timing differences between backends

v.b

Hi everyone,

I`m quite new to using open sesame and while trying to get used to the program I experienced some troubles concerning the timing of stimulus presentation.

As temporal precision is important for my experiment I decided to use either expyriment or psychoby as backend.
I´ve programmed a sequence (using an inline script, see attachment) with a simple structure (a cloud is presented, participants have to press space bar, as a result a flash is presented, ) to see which backend works best.

As variable I calculated the variable delay, which is the delay between the time of the keypress and the flash. I had expected the delay to be somewhere close zero (not zero of course due to the refresh rate of the monitor but somewhere close ). This is indeed the case if I use psychoby (the delay varies around 10-16ms). However, if I use expyriment the delay is somewhere around 76 ms.
Timing is slightly better if I run the experiment on a Mac computer, but still the difference between those backends is notable.

Does anyone have an idea what might cause this timing differences between the backends?
Thanks a lot,

sebastiaan

Hi VB,

The Expyriment backend can indeed be slow when presenting complex stimuli (such as images), in the sense that the Canvas.show() function takes some time to execute. (The timestamps will still be accurate though, so at least you know about the delay.) The PsychoPy backend suffers much less from this, for technical reasons that are also not entirely clear to me. If Canvas.show() takes unusually long, you'll also see a warning in the debug window with recent versions of OpenSesame.

This is indeed the case if I use psychoby (the delay varies around 10-16ms).

That seems healthy, given that you have a 60 Hz display.

Cheers!
Sebastiaan

v.b

I see. :-) Thanks a lot!