search

klabhub / neurostim

Design and run visual neuroscience experiments using Matlab and the Psychophysics Toolbox.
MIT License
5 stars 4 forks source link

Offline tool for reconstructing movie stimulus #186

Closed dshimaoka closed 3 years ago

dshimaoka commented 3 years ago

Added a new function to reconstruct a movie of a specified trial. Cf. noiseclut.reconstructStimulus

adammorrissirrommada commented 3 years ago

Hi @dshimaoka. I don't have the capacity to go through this in detail at the moment, or a means to test it, but it looks OK (very similar to my noiseclut solution, as you say). It doesn't interfere with anyone else's code, so it can't really cause immediate harm. Can you describe what you have done to test whether it is getting the image sequence right (e.g. a photodiode recording)?

I am OK with merging it as is, but perhaps include a warning that the reconstruction is relatively untested and should be confirmed by the user. On the other hand, that is always true and I probably should have included the same warning for my reconstruction!

dshimaoka commented 3 years ago

The test I had done is minimal: the function did not clash using a cic record produced by a 10minute presentation of movie. Any suggestion for more rigorous testing would be welcome. I was actually trying to confirm if the frame drop times could be inferred from inter-frame interval computed from digital signals sent with beforeFrame. The result was awful. Sending digital signal caused too many frame drops (~20times more than without digital output).