CRT vs. LCD

[Peter Quain]

Date: Fri, 19 May 2006 14:10:31 -0700
To: DMDX at psy1.psych.arizona.edu
From: "Jonathan C. Forster" <jforster at psy1.psych.arizona.edu>
Subject: [DMDX] TFT (or LCD) displays caveat emptor
Reply-To: DMDX at psy1.psych.arizona.edu
Sender: DMDX-owner at psy1.psych.arizona.edu


   As I was testing a LCD display (alternatively called an TFT or 
Flat Panel display) here I came across another previously 
unconsidered dimension of them that is a very serious consideration 
in their use as tachistoscopic display devices.  And that is that the 
LCDs all have a digital signal processor (DSP) in them sitting 
between the computer's output and the pixels of the display and the 
speed of that processor can interfere with tachistoscopic 
displays.  For instance when I first started testing this LCD the 
script I chose was using a 1024x768 display mode whereas the native 
resolution of this panel is 1600x1200.  So the DSP has to take the 
1024x768 signal and stretch it onto the 1600x1200 array of pixels 
that actually constitutes the display.  It turns out that this device 
can't keep up with even a moderately fast tachistoscopic display and 
was producing tearing at even pedestrian 4 tick displays.  I was 
toggling the background color between black and white and despite 
DMDX detecting no display errors I could plainly see tearing on the 
display.  Once I realized what was up and switched DMDX to using the 
native resolution of the panel the tearing went away.

   So one must be very careful to use the native resolution of an LCD 
when using DMDX.

   A further consideration that hasn't been tested yet but is likely 
to be the case is that my monitor is using a DVI cable and not the 
old 15 pin analog VGA cable that most displays are using.  The DVI 
cable is used here because high resolution displays (ie 1600x1200 and 
beyond) start to jitter and have all sorts of nasty 
problems.  Because the same DSP mentioned above has to take the 
analog video signal and process it it stands a very good chance of 
running into the same problems mentioned above.  So even if someone 
was using the native resolution of the panel the use of an analog 
signal cable could very well introduce the same sorts of tearing problems.

   Fortunately this is relatively easy to detect without additional 
test equipment beyond the human eye.  Basically toggle the background 
color at a rate DMDX can handle without display errors and you should 
see an even flickering display.  If you see periodic horizontal 
discontinuities then your LCD has a problem.

f30 <!VideoMode desktop> <vm 1024,768,16,0> d3
0 <dbgc 0> <dwc 255255255> <ln -1> "display tearing test" <cr> <set c1=100>;
3   <delay 3>  %4 <bgc 255255255>  /  <dec 1> <bicgt 1,0,3>;
0 "end" l;


                                                /"\
  -jonathan (j.c.f.)                            \ /
                                                 X
     ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL  / \

   Any given program, once running, is obsolete.



At 01:40 AM 5/03/2010, you wrote:
>I had a discussion yesterday with a collegue from another lab about
>display types. I am always using CRT displays for my experiments
>because they have higher refresh rates and no decay in which color
>values might differ from what you programmed.
>However, this guy was bringing forward an argument I have not thought
>about so far: For LCDs you have a stable overall picture at any given
>moment. A CRT display virtually draws one point after another, i.e.
>you never see an entire, let's say circle but your brain makes you see
>it as the point and lines are drawn so rapidly. I was shown photos
>made by a digital camera with really short shutter times (less than 5
>ms) and for CRT displays you can never see entire stimuli whereas for
>LCDs you do.
>
>So I was wondering what kind of display you are using in your labs and
>what your experiences are. Maybe there will develop a fruitful
>discussion.
>
>Cheers,
>Tobias
>
>--
>You received this message because you are subscribed to the Google 
>Groups "E-Prime" group.
>To post to this group, send email to e-prime at googlegroups.com.
>To unsubscribe from this group, send email to 
>e-prime+unsubscribe at googlegroups.com.
>For more options, visit this group at 
>http://groups.google.com/group/e-prime?hl=en.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.