Question: Minimizing between Trial loading Time (also script for a slider included)

AlexanderGiesche

Hey everyone,

recently i programmed an experiment in which you see a fixationcross (4s), a task (15s, text) and and a rating-scale (5s). It works ok (although there is a 10ms delay between the single objects). The bigger problem for me is that there is a about 350ms delay between trials.

I read the article about timing on the homepage but the hints seemed to not help my needs.

Anyone has an idea how to minimize between trial loading time? Also is it possible to run the objects of the sequence even more precise than 10ms delay? (i prepared all canvas objects in advance)

I attached my experiment here. Anyone feel free to use the slider for his/her own project (programmed it with the help of another forum member-eduard :)

paradigm.osexp

eduard

Hi Alexander,

The bigger problem for me is that there is a about 350ms delay between trials.

No, I don't think there is anything to be done about this delay. There are 2 workarounds though

1. take this delay into account and adjust the duration of the last parts of a trial to last ~350ms less that it should be, so that the effective duration is then what you would expect (that requires that the cross-trial delay is consistent)
2. drop the prepare-run layout and code everything in a single inline_script and a single loop. That will give you more control over the timing, but also more responsibility to make sure the timing works.

From looking at your experiment, I see some inefficiencies, that you could get rid of. Not sure whether it will reduce all of the delay though:

Create objects only once - right now you create keyboards and canvasses on every trial, which is a rather costly process. Keyboards you can initialize only once in the beginning of the experiment. They will be available throughout the course of the entire experiment. Canvases, you can clear if needed (canvas.clear()), but also here it might make sense to only create once and later only adjust settings of it, e.g. position, color, stimulus, etc.

The form might also be something you can create a priori, or at least in the prepare phase. Not sure how much time you can save with it.

Hope this helps,

Eduard