Thought I'd share a comment on ESR, might be interesting to some people.
Comments, your LCD experience etc. appreciated.


http://www.esreality.com/?a=post&id=1296217#pid1296217

Ghosting effect, or tearing?

There is no refresh (hz) on an LCD display, so it's a moot point.
However, the "response time" on the pixels and "input lag" on your LCD display are *very* important. It's nothing you can configure or change, so you have to check this *before* you buy the display.

The math seems simple enough. For instance:

1000ms/77fps=12.9ms per frame

However, your LCD manufacturer is most likely stating a *best-case* response time for a full black to white pixel transition. This used to be the norm, but gray to gray transitions are much slower. An advertised b&w transition of 8ms might take up to 30 or 40ms for a g2g transition (keep in mind that color is added by a filter, the pixels themselves are 256 levels of black to white).

So if possible look for g2g response times in the specs, as those are crucial.

So, where does it leave us with respect to your wonderful brand spanking new 8ms LCD display. Well, it *might* be able to display that 12ms/77fps frame before the next frame is rendered *if* there aren't too many g2g transitions. Better still, response time, as defined by ISO, includes both the *raise time* and *fall time* for the pixel change. Some manufacturers will only state one of them meaning you could end up thinking you bought a shiny 8ms display when it's actually closer to 16ms -- best case.

Confused? You're in good company, but it doesn't end there. Recent technologies, commonly called Overdrive, have reduced g2g response times. Each manufacturer has it's own technique and name for it. Samsung calls it *MagicSpeed*. Look here for an overview and check out the excellent articles on response times and LCD screens tested head to head:

http://www.tftcentral.co.uk/advancedcontent.htm#response
http://www.tftcentral.co.uk/specs.htm
http://www.tftcentral.co.uk/panels.htm


That leaves *input lag*. Yes, you read it correctly. LCD input lag refers to the time it takes from when the signal enters the LCD monitor, becomes processed and displayed on screen. It *can* be as high as 50-60ms on some LCDs and manufactures generally don't state this figure in their monitor specs.

*Scaling* is one of the pre-processing techniques involved to set up your image. Your LCD screen comes with a fixed physical resolution (1024x768 or 1280x1024 are common unless it's widescreen). Using any other screenmode requires complex scaling to match the physical resolution of your display. Keeping that in mind it is *plausible* that using scaled screenmodes will induce additional input lag as it requires more processing time. The lag might be negligeble, but it's *probably* a good idea to use a mode matching your LCD's physical mode if your PC hardware can deliver your desired framerate. Also keep in mind that scaling up is always more difficult than scaling down.


As you can see even an advertised 4ms response time is no guarantee your LDC will be able to accurately and quickly display our 77fps game. I've only tested 8ms displays myself and those have been disapointing. Maybe 2ms will change that, but because of all the uncertainty I would not reckomend LCD for hardcore gaming. But at the end of the day the only way you can be sure is to test it for yourself and make up your mind. Well that's easy enough, I'll stick to my 1995 vintage Dell 17" for now . <sigh>



/PaRadiZer

i have a 21" elements crt that ive used for years, i bought it back in 99 i think, for like 350. ive also got 2 apple cinema displays, anyone have any luck getting a smooth game on the apple screens?

I think my Cathode Ray Tube is radiating too many cations at me, I feel frail and ill after spending hours in front of it.
Keep us posted about decent LCDs. B)

Judge by your own eyes, don't listen to others opinions. Check out some LCDs at your friends or stores.
Modern LCDs not as bad as some people want them to be.

Once I was trying to use 40ms IPS LCD and it was playable, but some ghosting was noticeable with fast movement.
Right now I play at 20ms IPS LCD and can't see the difference between it and CRT. Go figure - 12.9ms rule seems not working in my case.

I think most of the modern IPS/TN LCDs will be fast enough for QW, just stay away from PVA/MVA, they tend to be slower.

