LCD lag, who *here* can test it

[davet11]

I am personally still frustrated with the questions around LCD lag. Some people say they notice it right away, some people say they notice it but can live with it, others say that it doesn't exist at all and that the people who are claiming it exists don't know how to properly set up hardware. I personally think there needs to be some kind of effort put forth by the review-site community to at least get some answers.

For those of you who aren't familiar with LCD lag, it isn't ghosting and it isn't motion blur, this is an actual delay (albeit a small one) from when you move your mouse, type something, etc. to when it appears on the screen.

Following are as many links as I can remember at the moment relating to the issue, as you can see, there's always a huge dispute. I did find it interesting in one of the threads that the dispute turned from "there is no lag LCD monitors, you don't know what you're talking about" to "well, 30ms doesn't really matter". As somebody who doesn't have a tonne of cash to drop, but still likes to make solid investments, this is really something that I would like to have more information on. I'm hoping that some people here still have LCDs as well as CRTs with a dual head card and can do some tests for us. Maybe if we actually address the issue intellectually we can find out if:

a) this is a problem in all LCDs
b) this is a problem in larger LCDs
c) this is a problem within certain brands of LCDs
d) this is a defect that only effects some LCDs

I'm sure somebody on this forum could also give a real reason for why this occurs, if it occurs. After all, we have the highly technical forum which sometimes brings me to tears because I feel so inadequate.

The links:

http://forums.whirlpool.net.au/forum-replies-archive.cfm/377111.html

http://www.hardforum.com/showthread.php?p=1028035239#post1028035239

http://weblogs.asp.net/ryanw/archive/2004/09/23/233681.aspx

http://s88012115.onlinehome.us/TFTDelayBeweis.avi

There's more on the web, but those links will lead to others as well ... let's try to get a real solid answer, I guess I could start with my CRT and my gf's Toshiba laptop!

 

[n7]

I don't really notice it anymore, but i have the lag issue on my 204T.

When i first got it, i sucked hard in UT2k4 till i finally adapted my shot to make up for the delay, though back then, i'd have been called crazy if i said it was lagging.
Ironically enough though, the very first thing i though was, wtf, my monitor is lagging...

It doesn't ghost at all, but for some reason my 20ms 17" LCD is better for fps games than my 16ms 20" LCD...because of that "lag"

 

[Matt2]

No lag here. I think maybe lag is an issue on the bigger LCDs??

 

[akugami]

Hrm...I'll have to test again on my 2005FPW. I always thought it was a little slower...but I attributed that to a larger LCD and the fact I'm playing twitch games on it. Using a 17" Dell LCD at work and no lag here, but this is work related stuff. Or as work related as you can get browsing a forum at work.

 

[n7]

Originally posted by: Matt2
No lag here. I think maybe lag is an issue on the bigger LCDs??

I'd tend to agree.
From what i've seen, the smaller ones seem to have no or very few issues, whereas the larger ones have more...

 

[davet11]

I'm thinking there has to be a way to determine which LCDs are better/worse for this, aside from large panels being worse than small ones. As well, does this affect LCD tvs? That would really piss me off if I had bought one planning to play console games ... it just seems like no manufacturer has really copped to this.

 

[xtknight]

Are you sure it isn't a pixel rise/fall time problem, and if so how are you sure? Too much of a lag? What I notice between a CRT and LCD is the speed of typing text. It always instantly appears on the CRT, but that's probably because of the 12x (~avg.) slower pixel rise time on the LCD. Particularly on S-PVA panels like the 2405FPW, in the darker transitions (0-50 RGB) reach upwards of 60 ms. 95% of 17" monitors use TN, which doesn't have the problem (usually 16 ms. in that range).

 

[mcmikemc]

Months ago my awesome 19" Sony CRT started to crap out on me. I went looking for a brand new Trinitron CRT and could not find one anywhere so I looked into LCDs. After reading a ton of info I bought a L90D (considered a great gaming LCD panel). I know for a fact that my L90D had lag in it. You see I play Counter Strike 3+ hours a day. Once I started using the L90D my game playing skill seemed to go away. I could not for the life of me figure it out. Then one day my buddy brought his CRT over and I hooked up both at the same time and I switched between the two, rather quickly, with CS. Once I had them booth side by side it was painfully obvious that my L90D lagged.

