/Net/www1/CERN_WWW/CERN/Divisions/ECP/PT/PINS/DALI.html
DALI
Hans Drevermann 1993.03.26
Goal
The graphics program DALI has been
developed to visualize events from
the ALEPH detector at LEP for presentation
purposes and for visual analysis.
Most event display programs are based
on conventional 2D and 3D methods
and are designed for general applications.
In contrast, DALI offers SPECIAL
PROJECTIONS that are adapted to the
cylindrical geometry of the detector
and to the track geometry of charged
particles originating from a common
vertex and moving in a homogeneous
magnetic field. For the selection
of convenient pictures, colors etc.
HUMAN PERCEPTION issues are considered.
Special 2D Projections
Two predefined STANDARD PROJECTIONS,
a front view and a special side view,
allow to visualize all tracks and
showers from ORTHOGONAL sides. Through
a ªFISH EYEº transformation the front
view can be modified in a way to
visualize simultaneously inner and
outer detectors, i.e. by enlarging
the inner and compressing the outer
detectors. In these projections the
different subdetectors do not overlap,
so that hits can be uniquely associated
with the subdetector, by which they
were recorded.
Through LINEARIZATION of track images
by use of NON-LINEAR ANGULAR PROJECTIONS,
track identification, analysis and
extrapolation are improved. This
is further ameliorated by TRACK COMPRESSION
through adequate subsequent linear
transformations. These methods strongly
improve the capabilities of human
pattern recognition. Selected regions
containing the full track may be
magnified to such an extent that
the errors of the tracking devices
become apparent. This means that
the LIMITATIONS DUE TO SCREEN AND
EYE RESOLUTION are overcome.
Special non-linear transformations,
which transform straight lines into
straight lines, allow to ESTIMATE
LOCALLY the relevant track features,
i.e from the measured points alone
without drawing the interaction point.
Special 3D Projections
These concepts, which can be applied
for 2D and 3D data, become more powerful
when taking advantage of the full
spatial information of well measured
3D data. This is accomplished by
a picture called V-PLOT, which visualizes
in one picture all tracks in 3D,
i.e. track curvature and spatial
direction, as well as the correlation
between tracks such as track crossings
in space. All these features can
be estimated locally from the displayed
tracks or track hits only. One can
also estimate if tracks originate
from the main vertex or not. The
V-PLOT concept does not rely on track
identification by a pattern recognition
program.
Due to a rigorous track compression
the V-PLOT improves track identification
by human pattern recognition. Therefore,
the V-PLOT is especially powerful
as compared to other projections
for the VISUALIZATION of SUPER HIGH
MULTIPLICITY EVENTS, i.e. events
with hundreds of tracks. The concept
of the V-PLOT can be generalized
and applied to other experimental
setups than ALEPH.
Representation of Calorimetric Data
Calorimetric data are represented
in the standard projections via RADIAL
HISTOGRAMS. The data can be analyzed
more precisely through lego plot
like projections of the different
calorimeter layers. The association
of showers in the different layers
is much facilitated by the use of
ISLAND REPRESENTATIONS. In the case
of a large number of layers and large
showers, the PUZZLE PLOT allows the
representation of parts of a calorimeter
in all three dimensions, so that
the connection of cells to a cluster
can be examined. It also allows to
represent the energy deposit of the
individual cells, thus visualizing
a 3D scalar field.
Colors
In DALI much care is given to the
USE OF COLORS both for events and
detector units. The standard projections
allow to draw all detector units
as solid areas, thus providing an
easy and fast understanding of the
pictures. This feature is especially
appreciated when using DALI pictures
for PUBLIC RELATIONS.
In most cases, colors are used to
CORRELATE hits, tracks or detector
units between different pictures,
which may also contain residual plots
or lists of tracks, etc.. DALI allows
the display of different pictures
SIDE BY SIDE.
A sophisticated way of FRAMING hits,
tracks and histograms allows LIGHT
BACKGROUND COLORS, if desired. The
framing is done in such a way that
resolution is not deteriorated. By
drawing histograms in several passes
on top of each other, as solid areas
and as wire frames, overlapping histograms
are still fully visible. Furthermore,
special ways of framing histograms
emphasize their structure.
Special Interactive Methods
The visualization concepts described
above are supported by a series of
SPECIAL INTERACTIVE METHODS. The
radial event and the cylindrical
detector structure are taken into
account for the implementation of
GENERALIZED RUBBER BAND CURSORS,
shaped as rectangles or parallelograms
both of which can be rotated. The
areas defined thereby are subsequently
linearly transformed to a predefined
window. In this way all possible
linear transformations can be defined.
RUBBER BAND HELICES may be used for
visual track fitting simultaneously
in different projections.
The V-PLOT is a powerful interactive
tool for INTERACTIVE TRACK SEPARATION
and for the definition of spatial
regions to be displayed in other
projections. It allows objects or
regions to be indentified across
different views through the application
of SYNCHRO CURSORS, i.e., several
interactively defined, synchronously
moving cursors.
In order to retrieve the relevant
information of picked objects, the
well known concept of picking is
implemented for hits, tracks and
histograms. The INVERSE ACTION TO
PICKING is also available, namely
the positioning of the cursor to
objects selected beforehand by picking
on another picture or by other means.
In the case of extended objects such
as tracks or in the case that the
same object is drawn several times
on the same picture, the cursor is
moved to a point on the object closest
to the current cursor position or
to the closest occurrence of the
same object. This is useful for escaping
from crowded regions.
The color mixing may be optimized
by editing. TEXT, comprising letters
and special symbols, can be inserted
at any position of the picture in
different sizes and colors. Polylines
and areas can be drawn.
Combination of Menus and Manuals
Commands may be given by clicking
on a HELP MENU. This combines extensive
explications and highlighted commands,
which can be invoked by clicking.
Thus, there is little need for the
parallel use of a written manual,
which simplifies the task of the
operator to find the relevant explanations
and the task of maintenance, as only
one help menu must be maintained.
Instead of clicking, one may type
a command in order to save time.
The COMMAND language is tuned to
be short. When clicking on the menu
the command in question is echoed
on the terminal window, which helps
the learning of commands and shows
the last actions. DALI also preserves
a LOG FILE of the actions. In this
file the terminal output and the
operator commands are stored, independent
of their input source. The commande
language is especially helpful for
passing information on a hot line
and to define MACROS. A TRAINING
SCHEME presents to the user the tools
in different user selected levels
of complexity. The inexperience user
is thus not intimidated whith too
much information.
PostScript Format
Pictures may be stored in PostScript
format, either as bit maps or in
basic PostScript language. The PostScript
file contains a general setup section
which by simple editing, allows pictures
to be reproduced with different colors,
different intensities or in black
and white.
Slow Motion Picture
For public relation purposes a sort
of animation is used to show the
creation of events, i.e. its recording
in the detector, as a SLOW MOTION
PICTURE.
As DALI does not require large resources
(it does not offer smooth 3D rotations),
it runs on simple workstations. A
picture is typically drawn in at
most a few seconds. The program was
originally developed for VAX stations
with UIS. The present version is
based on the X WINDOW SYSTEM and
runs on VMS and ULTRIX workstations.
It contains roughly 90 000 lines
of FORTRAN code and 7 000 lines of
C code.
References
H.Drevermann and W.Krischer, Nucl.Instr.Meth.
A239 (1985) 160. H.Drevermann and
C.Grab, Int. Jour. Mod. Phys. C1
(1990) 147. H.Drevermann, C.Grab and
B.S.Nilsson, How to Represent Three
Dimensional Data of Events from High
Energy Physics? Proceedings of the
international conference on Computing
in High Energy Physics 91,Tsukuba,
Japan, page 545. Universal Academy
Press, INC.-Tokyo H.Drevermann,
C.Grab, D.Kuhn, B.S.Nilsson and R.K.Vogl:
A New Concept for Visual Analysis
of Three Dimensional Tracks John
Wiley Sons: New Trends in Animation
and Visualization, edited by N.Magnenat
Thalmann and D.Thalmann
H.Drevermann, D.Kuhn and B.S.Nilsson,
Is There a Future for Event Display?,
to be published in Proceedings of
the 1992 CERN School of Computing,
Aquila, Italy