Samsung SyncMaster 930BF here and it works just fine. IIRC It's 8ms or something like that..No ghosting or tearing or anything. Just really smooth and nice, easy on the eyes too.

samsung syncmaster 206BW here, 2ms gtg, 1000:1 - smooth ride, no ghosting and stuff....sweeet beauty

thanks.

Bought Samsung 245BW, TN panel, 1920x1200.
It's the fastest 24" LCD by TFTCentral.

Set up together with CRT Dell 17" to measure lag, both at 60Hz. Run http://tft.vanity.dk/inputlag.html

From 15 shots taken at 1/500 sec:

- 11 were equal (example)
- 2 shots CRT was ahead (example)
- 2 shots LCD was ahead (example)

So lag can be considered as zero comparing to old 60Hz CRT.

Comparing to previous 19" LCD, which was on average 1 frame behind CRT, 245BW has that CRT-like instant feel when moving crosshair.

In-game shots:

LCD don't lag at all, but shows some ghosting

I'm sorry but there is a massive difference in lag between any CRT and LCD monitor. I have a 2ms 22" Samsung 226BW and next to my old 19" crt running 800x600 at 144hz, it feels like shit in game. It's not so bad now that I'm used to it, but seriously, CRT feels SO much better.

You mentioned you ran both at 60hz.. That's silly. You aren't going to see any differences there. The advantage with CRT is the refresh rates and the fact that you can use lower resolutions without it looking like shit due to native resolution etc.. Put a CRT at low res and 100+hz next to any LCD at 60/75 Hz playing the same MVD demo at over 100 fps and tell me there's no difference :p (or even better, play with over 100 fps on those and feel the difference)

Massive? Well, let's see.

You say you playing at 144Hz at CRT vs 60Hz at LCD.
Means average delay for frame at LCD is (1000/60 - 1000/144)/2 = 4.86 ms
5ms difference in ping is truly earth-shattering, guess you can rape dag with such advantage ... or may be not? What do you think?

I don't have any problems using native 1920x1200 resolution, since timedemos show 1000+ fps (yes, thousand) even with great-looking QRP textures.

CRT pros:
- 5ms better response time
- doubles as a space heater

LCD pros:
- high-resolution
- widescreen, wide fov, you see more stuff on the screen
- less strain on the eyes
- takes less space on a table

While 5ms may theoretically affect result of the match, wider fov for free can also be an serious advantage.

You forgot to consider the length of time it takes for an LCD to visually change the screen, so it feels even worse

Minimum latency between frames is 17ms at 60Hz and 7ms at 144Hz, that's 10ms difference. In addition there is image processing latency and the latency for the pixel to change on LCD, so it's more like 20+ms vs 7ms and in quake you can call that massive. The best way though is simply play with both and feel the difference, I know at least 3 players who switched back to CRT for quake and no one would do that unless it'd be clearly better for playing.

i have both crt and lcd connected, i use crt for quake
and yes, the difference is massive. That's why i have this huge "box" occupiing 1/3 of my table's area

I didn't. It's likely 7.5ms for Samsung 245BW, though I can't catch it with my measurements.


1000ms/60Hz = 16.(6) not 17
Frames not always will be full 10ms behind, on average delay will be half of that or 5ms.


13ms is massive ... could be, but certainly not for me. My ping is usally 40-100ms and reaction time is 160-200ms. I would not play any better or worse because of these 13ms. And less eye strain is a bonus.

Anyway, I don't want to convert anyone to LCD or something like that. I did some tests, posted results and that's it.
If you like your CRT and think it gives you massive advantage ... by all means go for it.

High framerate is not just about latency. It also helps in sensing the speed of movement. Quake is not whack-a-mole. Have you ever played with 154Hz? Even if there was no performance gain, it just feels so damn nice that I'd never trade it.

I bought a non-TN panel that is great for anything else and kept CRT for games and as extra desktop space.

