List Serv Threader V1.4.02

Date: Thu, 13 Oct 2005 14:05:19 -0500
From: John Curtin <jjcurtin@wisc.edu>
Subject: [DMDX] Re: Display error oddity

------=_Part_2421_7062338.1129230319124
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Thanks Jonathan,
Yes, the machine that was not generating the problems was running
3.1.2.1<http://3.1.2.1>All makes sense now.
 J

 On 10/13/05, Jonathan C. Forster <jforster@psy1.psych.arizona.edu> wrote:
>
> At 12:42 PM 10/13/2005 -0500, you wrote:
>
> I have a question about display error messages that are generated in the
> azk output. Basically, the same script generates a very different pattern=
 of
> display errors (or no errors) on two different machines. Here is the
> scenario. This is the script:
>
> <ep> d30 f1 <cr><t 2000><id keyboard><id mouse><vm
> 1024,768,768,32,60><nfb><eop>
> +1 * "Test1" <msfd 500> /;
> +2 * "Test2" <msfd 500> /;
> 0 "End";
>
> If I run this script on one computer and let it run with no responses, it
> generates the following output:
>
> Subject 1, 10/13/2005 12:24:57 on HEINEKEN, refresh 16.66ms, ID 1
> Item RT
> 1 -2000.00
> ! Display error at msec 2561.41, tick 152 in item 2, frame "Test2"
> ! moved into video memory 60 ticks late
> ! (previous frame's duration will have been longer)
> 2 -2000.00
> ! Display error at msec 4610.93, tick 275 in item 0, frame "End"
> ! moved into video memory 61 ticks late
> ! (previous frame's duration will have been longer)
> This surprised me b/c using a d of 30 should be way more time than was
> needed to prepare to present text. Also the 60 & 61 additional ticks list=
ed
> in the display error seemed way too long. Then I realized that the time o=
ut
> of 2000 would obviously be delaying things if the participant responded r=
eal
> slow (or didnt response as in this simulation). If I change the time out =
to
> 500 the delay errors go away.
>
>
> Yeah, the time for the next frame is calculated from when the previous
> display ended so a timeout that's longer than the display will screw up t=
he
> timing. A 500ms frame plus a 500ms delay minus a two second response time=
out
> gives you around 60 ticks.
>
>  I thought problem solved. But, if I run the same original script on a
> different machine (with the time out of 2000), I DO NOT get any display
> errors:
> Subject 1, 10/13/2005 12:28:35 on REDSTRIPE, refresh 16.68ms
> Item RT
> 1 -2000.00
> 2 -2000.00
>
> If the time out was the source of the display errors, then it should occu=
r
> on both machines, correct? But if the timeout is not the source of the
> error, why does making the timeout <=3D to the duration of the item make =
the
> display errors go away on the first machine?
>
>
> Could be a version thing, I just recently fixed the way <cr> was working
> with <d> and your machines may have different version of DMDX on them. Pr=
ior
> to 3.1.4.1 <http://3.1.4.1/> a pretty un-rigorous piece of code would hav=
e
> not had display errors when the subject's response went over the display
> duration when <cr> and <d> were used together. As long as the display was
> longer than the RT things will have been fine (as long as you didn't have=
 a
> really fast CPU) but for poorly formed items like you have there it will
> have been scheduling the display from the time that the previous item
> finished, not it's display.
>
>
>  /"\
>  -jonathan (j.c.f.) \ /
>  X
>  ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL / \
>
> You probably wouldn't worry about what people think of you if you could
> know how seldom they do.
>
>  - Olin Miller.
>
>


--
John J. Curtin, Ph.D.
Assistant Professor of Psychology
University of Wisconsin
1202 West Johnson St.
Madison, WI 53706

Email: jjcurtin@wisc.edu
Website: http://dionysus.psych.wisc.edu
Office: 608-262-0387
Lab: 608-262-5621
Cell: 608-217-6221
Fax: 608-262-4029

------=_Part_2421_7062338.1129230319124

John J. Curtin, Ph.D.
Assistant Professor of Psychology
University= of Wisconsin
1202 West Johnson St.
Madison, WI 53706

Email: jjcurtin@wisc.edu
Website: http://dionysus.psych.wisc.edu
Office: 608-2= 62-0387=20
Lab: 608-262-5621
Cell: 608-217-6221
Fax: 608-262-4029=20 ------=_Part_2421_7062338.1129230319124--