[visionegg] Re: time and OS es

Mark Halko a écrit :

The other solution to this is to only use the first TTL pulse, as a "trigger" for your experiment. You can then use the computer's own internal timing to display your stimulus. If your experiment doesn't start on time, but starts 2 seconds late, you'll know you missed the pulse.

Yes. I use that approach now. That is how I discovered that one of our scanners has an actual interscan delay which is different than the requested one, leading to a drift which ruined an audio experiment where stimuli were supposed to be presented in the one second gap between consecutive scans... Now, we measure the interscan delay rather than trust the vendor's program.

Which brings out one of the points I made but didn't emphasize. Many experimenters do not record the error of the measurement. A long tradition of canned packages that just give you back what you put in has really promoted this. I'm glad some of the newer ones like Experiment Builder (SRI) and Presenter(?) record what happened.

Call me paranoid, but can you trust that the software knows everything? The hardware could introduce some additional buffering, no?
This can matter or not, depending on the experiment. Audio sound boards, for example, have actual sampling rates that can differ markedly from the requested one. If not taken into account, this can introduce systematic differences in reaction times measured from targets appearing at different moments in the audio stream.

In the end, for some experiments, it is necessary to check with external means if the timing of stimuli matches what is intended.

A very useful contribution would be a toolbox to perform this sorts of checks. It could be ran on second computer connected to the stimulation PC (to its audio ouput, its parallel port, and to its screen via a photodiode). Not everyone knows how to use an oscilloscope or wants to use one.

One possibility may be to build some simple hardware to record the various events in an experiment and mix them as an audio stream on the second computer. This "poor man's" approach once allowed me to check the delay between the display of an image and the start of a sound file. It actually worked quite well (the stimulation program was a commercial software which has not been cited here yet, and which I do not like at all, but I must recognize it handled timing quite well in this case)

The idea of a stripped down distribution is a good one (I usually only disable power management).
Once latency problems are solved, a project that would get me excited would aim at creating a Linux-Live CD for Psychophysics (a la Knoppix). 

Cheers (and apologies for this long series of off-topic posts)

Christophe Pallier










======================================
The Vision Egg mailing list
Archives: http://www.freelists.org/archives/visionegg
Website: http://www.visionegg.org/mailinglist.html

Other related posts: