# Question about the timing of Screen Flip
**URL:** https://psychtoolbox.discourse.group/t/question-about-the-timing-of-screen-flip/812
**Category:** Uncategorized
**Created:** [April 18, 2013, 3:16pm UTC](https://psychtoolbox.discourse.group/t/question-about-the-timing-of-screen-flip/812 "2013-04-18T15:16:56Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@Allen\_Ingling](https://psychtoolbox.discourse.group/u/Allen_Ingling)
#### Post date: [April 18, 2013, 3:16pm UTC](https://psychtoolbox.discourse.group/t/question-about-the-timing-of-screen-flip/812/1 "2013-04-18T15:16:56Z")
I tried a simple code including the following lines. So the screen was flickering at 15 Hz, with 60 Hz of monitor refresh rate.
...
waitframe = 1;
ifi = Screen('GetFlipInterval',w);
tvbl = Screen('Flip',w);
tvbls = [];
k = 1;
while ~KbCheck
Screen('FillRect',w,128-127\*(-1)^k);
tvbl = Screen('Flip',w,tvbl+(waitframe-0.5)\*ifi);
tvbls = [tvbls tvbl];
k = 3-k;
end
...
I though the result of diff(tvbls) would be around .033 (i.e. 2\*ifi). But I found they were around .0267. But if waitframe was set to zero, i.e. the screen flip at every vertical retrace, diff(tvbls) became closely around .016. Why?
---
### Author:  [@dcnieho](https://psychtoolbox.discourse.group/u/dcnieho)
#### Post date: [April 18, 2013, 3:46pm UTC](https://psychtoolbox.discourse.group/t/question-about-the-timing-of-screen-flip/812/2 "2013-04-18T15:46:32Z")
That is funny. What operating system are you running (version, etc all the details) and what graphics card? You are using a fullscreen window, right?
Best,
Dee
> On Thu, Apr 18, 2013 at 11:16 PM, ptbmin \<[ptbmin@...](mailto:ptbmin@...)\> wrote:
>
> > I tried a simple code including the following lines. So the screen was flickering at 15 Hz, with 60 Hz of monitor refresh rate.
> >
> > ...
> > waitframe = 1;
> > ifi = Screen('GetFlipInterval',w);
> > tvbl = Screen('Flip',w);
> > tvbls = [];
> > k = 1;
> > while ~KbCheck
> > Screen('FillRect',w,128-127\*(-1)^k);
> > tvbl = Screen('Flip',w,tvbl+(waitframe-0.5)\*ifi);
> > tvbls = [tvbls tvbl];
> > k = 3-k;
> > end
> > ...
> >
> > I though the result of diff(tvbls) would be around .033 (i.e. 2\*ifi). But I found they were around .0267. But if waitframe was set to zero, i.e. the screen flip at every vertical retrace, diff(tvbls) became closely around .016. Why?
---
### Author:  [@Allen\_Ingling](https://psychtoolbox.discourse.group/u/Allen_Ingling)
#### Post date: [April 19, 2013, 8:41am UTC](https://psychtoolbox.discourse.group/t/question-about-the-timing-of-screen-flip/812/3 "2013-04-19T08:41:35Z")
I tried the same code on another computer, and the results looked OK. Both computers are installed with Windows 7. The difference is the one I tried yesterday has two graphics cards with the nVidia Optimus technique. It seems the Optimus technique messed up everything.
> --- In [psychtoolbox@yahoogroups.com](mailto:psychtoolbox@yahoogroups.com), "Diederick C. Niehorster" \ wrote:
> \>
> \> That is funny. What operating system are you running (version, etc all the
> \> details) and what graphics card? You are using a fullscreen window, right?
> \>
> \> Best,
> \> Dee
> \>
> \>
> \> On Thu, Apr 18, 2013 at 11:16 PM, ptbmin \ wrote:
> \>
> \> \> \*\*
> \> \>
> \> \>
> \> \> I tried a simple code including the following lines. So the screen was
> \> \> flickering at 15 Hz, with 60 Hz of monitor refresh rate.
> \> \>
> \> \> ...
> \> \> waitframe = 1;
> \> \> ifi = Screen('GetFlipInterval',w);
> \> \> tvbl = Screen('Flip',w);
> \> \> tvbls = [];
> \> \> k = 1;
> \> \> while ~KbCheck
> \> \> Screen('FillRect',w,128-127\*(-1)^k);
> \> \> tvbl = Screen('Flip',w,tvbl+(waitframe-0.5)\*ifi);
> \> \> tvbls = [tvbls tvbl];
> \> \> k = 3-k;
> \> \> end
> \> \> ...
> \> \>
> \> \> I though the result of diff(tvbls) would be around .033 (i.e. 2\*ifi). But
> \> \> I found they were around .0267. But if waitframe was set to zero, i.e. the
> \> \> screen flip at every vertical retrace, diff(tvbls) became closely around
> \> \> .016. Why?
> \> \>
> \> \>
> \> \>
> \>