UniServe€Science News: Newsletter of the Science Software Clearinghouse Vol. 3, March 1996

InSights 2 (Color)

Peter Wenderoth

"In sum, this is a terrific package which is easily worth the US$169 for a site licence."

The package contains the following 10 self-contained modules on visual perception: Stereograms (which uses red-green glasses - one pair supplied); Additive Color Mixture; Illusions and Aftereffects; Mach Bands; Stereopsis and Depth; Contrast Sensitivity; Spatial Vision; Color Arrangement Test; Form and Motion; and Feature Analysis (visual search). Real data can be collected in the experimental modules - Stereopsis and Depth, Contrasts Sensitivity, Color Arrangement Test, and Feature Analysis.

The package is a lot of fun. In using it with students in 1995, I found that even those who claimed to be terrified of things "scientific" were captivated by the excellent visual demonstrations. The Stereograms module contains ordinary picture, outline shapes, and random dot stereograms and, except for the pictures, disparity is choosable. This and the 18 demonstrations in the Illusions and Aftereffects module alone make the package worth its price. The Form and motion module makes accessible a range of demonstrations which cannot easily be explained merely with words.

One of the nicest things about the program is that it can be used at different levels. For example, the Contrast Sensitivity module allows one simply to demonstrate the contrast sensitivity function (the opening screen displays a contrast x spatial frequency envelope), to measure it using a crude adjustment method (using a window slider) or to measure it using a sophisticated two interval forced choice procedure. One problem with the latter - not foreseen by the authors - is that the two intervals are defined by beeps, as is the feedback on correct/incorrect responses. As a result, when one tries to run this experiment in a lab class, the result sounds like a bad orchestra warming up and the students cannot tell which beeps define their interval or someone else's. It would have been clever to allow the user to choose the method of interval definition - visual perhaps (despite possible artefactual interference) or even to choose the beep frequency. The other problem with this experiment as well as some other modules - and as noted by Blake and Rose (1994) - is that there is no simple way of exiting once the experiment has started. It is therefore necessary to teach students that Control-Option-Command-Escape can be used to quit any Macintosh application.

Perhaps the best module of all - especially powerful for teaching - is Spatial Vision. Normally, students are not highly enthusiastic about Fourier analysis and synthesis. Here they can import all sorts of supplied PICT images and filter them with a two-dimensional fast Fourier transform using low pass, high pass, notch and bandpass filters, all of variable width. It is a particularly useful way to demonstrate the fundamental and harmonics of a square wave and to show that different spatial scales provide different kinds of information (low pass: overall shape; high pass: edges). One menu also allows blocking of faces and students are interested to discover that blurring the eyes restores recognition and to hear that this is why the blocked faces on TV news programs now jiggle to disable this recognition ploy. It is worth reiterating the disappointment voiced by Blake and Rose (1994) that there is no inclusion of orientation-domain variability, although it is possible to make one's own set of grating PICTs with different orientations and frequencies for import into the program.

There is very little positive which can be added to the excellent and detailed review of Insight 2 (Color) by Randy Blake and David Rose (Blake and Rose 1994) . They also mention some of the program's difficulties which are, however, relatively minor compared to its strengths. Because their review covers all of the modules fully, only some were discussed here.

The Blake and Rose review points out that some modules will not run with certain control panels. While this is true, I have also found that the Feature Analysis module will not run on a Centris 660AV even when all the extensions are turned off. Everything runs perfectly, however, on the very cheap and value-for-money Macintosh LC630 which is the laboratory machine I use. A maths co-processor is required only for the Spatial Vision module which will still run on a Mac 8500 using a software FPU, albeit very slowly.

If the on-line help and information is a little difficult to navigate, it is easy to print out a complete manual, complete with pictures of the screens (using the print screen command -- Shift-Command-3). In sum, this is a terrific package which is easily worth the US$169 for a site licence.

Peter Wenderoth
peterw@vision.bhs.mq.edu.au

Peter Wenderoth is a Professor in the Psychology Department at Macquarie University

REFERENCE:
Blake, R. and Rose, D. (1994) Perception, 23: 1102-1104.

Requirements: Macintosh computer with at least 3.5 MB of space on its hard disc, 2 MB RAM, System 6.05 or higher (4 MB needed if using System 7), HyperCard 2.0, 13" colour monitor, several features of the program require a maths co-processor.
Authors: John Baro & Stephen Lehmkuhle, University of Missouri, St. Louis
Distributor: Intellimation Inc., Dept 5CKF, 130 Cremona Drive, Santa Barbara CA 93116, USA
email: intellLFM@aol.com Tel: 1 805 968 2291
12 pp + software.
Cost: Single copy US$39, site licence US$169