I tried the L90D with VGA and DVI at every possible resolution and it was still there.

I could play single player games and I could get used to the lag and really enjoy the game. I played the single player portions of Painkiller and Far Cry on the L90D and loved it. But with Counter Strike, where every ms matters, I could not stand it.

So what did I do?

I sold the L90D and bought a Viewsonic G90F 19" CRT, 2005FPW and found a few used CRTs in town. I now have 4 CRTs in the attic so I should be set for competitive gaming for 10 more years.

 

[L00PY]

2405FPW (1920x1200)+ 20" CRT (1600x1200) via dual head. I did the screen drag across spanned desktops and couldn't see any lag.

Two caveat:

The CRT isn't at the same level as the LCD -- it's much too heavy to place up there.
I'm not a huge FPS fan.

 

[imported_thefonz]

I do notice LCD lag on my screen, It's a 25 ms response, and I can tell its monitor lag becuase, when I play older, like the desert combat mod for battlefield, it seems to skip a frame, or delay a frame, even though my system is doing well over 100 frames easily.
Next paycheck im getting a 19 widescreen with 12ms though

 

[davet11]

Originally posted by: thefonz
I do notice LCD lag on my screen, It's a 25 ms response, and I can tell its monitor lag becuase, when I play older, like the desert combat mod for battlefield, it seems to skip a frame, or delay a frame, even though my system is doing well over 100 frames easily.
Next paycheck im getting a 19 widescreen with 12ms though

Please read the original post, this isn't the "ms" that the manufacturers advertise. This is something else, another delay that is basically unaddressed by LCD manufacturers ...

 

[xtknight]

Skip a frame/jitter isn't pixel rise/fall. Thefonz, it's probably something else like the CPU/mem bottlenecking the GPU. Turn on VSync or a maxfps setting.

 

[MrJingleBells]

triple buffering will also cause input lag, so make sure you arent using such an option. not saying thats the problem, but it could account for why some people have more of an issue than others. the tendency is to get a huge lcd, and either 1 monster card, or 2 in sli, and run with every option enabled. in the case of triple buffering, it immediately will cause gameplay issues.

 

[clickynext]

My LCD has ghosting with its 40+ms response, but I'm not really bothered by it, and don't really notice it. It doesn't really affect my CS gaming at all, compared to the CRT I used to use.

 

[darXoul]

Being sarcastic or serious?

 

[clickynext]

Originally posted by: darXoul
Being sarcastic or serious?

Serious. It's a noticable thing, but honestly it doesn't cause any lag in movement, and the afterimage stays on a bit longer, but the new image is displayed just as quickly, so it doesn't really delay anything. It's an old LCD I got for free, 15". I came from a 19" samsung CRT which was good quality, but really I havn't missed anything about it except the resolution.

 

[davet11]

Originally posted by: clickynext

Originally posted by: darXoul
Being sarcastic or serious?

Serious. It's a noticable thing, but honestly it doesn't cause any lag in movement, and the afterimage stays on a bit longer, but the new image is displayed just as quickly, so it doesn't really delay anything. It's an old LCD I got for free, 15". I came from a 19" samsung CRT which was good quality, but really I havn't missed anything about it except the resolution.

Once again, lag is different from ghosting or motion blur, please have a look at the links above.

 

[xtknight]

So, any videos of 2 CRTs in clone mode with the digital clock method?

