Global frame counter and accurate timing programming

Dear members

I am currently programming for a very timing-snsitive experiment. I was wondering whether there is a way to setup a global frame counter, which starts upon the program initiated, keeps counting till the program terminated, so that I could check and specify different actions(e.g., present images or text) accourding to the specific frames reached. In other words I am looking for a tic toc timer or internal cpu clock alternative(e.g., using the function GetSecs).

The reason behind trying to sdopt this rather cumbersome method is due to the fact that according to the Accurate timimg demo it seemed better to count the number of frames rather than specify the duration of each event. At the same time, a potential concern for using this frame counting strategy is I am not sure whether dropping frames (perhaps due to hardware limitations) would messed up the sequences entirely. For example, using a 240Hz display dropping 1 frame would advance the entire sequence forward by 1/240 second, as more frames were dropped the effect would accmulate resulting further timing deviations? I am pretty new with this type of programming, perhaps this is a misunderstanding?

Any comments or suggestions would be great! Thanks in advance!

p.s. I am using psychtoolbox for better timing accuracy

[Jan]

What are the "frames" you want to count? I cannot guess, what a "global frame counter" is.

[Yu Takahashi]

Thanks for the reply, I was trying to count the total number of screen refreshes/total number of flips

[Jan]

Please remember, that the readers do not know anything about what you are doing. I cannot know, what you call "screen refresh" or "flip".

[Yu Takahashi]

@Jan Sorry about the confusion, I might have skipped some crucial explainations.

By "screen refresh", I was refering to the refresh rate of the computer monitor, e.g., for an ideal 240Hz gaming monitor, the monitor would theoretically refresh 240 times every second. "flip" is just another way (perhaps only in my field) of refering to the same thing.

I was just tryin to find a way to count the total number of times the monitor actually refreshed, so that I could specify, for example, present an image/texture from the first refresh to the 500th refresh upon the initiation of my script.

[Jan]

We needed an equivalent count and failed to implement it eliably enough in software. Therefore we measured the display frequency externally with a small photo-sensor in a corner of the screen.

[Yu Takahashi]

@Jan I see, thanks for the advise. Would you mind directing me to some further readings?

My instinct would be a oscilloscope with a photodiode, but is there a way to ensure that the image in the conor is in-sync (i.e., refresh at the same moment) with each monitor refresh ?

[Walter Roberson]

There is very slim hope of doing any of this with regular MATLAB; you would either need psychtoolbox facilities or build your own that are similar to psychtoolbox. Psychtoolbox already works are to ensure that screen flips are in synch with monitor refreshes.

[Yu Takahashi]

@Walter Roberson Thanks! I think I might try to contact someone from the dev team.