LCD-TVs and display lag

N

Npl

Veteran
Since I probably aim for getting a TV sometime soon, I would like to know about the lag many LCD-TVs introduce. Are specific manufacturers worse than others, can the lag be reduced by disabling postprocessing (100Hz+ [strike]conversion[/strike] guesswork), do the TVs actually allow disabling postprocessing ... are there any sites testing for this or atleast lists dealing with this issue?

Im looking at a Sony,Toshiba or LG 1080p set. Will be used for PS2 and PS3 games and maybe very rarely a movie (I haven`t had a TV for 2+ years and I ain`t missing them much)
 
Unfortunately, you're always going to have some lag on a 1080p set playing 720p games on PS3 (or 480i/p games on PS2) because the TV must enable the internal scaler to convert the video input to the native resolution, even in "game mode". Testing on my Sharp Aquos 52" 1080p set shows that Rock Band 2 on Xbox 360 has no measurable lag when the Xbox is set to output 1080p, but about 40 ms when the Xbox is set to output 720p. I measured about the same on PS3 with Guitar Hero Metallica.

I think that if you buy a TV with a "game mode" and play at the native resolution, you're very likely to have acceptably low lag. Anything else and it's hard to say without testing.
 
Just out of curiosity, how would one go about measuring input lag on a console? I'd like to see what sort of input lag my Panasonic 42" plasma creates. I can't say I've ever noticed any, but maybe my reactions just aren't fast enough :p
 
well, I can implement bilinear scalers that are only 1-2 scanlines "behind", so I still have hopes that some people in the CE industry can do the same.

Thanks for the reply, I feared there would be no simple answer to know a TV-Sets delay. But I hoped for a list of TV-Models with user-measured delay

@ShaidarHaran: I seen some vids where the controller and the TV are filmed at 60fps and the delay of a buttonpress and the reaction on screen is measured in frames.
of course you measure the whole chain from the controller through gamecode and display that way, so i`d would help knowing how much time the game/Console itself takes till you get the visual confirmation of your actions.
 
@SH: To measure lag on my TV, I just used the Rock Band/Guitar Hero calibration feature, where you press buttons on the controller when you see a visual cue on the screen.

@Npl: I've never understood why TV scalers are sometimes so slow either. My parents' Panasonic plasma (2006 model), for example, takes ~160 ms (!) just to scale the 720p output of the Xbox to the panel's native resolution (1366x768). Why?
 
I just bought a Sony Bravia V-Series, and do not notice any lag. I've heard the lower-end LG panes were also low lag. I've seen _lots_ of forum posts that indicate Samsung TVs are very high-lag across the entire product line (even with game mode).
 
The funny thing is on my 50" DLP I'm not even using the gaming mode. The problems with the Samsung tv's are with the LCD ones and not the DLP's.
 
For fun I just tested my new Samsung PN50B550 plasma TV, using Rock Band 2 on my Xbox 360. My results, repeated several times, ranged from 0-20 ms of lag, with the Xbox outputting at the display's native 1920x1080p resolution. To my surprise, the lag was also 0 ms (and more consistently so, which is probably more attributable to my reaction times than anything else) when I changed the 360's output to 720p and let the TV scale the picture. Apparently, Npl, at least one CE manufacturer finally did implement a fast scaler!
 
paawl said:
@SH: To measure lag on my TV, I just used the Rock Band/Guitar Hero calibration feature, where you press buttons on the controller when you see a visual cue on the screen.
Click to expand...
Dunno about you, but I have about 200 msec reaction time ... even if you are well at the edge of the curve you are still looking at 125 msec at the very best.

I wonder what exactly Rockband is measuring ... does it simply subtract an average reaction time from the real measurement? (Given the rather large spread of reaction times that's not a very good measurement then.)
 
MfA said:
Dunno about you, but I have about 200 msec reaction time ... even if you are well at the edge of the curve you are still looking at 125 msec at the very best.

I wonder what exactly Rockband is measuring ... does it simply subtract an average reaction time from the real measurement? (Given the rather large spread of reaction times that's not a very good measurement then.)
Click to expand...

Rock Band displays a "metronome"--a little swinging arrow--, and you're supposed to press a button when the arrow swings past vertical. That takes the reaction time out of it, because you can anticipate when it's going to pass vertical.
 
but the program has no way of knowing if you are anticipating or actually waiting till it goes past vertical
 
paawl said:
For fun I just tested my new Samsung PN50B550 plasma TV, using Rock Band 2 on my Xbox 360. My results, repeated several times, ranged from 0-20 ms of lag, with the Xbox outputting at the display's native 1920x1080p resolution. To my surprise, the lag was also 0 ms (and more consistently so, which is probably more attributable to my reaction times than anything else) when I changed the 360's output to 720p and let the TV scale the picture. Apparently, Npl, at least one CE manufacturer finally did implement a fast scaler!
Click to expand...
Plasmas generally have faster and better quality scaling than LCDs. Don't know why this is(maybe LCD manufacturers are more likely to skimp on scaling hardware costs). It was unnecessary but plasma manufacturers still added a game mode so people can see they have the same feature LCDs have.
 
Davros said:
but the program has no way of knowing if you are anticipating or actually waiting till it goes past vertical
Click to expand...

True, you can press the button any time you like, even before the arrow passes vertical, resulting in a negative lag measurement, but why would you try to cheat a program that's measuring the lag of your display? :?:
 
Still, if you wanted to be somewhat rigorous you can always record yourself pressing the button with a video camera that has both your hands and the screen in view.

Regards,
SB
 
Davros said:
you may not be doing it consciously
Click to expand...

While it is, I suppose, certainly possible for one to be unconsciously pressing a button early, out of a desire to minimize the lag measurement of one's fancy new TV, how would one know how early to press the button? It's not like there's any visual feedback of measured lag during the test. You simply press a button when the visual cue appears, then, after doing so several times, some sort of mean or median is computed by the program, and your game's input window is calibrated to that lag measurement. If you press too early, you're likely to get a negative lag measurement, which would clearly indicate that you're doing it wrong.

It's like pressing the button to stop a stopwatch when a runner crosses the finish line. You see the runner coming, estimate his speed, prepare yourself, then press the button when he crosses the line. Reaction time doesn't play into it, and, provided that you have the desire to make an accurate measurement (and not cheat), I see no reason why you would consciously or unconsciously stop the timer early.

Is there some other reason that you doubt the veracity of this measurement method?
 
You must log in or register to reply here.

Similar threads

J
4K displays announced at CES 2011
5 6 7
Replies
128
Views
32K
Silent_Buddha
S
Berek
Display Performance: Input Lag, Response Time, Latency
Replies
0
Views
4K
Berek
Berek
T
Sony PS3 Q+A (Leaked)
11 12 13
Replies
251
Views
70K
one
O
Back
Top