[Scott_Arm]

The way network lag affects gameplay changes quite a bit depending on the actual game. For instance, when I play NHL, there is a big delay between an input and when my skater actually moves on the screen. As described above, latency might be hidden very well in an online shooter or other game. Still, input lag is not a big issue on newer tvs. I have a friend with a DLP where the input lag is noticable. I'm just saying that for the most part, network lag will have a far greater impact on play than input lag, which for the most part will not be visible to the user.

[sebbbi]

Multithreaded game engines also usually have more input to screen lag than single threaded game engines. To get max power out of the current console platforms, multithreading is required.

Usually you are processing game logic and physics on one (or several threads) at the same time as you are rendering (separate render & render setup & culling thread). The rendering must be done using last frame game state, as the current frame game state is not ready yet (physics and game logic threads are processing it on different threads).

Single threaded game game loop:
1. Process input
2. Process game logic & physics
3. Render frame

Multithreaded game loop:
1. Start rendering graphics by using last frame finished game state in it's own thread(s)
2. Process input
3. Process game logic & physics (usually in multiple threads)
(4. Graphics thread is finished around here and updates the visible frame independently on game logic & physics)
5. Synchnonize threads

With the multithreaded setup you get slightly longer input lag. The maximum input lag equals the time it takes the graphics chip to render the frame.

[PeterT]

Originally Posted by homerdog

It's not just that TV, most fixed pixel HDTVs I've ever gamed on have had _noticeable_ lag. And I'm not talking about Vizios; Samsungs, LGs, Sharps, all lagtastic to the point of hilarity.

I agree. And in many cases the situation hasn't been getting better, but worse (with more sophisticated image scaling/"improvement" techniques). The only hope are dedicated "game" modes.

I find it amazing that (well-reviewed!) monitors are being sold with lag approaching 80ms.

[Entropy]

Originally Posted by Scott_Arm

In a heavy twitch game it's definitely possible that even a small imperceptible difference could be a deciding factor in play, but for the most part people play online, and the difference in network lag would dwarf any display lag. If you're into lan play, then I could see it being an issue. For single player, if you can't notice it, then it really is irrelevant.

Back when I was a div1 Quake 3 player, having my network lag go from 50 ms to 70 ms was sufficiently debilitating that I resigned from the team. Essentially, it dropped me from decent div1 to div2 quality.
On the other hand, playing WoW, I could go away and fetch coffee in the middle of a fight without affecting the outcome.

Moral: The demands of the game decide how critical lag is for gameplay.

Reasonable extention: Games that are designed so as to be more tolerant of lag, also leave less room for skill building based on precise control, which is a critical component of all physical games.

[homerdog]

Originally Posted by PeterT

I agree. And in many cases the situation hasn't been getting better, but worse (with more sophisticated image scaling/"improvement" techniques). The only hope are dedicated "game" modes.

I find it amazing that (well-reviewed!) monitors are being sold with lag approaching 80ms.

Yup. My friend has an old (~5yrs) Westinghouse LCD, doesn't even have DVI in, but it doesn't have any discernible lag. I guess back then they didn't have all that postprocessing going on. So despite the panel's atrocious response times (ghosting is insane!) it's the only LCD I can play Halo on without getting smashed.

From my experience the newer the TV, the worse the lag, but maybe that's starting to change.

[paawl]

Originally Posted by homerdog

From my experience the newer the TV, the worse the lag, but maybe that's starting to change.

That's the opposite of my experience. My parents ~3 year old Panasonic Plasma has ~160 ms input lag (no game mode); my ~2 year old Vizio LCD has ~80 ms lag (also no game mode); my brother's ~1.5 year old Samsung DLP has ~20 ms lag (in game mode); and my brand new Sharp LCD has ~0 ms lag (in game mode).

What's frustrating to me is that it's very nearly impossible to find out what a display's input lag is, unless you buy it and test it yourself. I got unlucky on the Vizio (but that was never my primary gaming TV, while I still had my Sony XBR CRT), and I looked forever until I found the 52" Aquos that had the least lag.

[homerdog]

Originally Posted by paawl

What's frustrating to me is that it's very nearly impossible to find out what a display's input lag is, unless you buy it and test it yourself.

Yep that's the main problem. Seems like some manufacturer would start advertising this, but maybe people just don't know (or don't care) . Was hoping rhythm games like Rock Band would bring this issue into the mainstream but that doesn't seem to be happening.

[Mr Deap]

Originally Posted by paawl

That's the opposite of my experience. My parents ~3 year old Panasonic Plasma has ~160 ms input lag (no game mode); my ~2 year old Vizio LCD has ~80 ms lag (also no game mode); my brother's ~1.5 year old Samsung DLP has ~20 ms lag (in game mode); and my brand new Sharp LCD has ~0 ms lag (in game mode).

