LCD monitors and Input lag

[Peter Quain]

The issue of LCD monitors has been raised periodically on this list. 
Because CRTs have gone way of the dinosaur many labs use LCDs now for 
experiment presentation, and my guess is that some aren't aware of 
some possible pitfalls of doing so. My understanding is, basically, 
LCDs work differently to CRT monitors, and the key issue for display 
timing accuracy is that LCD monitors have onboard image processing 
engines that do their own thing with the frames sent to them by 
graphics card, prior to displaying them, and this processing can take 
(varied) time, with range that can be perhaps 0 to 70 ms across 
frames, and average lag which can be in the 30-40 ms range. The 
variation compounds the problem LCD screens may pose for some types 
of experimental psychology.

This means 1) that you don't have a clue what is going on with 
display timing unless you test your LCD monitors for input lag (and 
then it likely will fluctuate across trials anyway) ; and 2) Just 
because it shows a picture, you can't just treat an LCD monitor as 
though it is a CRT monitor for purposes of time critical paradigms.

Additionally, any concurrent audio would be out of sync because the 
audio is not routed through the display, and so would experience no delay.

Also, without testing the monitor, it may not be possible to trust 
the "refresh rate" setting at anything other than the native refresh 
(mostly 60Hz) even though Windows may provide an option for the 
monitor to run at a higher "refresh" (say 75Hz), which some 
researchers may choose in their experiment. I have seen results 
showing that when refresh is set to 75Hz on a 60Hz native refresh 
LCD, frames are redrawn every ~13ms (instead of ~17ms), however 1 in 
every 6 frames was skipped (no display). So, in 167ms only 10 frames 
were displayed, not 12. Looked like the engine was correcting back to 
native refresh. You test this with high speed camera (same as to test 
input lag).

As far as e-prime goes, this means that you could write a tidy 
paradigm where timing was tested as perfect on a good PC, and e-prime 
would log all durations as being so. However, at the display level 
the timing could be all over the place. Nobody would know, and 
effectively all the time taken to use e-prime for millisecond 
precision would be wasted. On monitors that have a big range of input 
lag, some paradigms would really be impossible to implement accurately.?

I'm no expert in this, have just been fishing round on the net. For 
anyone who might be interested, below are some useful links providing 
a little digestible background on how LCDs work, how they differ from 
CRTs, and how to go about testing for input lag. Note that to do this 
properly you need a CRT monitor as baseline. Don't be tempted to use 
another LCD, which would include using a laptop screen, and my advice 
would be to definitely use a PC (with dual head graphics card in 
clone mode) not a laptop.

----------------------------------------------------------------------
Good old Wikipedia defining input lag:

http://en.wikipedia.org/wiki/Input_lag

A basic primer on how LCDs / CRTs function, and differences:

http://www.bit-tech.net/hardware/2006/03/20/how_crt_and_lcd_monitors_work/

Here is an interesting site describing CRTs and LCDs (from gamer 
perspective). Navigate through the next few pages forward / backward 
using buttons down the bottom:

http://www.tweakguides.com/Graphics_7.html

How to test your monitors for input lag?

Here is a brief description, and a little counter program you can download:

http://www.flatpanelshd.com/focus.php?subaction=showfull&id=1229335064

Here is another description with some useful info about type of 
camera that is suitable:

http://www.avsforum.com/avs-vb/showthread.php?t=1131464

A neat site who say they have done a lot of testing re: input lag, 
and provide comparison output for many LCD monitors. Note, lags might 
be different on your monitor even though it is same model tested. You 
need to test each individual monitor:

http://www.digitalversus.com/duels.php?ty=6&ma1=35&mo1=121&p1=1303&ma2=284&mo2=326&p2=3097&ph=12

-----------------------------------------------------------------------------
Peter

-- 
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.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/eprime/attachments/20110712/92432089/attachment.htm>