What do you mean? How do you come at this 'average'? The important thing is the time between frames, how quickly you can respond to a change you see on the screen. Input can be concidered continuous and if fps >> refresh rate then we can think that the input values are read at the 'last moment' before the display adapter refreshes. At 1000 fps input values can be 1ms old at highest. (And 16.666... rounds up to 17 )

Oh, cmon blAze, it's simple math.
- when CRT refreshes, LCD could be behind by 0-17ms, average 8.5ms
- when LCD refreshes, CRT could be behind by 0-7ms, average 3.5 ms

8.5ms - 3.5ms = 5ms LCD behind on average



To test above calculations you can take a Riemann sum of timeline.

Length is 660 pixels from 1st to last CRT refresh.

Sum = 10/660 * 15ms + 23/660 * -2ms + 33/660 * 5ms + 24/660 * 12ms + 9/660 * -5ms + 33/660 * 2ms + 33/660 * 9ms + 5/660 * 16ms + 28/660 * -1ms + 33/660 * 6ms +19/660 * 13ms + 14/660 * -4ms + 33/660 * 3ms + 33/660 * 10ms + 33/660 * 0ms + 33/660 * 7ms + 14/660 * 14ms + 19/660 * -3ms + 33/660 * 4ms + 28/660 * 11ms + 5/660 * -6ms + 33/660 * 1ms + 33/660 * 8ms + 9/660 * 15ms + 24/660 * -2ms + 33/660 * 5ms + 23/660 * 12ms + 10/660 * -5ms = 5ms



As I you can see above, sometimes CRT is ahead, sometimes LCD is ahead, on average difference because of refresh rates will be 5ms for 60Hz vs 144Hz. Plus input lag, which can be as low as 1.6ms for 226BW or 7.5ms for 245BW or as high as 59ms for 245T (PVA panel).

About respond to a change you see on the screen - catch some sheep here.

Yeah all these numbers are nice hell but I'd rather trust my eyes thanks.

200hz =! 7ms

Sure, it's your choice.

A and B squares are the same shade of gray. Would you trust your eyes on this one too?




Someone said that 200Hz == 7ms? Who?
200Hz or 5ms
(16.67ms - 5ms) / 2 = 5.84 ms faster than LCD

It would be nice to have 120Hz 1920x1200 LCD with zero or close to zero input lag, overall it can be within 1-2ms from the fastest 200Hz CRTs.
120Hz already there at TV LCD sets, so PC market should catch up in a year or two.

Isn't that an optical illusion? Come on mate this is completely different. To any gamer CRT would feel smoother and "better" when setup correctly than any 2ms LCD monitor. Yes newer monitors such as 226bw are generally fine but they just can't beat a good CRT for games.

I compared my LG Flatron T910B to my mates Samsung 226BW and we both agreed that for fast paced fps games and ESPECIALLY Quakeworld CRT is better.

BUT the difference is not that HUGE as some people claim. I could play qw easily on 226bw.

Yeah, it's an illusion and it's different, just to illustrate that our eyes isn't absolutely precise instrument.
And because difference in response time between modern LCD and CRT is getting narrower, numbers can help us understand it better - you can't rely only on eyes to measure small intervals like 10ms or so.

If difference is small enough to be compensated by wider fov advantage ... why not?

4:3 vs 16:10 (fov 115 vs 127)

Ok so you are trying to convert us

Well when this one dies I'll buy one for sure.

16:10 vs 4:3

Math may be simple, but I'm still not at all convinced that your calculation accurately reflects the real world situation in it's entirety. At 30Hz your average would be at ~13ms which doesn't sound like a huge number either, yet which of us would concider 30Hz a competitive setup? Like Kalma said, Quake is not a whack-a-mole, it's a continuous game. Perhaps the percentage of the times a crt player gets an update before an lcd player would be a more descriptive number?

You can set any fov you want on 4:3, perhaps I'm missing your point on that.

Hm, have you read http://wiki.qwdrama.com/Widescreen_Guide ?

Your screenshot has the same width as with 4:3 resolution, but less information vertically - that's not correct setup. Increase fov to keep the same height and get some extra width from 16:10.

Also, if you post resized screenshots, please resize them with the same factor, 0.5 for example, otherwise it's misleading.


