
The Black Box Toolkit Page 75 of 98
Copyright © 2004 The Black Box Toolkit Ltd. All rights reserved.
Once everything is ready on both the host and remote PC DSCAR is set running by
pressing “F5”. As each bitmap is displayed by the paradigm a response is automatically
triggered at 300ms with the actual sequence of bitmaps only lasting a few seconds.
When data for a given device has been collected the next response device was tested.
The identical E-Prime script was used each time, bar that the response device was
changed. For the voice key DSCAR was programmed to generate a response on line 5 to
trigger the BBTK Tone Generator. A duration of 150ms was used meaning that the trigger
tone was output for 150ms.
Once a device has been tested the following data is available for analysis:
BBTK .rtl file (host PC) E-Prime E-DataAid .edat file (remote PC)
• Data on the onset of each image
• Offset of each image
• Duration of each image
• Refresh rate
• Onset of the simulated response
• Offset of the simulated response
• Duration of the simulated
response
• Data on the refresh rate
• Duration of the image display prior
to termination
• Reaction time
• Any other logging data that has
been chosen to be logged
Primarily we are interested in two key measures of timing error:
1. The display error – the image display was scripted in the paradigm to terminate to a
black screen immediately after a response. Ideally this should have been at the end
of the same refresh. The target display time should be as close to 300ms as
possible.
2. The response time error – this is calculated by looking at the response times
recorded by E-Prime. Any difference from 300ms is error.
Here we need to make use of data collected by the BBTK together with the response times
recorded by the paradigm. To analyse data collected by DSCAR we use the data analyser
module.
The example below shows the .rtl file for an OEM unbranded mouse.
Kommentare zu diesen Handbüchern