Black Box IEEE 1284 Guía de usuario descargar pdf (Pagina 6)

Page 6 of 98 The Black Box Toolkit

1.2. How can I ensure my timing accuracy is up to scratch?

The ethos of the BBTK is to allow researchers to benchmark their paradigm in-situ and

without modification by means of an easily programmable “virtual human”. By making use

of a wide range of external sensors the toolkit can detect a variety of stimulus materials

when presented. Depending on programming it can generate a response at a known onset

and for a given duration. The toolkit can detect visual stimuli, auditory stimuli or any TTL

signal. Responses can be made using TTL signals, switch closures (button/key down) or

through a tone generator to trigger voice keys etc. Stimulus detection and resulting

responses are recorded with sub-millisecond accuracy.

Conceptually the BBTK offers much the same functionality as a four channel digital signal

generator and a four channel digital oscilloscope. It is easily capable of sub-millisecond

sampling rates on a typical host PC (around 48kHz, or 48 samples per millisecond). Unlike

a signal generator and oscilloscope which typically costs many thousands and are difficult

to use, the BBTK enables the researcher to check most paradigms in-situ in less than 30

minutes. Even with a modern signal generator and scope you cannot hope to virtualise

human senses and response characteristics – with the BBTK you can do just that! Timing

analysis of events is accomplished using a virtual 8-channel oscilloscope style display.

Moveable cursors allow event timing to be measured relative to any two points. Four lines

show detected stimuli and four lines show simulated responses made by the toolkit and

fed into the remote PC running the paradigm being benchmarked.

1.3. How can you ensure millisecond accuracy and experiment generators typically

can’t?

First the BBTK will only run on Microsoft Windows 2000 and XP based operating systems.

Multi-tasking operating systems typically use threads with each piece of software utilising

one or more of these threads. The operating system typically gives each application a

share of the CPU time in a round-robin fashion based on each applications threads and

the priority of those threads. Typically experiment generator software tries to increase the

priority of its own threads so that it has more time to process its own tasks and is less

likely to be interrupted by other applications that might disturb its progress. However under

Microsoft Windows for example the priority can only ever be boosted to “Above Normal” or

“High Priority” as boosting threads any higher will have undesirable side effects. All BBTK

software runs with its threads switched to “Realtime” priority. This means that our software

gets nearly all the processor time and is not interrupted by other software or the operating

system itself. Experiment generator software cannot run with “Realtime” priority due to

“undesirable” side effects. Realtime priority “locks out” non-critical operating system

threads meaning that the mouse, keyboard and background disk flushes are turned off.

Obviously there is little point in having an experiment generator that cannot accept

responses from participants!

Once a piece of toolkit software is running it cannot be stopped as it takes over the whole

system until it reaches a predetermined time limit and control is relinquished. The mouse

and keyboard become inactive and the only way to stop a module is to turn off the PC!

Whilst event data is being collected timestamps taken from the PC’s internal high

resolution timer are stored within a pre-allocated memory buffer. Once a run has finished

data is written to the hard disk ready for analysis. This helps ensure we are millisecond

accurate.

1 2 3 4 5 6 7 8 9 10 11 ... 97 98

Comentarios a estos manuales

Sin comentarios

Black Box IEEE 1284 Guía de usuario Pagina 6

Comentarios a estos manuales

Relacionado con productos y manuales para Redes Black Box IEEE 1284