 

[tm101]

There has been reports of the Fujitsu-Siemens P20-2 lagging behind a CRT and a laptop TFT. The owner has made the following pictures for comparison, which clearly show the lag of the big TFT: http://www.stud.tu-ilmenau.de/~maei1ii/p20bilder/

Generally I think that most if not all TFTs use a buffer of at least 1 frame, which means a 16.6ms lag at 60hz or a 13.3ms lag at 75hz. Since Overdrive displays have to calculate the next frame out of the last ones (according to Eizo they use the last 3 frames to calculate the next one) they will surely need some time for the calculation and some buffer to prepare the next frame to be displayed. When comparing my VP930 to a CRT via dual-head I am not able to notice any obvious lagging behind so I guess the maximum lag is 1 frame. Most people are not able to discern less than 20ms anyway. As a musician with somewhat trained ears I can barely identify an audio lag of more than about 10ms, so a 1 frame lag should still be fast enough for my untrained eyes.

 

[xtknight]

If the Overdrive chip could not process pixels in real-time, wouldn't the effect of lag be cumulative over time?

 

[Matthias99]

Originally posted by: xtknight
If the Overdrive chip could not process pixels in real-time, wouldn't the effect of lag be cumulative over time?

It's not that it can't process them in real time -- but instead that it has to spend some time doing additional processing on the image (which would introduce a fixed lag in the image), or that it buffers multiple frames (in order to do more accurate processing). So the monitor is actually displaying what the video card output one or more frame 'ticks' ago.

That said, I tried to test this between my Viewsonic PS970 (19" CRT) and my VP930b (which uses an overdriven panel), and while I can see a *tiny* amount of 'lag' on black->white transitions, I am pretty sure this is just the latency of the pixels themselves rather than any additional lag being introduced. If there's an additional latency, it's so small that I can barely see it (so it's probably only one frame, if even that).

 

[xtknight]

Would it be correct that because CPUs aren't always perfectly precise in timings, that you wouldn't be able to rely on those images to determine the amount of lag just by reading the value reported in the photo? Granted you would be able to tell if there was lag or no lag, just not how much.

 

[Matthias99]

Originally posted by: xtknight
Would it be correct that because CPUs aren't always perfectly precise in timings, that you wouldn't be able to rely on those images to determine the amount of lag just by reading the value reported in the photo? Granted you would be able to tell if there was lag or no lag, just not how much.

Generally speaking, the hardware CPU timer (what MS calls a 'performance timer') is accurate down to a microsecond. At least. It usually only takes a few clock cycles to read the value in from the hardware timer, and a clock cycle on a 2Ghz CPU is .0005 microseconds.

Now, the timestamp you get has no particular absolute value -- the system clock is usually only accurate to the second. And there could be variable lag in actually displaying the values on the screen, so it's not perfectly accurate in that sense (but that should be a small lag, since it's just transferring the data from the program to the video driver to the framebuffer). But if you read the value at time X and at time Y and the time values you get are 1,532.5 microseconds apart, you can be very sure that between 1,532 and 1,533 microseconds passed between when you read those values.

Unless the video card's driver has a significant delay in updating the second display (or something like that), I would think the timer being displayed on-screen would be accurate to at least the millisecond scale.

 

[bobsmith1492]

Well, LCDs are digital while CRTs are analog; digital processing will always be slower than analog processing, so there's really no getting around it.

 

[tm101]

I am quite sure Overdrive can process in real-time, but most likely it has to start processing at least the first frame (with the P20-2 maybe the first 3) before displaying it. So you would start with a lag of 1 to 3 frames before the monitor starts displaying from then on you have realtime-processing but keep carrying on the lag from the very beginning. But even without Overdrive most TFTs buffer at least 1 frame anyway. But even if the monitor analyses 3 frames to calculate the next one (like Eizo claims) it does not necessarily mean you have a 3 frame lag. The Eizos most likely buffer just 1 frame before displaying it like most other TFTs, but they probably also buffer the last three frames that already have been displayed to use them for calculation. That means that the very first frames the monitor displays after switching it on (or changing display mode) lack those 3 frames for calculation, but who would recognize that anyway?

Concerning the timing of dual-head outputs you just have to look at the pics posted by that P20-2 owner. He had both the TFT and the CRT connected to his laptop. Still the difference between the P20-2 and the laptop screen was higher than the difference between the CRT and the laptop screen. This corresponds to the actual lag of the P20-2. Besides that several other people have reported a considerable between moving the mouse and seeing the mousepointer move on screen.