by an earlier session of bit), they will
be loaded and shown in the browser. If thumbnail images do not exist, bit
will only show generic icons, i.e., the program will not generate
thumbnail images upon startup.
To view an image, double click on the filename icon. If the image is
successfully loaded, its thumbnail image will be made and used to replace
the original generic icon. Double click on a directory will spawn another
image browser. Total of 7 directories can be simultaneously displayed.
To generate thumbnail images for all visible filenames in the browser,
click on the Update button. If a thumbnail has a date that is earlier
than that of the image, the thumbnail is re-generated.
Directory entries are cached and to force a re-read after addition or
deletion, click on the ReScan button. Note however, ReScan will also do
an update. In addition, thumbnails whose original images no long exist
will be deleted.
To move forward and backward within the browser, the arrow buttons and
the slider can be used. For efficiency reasons, slider change is acted
upon only on its release, that is, program does nothing while you are
changing the slider. Arrows move one-screenful of files at a time, and
thumbnails, if they exist, are loaded as you go. This could be slow if
many thumbnails exist, to move back and forth quickly without loading
any thumbnail images, keep CNTL key down while pressing on the arrow
buttons.
To cleanup all thumbnail images, click on DelAll button. To cleanup
manually, remove ./.btn directory and all entries in it.
There is no built-in buttons to manipulate (for example, deleting the
original image) the original images. A mechanism via run time definition
is provided to achieve this operation. Once a definition is registered,
it operates on all marked files. To define an external binding, click
menubutton on the button next to the Close button; to execute the
binding, click leftmouse on the button.
The definition is consisted of two strings, and two booleans. The two
strings are the button label and a shell command respectively and two
booleans are Confirmation and MultiArgs. MultiArgs indicates if the shell
command accepts multiple arguments and if Confirmation is set, bit will
pause for a mouse click before the actual execution of the binding. The
shell command takes the form
cmd [options] [?] [options]
where ? will be replaced by the filenames marked. For example,
mv ? /usr/people/good_dir
Paint
This function offers standard paint functions, namely, pencil, spray,
brush and some other common operations.
Upon entering paint function, a panel containing all paint function
icons are shown. Click on the icon to activate a specific function.
On the lower-right corner of the panel, current line width and fill
pattern are shown. Two special entries are available to facilitate the
uniform operation of outlining and fill. By default, outlines are drawn,
possibly with a different color, around the filled area. To turn
outlining off, select None (or zero) for line width. Similarly, all
closed area are filled unless the fill pattern is None.
MPEG movies
This is one of the file formats that can hold more than one images in it
(PostScript is another). Upon loading and displaying the first frame, a
control panel is shown. Click on the Play button to show the movie non-
stop, interruptible by the Stop or Done button. To play and rewind/loop,
click on the Play&Loop button, again, interruptible by Stop/Done button.
Number of frames per second is reported at the end of a sequence.
To display frame by frame (stepping), click on the NextFrame button.
The default magnification factor for MPEG movie is 2x2. To change it,
click on the left-most box above the FrameControl box. Use menubutton for
random selection.
To introduce a delay between two frames, click on the box next to the
magnification box, click leftmouse for next entry, middle mouse for
previous entry and rightmouse for random access.
To convert the currently displayed frame to other format, click on the
Convert button. The destination format is selected by clicking the box on
top of the Convert button. To convert all frames, starting from current
frame, click on ConvertAll button.
All other buttons have the obvious meanings.
Also see sgimpeg(1L) for a simpler, stand-alone MPEG player.
Note: seeking of any kind does not work. Clicking on PrevFrame is the
same as NextFrame.
PostScript Files
PostScript is supported via GhostScript, see gs(1). Again, upon
converting and displaying first page, a control panel is shown if the
document is multi-paged.
The following functions are not directly supported by bit program but can
be done indirectly:
Measuring Pixel Intensities
In any mode where a color is prompted for, e.g., rotate, scale crop, text
etc, click inside the main window while holding down CNTRL key will read
the intensities of the pixel under the mouse. Again, the RGB, and
luminosity (or CI) will be reported. The mouse position can be gotten by
turning on the mouse reporting option through Setup menu. Also see Select
a Color.
ACKNOWLEDGMENT
JPEG support is based on the Independent JPEG Group's library Version 4.
The Graphics Interchange Format (c) is the Copyright property of
CompuServ Incorporated. GIF(sm) is a Service Mark property of CompuServ
Incorporated.
The FORMS library is developed by Mark Overmars.
MPEG support is derived from the Berkeley source.
FILES
/usr/local/lib/bit/*.hlp
All online helps. In case environment variable BITHELP exists, BITHELP
will be used instead of the default.
/*.dat
Some control files
/Setup.opt
Basic configuration. The first file bit reads unless -n is present.
~/.bitrc/Setup.opt
User configuration file which is read after reading the system default.
~/.bitrc/BIT_filters
The definition file for external filters. It will be concatenated with
definitions read from /BIT_filter.
~/.bitrc/BIT_convolve
The external convolution matrices.
SEE ALSO
djpeg(1), cjpeg(1), ppm(5), pnm(5)
BUGS
Command line options are single letters.
Unable to load two images simultaneously due to loading module's re-
entrant problems.
Can't run bit remotely, at least not reliably.
AUTHOR
Copyright (c) 1993, 1994 by T.C. Zhao (zhao@monte.svec.uh.edu)