#Stimulus duration framen psychopy drivers
#Stimulus duration framen psychopy update
#Stimulus duration framen psychopy software
#Stimulus duration framen psychopy trial

The biggest of these is the loading of images from disk.įor image stimuli where the image is constant the image should be loaded from disk at the beginning of the script (Builder-generate experiments will do so automatically for you). Even if you really know what you’re doing, it’s easy to make a silly mistake, and if you don’t really know what you’re doing then all bets are definitely off!!Ĭommon ways for this to happen are to forget the operations that are potentially time-consuming. It can be really easy, as a user, to introduce timing errors into your experiment with incorrect coding. Such a system means that your images all appear exactly one frame later than expected.Įrrors caused by triple buffering, either by the operating system or by the monitor, cannot be detected by PsychoPy ® and will not show up in your log files. Triple-buffering is a system whereby the images being rendered to the screen are put in a 3rd buffer, and the operating system can do further processing as the rendered image moves from this 3rd buffer to the back buffer. Triple buffering: In general PsychoPy ®, and similar graphics systems, are expecting a double-buffered rendering pipeline, whereby we are drawing to one copy of the screen (the “back buffer”) and when we have finished drawing our stimuli we “flip” the screen, at which point it will wait for the next screen refresh period and become visible as the “front buffer”.

#Stimulus duration framen psychopy update

The particularly annoying factor here is that your experiment might work with very good timing for a while and then the operating system performs and automatic update and the timing gets worse! Again, the only way you would typically know about these sorts of changes is by testing with hardware. That said, PsychoPy ® does aim to give you as high a temporal precision as possible and, in a well-configured system achieves this very well.Īll three major operating systems are capable of introducing timing errors into your visual presentations, although these are usually observed as (relatively) constant lags.

#Stimulus duration framen psychopy trial

a Labhackers Millikey) for response timing with a sub-ms precision, but note that there will still be an apparent lag that is dependent on the monitor’s absolute lag and the position of the stimulus on it.Īlso note that the variance in terms of response times between your participants, and from trial to trial within a participant, probably dwarfs that of your keyboard and monitor issues! The clocks that PsychoPy ® uses do have sub-millisecond precision but your keyboard has a latency of 4-25ms depending on your platform and keyboard.

#Stimulus duration framen psychopy drivers

the monitor taking additional time to process the image before presentationĭelays caused by drivers and OS: Windows, Linux and Mac all perform further processing on the images, depending on settings and this can delay your visual stimulus delivery by a further frame interval or more The following issues are extremely common and until you actually test your experiment you don’t realise that your timing is being affected by them:Īdditional delays caused by monitors: e.g. The sources of error in stimulus/response timing are almost never to do with the poor precision of the clock.

#Stimulus duration framen psychopy software

Many software and hardware manufacturers will suggest that the key to good timing is using computer clocks with high precision (lots of decimal places) but that is not the answer at all. Very often a computer is not optimally configured to produce good timing, and a poorly coded experiment could also destroy your timing (which is one reason we now recommend even good coders use Builder!). The longer answer is that you should test the timing of your own experimental stimuli on your own hardware. For detailed Studies of timing see Bridges et al., 2020

PsychoPy’s timing is as good as (or in most cases better than) any package out there. Can PsychoPy ® deliver millisecond precision? ¶