Some thoughts on event visualization -- JU - 9/19/02


First, some features of our events:

  • They are long and skinny,
    --> transverse and longitudinal scales in *** vs Z views should be independent
    --> problematic for symbol size - would like size to increase as zoom increases, but don't want distortion

  • We have dead space in the detector: ,
    U, V spacing 4.1 cm, but elements really are 4.1 cm wide
    Z spacing 6.x cm, but elements are only 1 cm wide.
    --> "realistic" detector representation "wastes" valuable real estate on display
    --> possible solution: allow symbols to overlap in U or V?
    --> for me, plane vs strip view is more useful than ** vs Z.

  • Cosmic ray events come in at all angles
    --> requirements on aspect ratio differ from those for nominal beam-neutrino events
    --> in particular, detector aspect ratio is not optimal for cosmics.
    --> Also not optimal for studying demux issues.

  • Some examples
    --> A MINOS beam event - old midad
    --> A MINOS atm event - new midad

  • Basic problem with long events:
    --> want to be able to examine detail, without losing complete picture of event.
  • Solutions?
    --> non-linear scaling: fisheye/wide-angle views?
    --> magnifying glass: position curser over area of interest, blow up small portion of display, blown up view contained within boundary of magnifying glass...

    Example from CLEO -- zooming on silicon detector

    A CLEO event - nominal display
    same event - zoomed onto silicon
    same event - full view, but now fisheye