[solved] How did you evaluate pygame's timing?
On this page (http://osdoc.cogsci.nl/getting-started/choosing-and-configuring-a-system) you note that pygame's "temporal properties are not as good as of those of the other back-ends. In practice, this means that the timestamp of display presentation has an error in the order of 5 to 10 ms."
I'd be interested to hear how you came to this evaluation.
Comments
Hi Mike,
Please see the benchmark experiment in our paper on OpenSesame: http://www.springerlink.com/content/n264513n66704v33/ (open access)
Basically, PyGame offers poor temporal temporal precision when it comes display presentation in non-OpenGL mode. In OpenGL mode things are different. In that case it's fine, as you can read in the paper as well.
The reason that legacy (non-OpenGL PyGame) is the default back-end is that it's much more stable than most other OpenGL-based options. But right now I'm testing OpenSesame with Expyriment, which is an OpenGL based library (also using PyGame btw). If that works out well, OpenSesame will probably switch to Expyriment as a default, and the timing issues will be of the past.
Cheers,
Sebastiaan
Check out SigmundAI.eu for our OpenSesame AI assistant!