If you see error in my calculations, let me know. I never tried QW at 30Hz, I assume that aim skill wouldn't suffer that much, but the problem with low Hz is that you will likely see choppiness during fast turning/moving. Nobody would want to play on such setup.
Can you notice such choppiness at 60Hz?


75% of time CRT ahead
20% of time LCD ahead
5% of time equal

But the whole point was how many milliseconds CRT is ahead. Say, 50 ms is unnaceptable but 10 ms could be.


Yes, but in this case you will compress most important central part of the screen. Should be obvious, but I can make a sshot.
The point of having 16:10 is to get some extra space on sides for free, without losing resolution in the middle.

Btw, what's your sheep score? I'm asking because it's interesting if CRT really improves reaction time. That would be an advantage.

If anyone with dual CRT/LCD setup can test it - please do so, and post both numbers. I can't get past 144ms on 245BW LCD.

http://www.bbc.co.uk/science/humanbody/sleep/sheep/

Yes they are resized with same factor.

My point in posting that pic was that aspect ratio does not really matter. You can change fov as you wish. Only resolution and size of the screen is relevant (and many players would argue they are not that important either). Or would you rather play on 20" 16:10 than 30" 4:3? Would you perhaps tape the 30" screen so that it has 16:10 aspect ratio?

There are practical and economical reasons to prefer wide aspect ratio screens, but 16:10 is not bigger or smaller than 4:3. If for some odd reason current upcoming screen tehcnology would require 1:1 aspect ratio, they would be marketed as having extra space on top and bottom.

Well, then from what models and resolutions they come from? Your left pic is 400x240, 1.6(6) aspect, so it's not 1920x1200 or 1680x1050 or 1440x900.


Agreed, aspect ratio only matter in a sense of bigger resolution/screen. But I don't see any high-res 4:3 24"+ monitors on sale. Can you find me one? Is there a such a thing as 30" screen that needs tape to become widescreen?

When I moved from 19" LCD to 24" widescreen LCD, I increased fov by 12. That resulted in the same vertical viewing angle (objects became a bit bigger since panel height is up to 12.7" from 11.8" and +10% more horizontal angle. Makes sense to me.
Increasing fov on 19-incher "as I wish" would compress central part of the screen and hinder aim.


They may market it however they like, but we all know that for QW extra space on top and bottom is not as useful as extra space on sides.

Just because there is no error in the calculations doesn't mean they give relevant information or an accurate picture of the real world phenomenon. At best, you have covered one aspect of the matter.


Of course. You can't!? 154Hz is extremely smooth compared to 60Hz, or even 75Hz.


I would say that getting information of game events quicker than your opponent 75% of the time, is much more important point than the actual amount of delay. Plus the additional smoothness and great feel of high refresh rate and fps.


Unless you go to extremes, I haven't noticed that it would create any disadvantage in the game, you can still see everything in the center very well. I suppose you can push the limit a bit further with widescreen though. Still, there is no way I'd switch the smoothness of CRT for that.


I don't have an LCD, so I can't give you a score of any relevance.

-edit-

Well here's the best I could do at this hour just for fun.

http://img254.imageshack.us/my.php?image=sheepnm1.png

Since this thread was started as discussion of LCD response times, this aspect seems important enough for people.


Looks like I get used to LCD so much that everything appears smooth to me.
cl_independentphysics 1 and cl_maxfps 900 helps too.


Disagree, lagging 10ms behind and 50 ms behind is not exactly the same, more on that later.


Agreed on additional smoothness.
Haven't we ruled out fps, since it's 1000+ on most modern PCs?


Objects will became smaller. It's a fact. You think it would not be disadvantage in the game, same way I don't think additional smoothness will be an advantage.
But again, we have this discussion not to convert anyone. You would not play on LCD, I would not come back to CRT. It's fine.


At least now you can see of what kind of numbers we talking about when considering human response to a screen change. Human is a slow animal.

Yes, your eyes will get information of game events 16ms quicker than your opponent 75% of the time. After that your eye-brain-finger system will take 140-200ms to press that mouse button. And with age we getting worse.
What that means is personal reaction time of the player will ultimately overweight any advantages of CRT/LCD.