What's frustrating to me is that it's very nearly impossible to find out what a display's input lag is, unless you buy it and test it yourself. I got unlucky on the Vizio (but that was never my primary gaming TV, while I still had my Sony XBR CRT), and I looked forever until I found the 52" Aquos that had the least lag.

The Sharp actually have lag, the video on youtube is actually misleading(test done in 480i, which is actually 45ms which is mesured by an AVS forum member). The only way to get that 0ms is to use the VGA input which most HDTV get no lag at all, but the picture look horrible.

On HDMI with the Vyper engine, the D64U get 15ms of lag in 1080P via HDMI.

Originally Posted by sebbbi

Multithreaded game engines also usually have more input to screen lag than single threaded game engines. To get max power out of the current console platforms, multithreading is required.

Usually you are processing game logic and physics on one (or several threads) at the same time as you are rendering (separate render & render setup & culling thread). The rendering must be done using last frame game state, as the current frame game state is not ready yet (physics and game logic threads are processing it on different threads).

Single threaded game game loop:
1. Process input
2. Process game logic & physics
3. Render frame

Multithreaded game loop:
1. Start rendering graphics by using last frame finished game state in it's own thread(s)
2. Process input
3. Process game logic & physics (usually in multiple threads)
(4. Graphics thread is finished around here and updates the visible frame independently on game logic & physics)
5. Synchnonize threads

With the multithreaded setup you get slightly longer input lag. The maximum input lag equals the time it takes the graphics chip to render the frame.

In big a games offline can actually add an extra 15 to 30ms of lag due to the process games logic.

Also in big a GPU with weak fillrate can add extra lag if the resolution & polygons are too high?

So the more it is multithreaded, the weaker the GPU, the more chance you will see lag.

[Quaz51]

Originally Posted by zed

halo3 has been tested (see gamasutra article)
input lag is 133msec -> 166msec

on a CRT but on a flat panel (like in the killzone 2 video test) is 166ms/200ms for Halo3 and others (200ms for GTA4, skate. 200/266ms for Harry potter)

http://www.gamasutra.com/view/featur...eo_.php?page=3

[Cheezdoodles]

Originally Posted by homerdog

Network lag affects input lag? I didn't think it worked like that.

Depends on netcode. Some games will not show you firing a gun and the bullet hitting before the server "confirms" the shot.

Others (most online fps games) will show your gun firing, but bullets hitting once server has confirmed the shots.

KZ2 will show your gun firing and bullets hitting instantly, regardless of your lag to other players\servers (i say players\servers because my experience suggest that KZ2 is actually pure P2P netcode - which means that you have different lag to every player).

[homerdog]

Originally Posted by Ostepop

Depends on netcode. Some games will not show you firing a gun and the bullet hitting before the server "confirms" the shot.

Okay I've never played a game like that, that I know of.

Originally Posted by Ostepop

KZ2 will show your gun firing and bullets hitting instantly, regardless of your lag to other players\servers (i say players\servers because my experience suggest that KZ2 is actually pure P2P netcode - which means that you have different lag to every player).

From my experience I think that's what Halo 3 does. Not P2P but with guns firing and bullets hitting regardless of network lag.

But as far as input lag when you're looking and walking around and such, network lag should not impact that at all in any FPS games, correct?

[Shifty Geezer]

Originally Posted by Quaz51

on a CRT but on a flat panel (like in the killzone 2 video test) is 166ms/200ms for Halo3 and others

200 ms is crazy lag! Like playing through treacle. And I'm shocked that 50 ms is considered good for a music game. Actually recording music on computer, 10 ms is bad! 50 ms lag between playing a note and hearing the result is unplayable, yet that's the standard of music games?! Having said that, I've played a taster of GH and it felt pretty good. But then I also remember feeling the timing was out.

[Entropy]

Originally Posted by Shifty Geezer

200 ms is crazy lag! Like playing through treacle. And I'm shocked that 50 ms is considered good for a music game. Actually recording music on computer, 10 ms is bad! 50 ms lag between playing a note and hearing the result is unplayable, yet that's the standard of music games?! Having said that, I've played a taster of GH and it felt pretty good. But then I also remember feeling the timing was out.

See, you do appreciate precision control!
I'll reinforce your message though, anyone who has been involved in "live" music would.

[Cheezdoodles]

Originally Posted by homerdog

But as far as input lag when you're looking and walking around and such, network lag should not impact that at all in any FPS games, correct?

correct (but what you see- as in where other players are and the action thats happening will ofcourse be affected by network lag)

[Turok]

Input lag is one of my biggest HDTV pet peeves, something that is only exacerbated by my OCPD. Thankfully my Sony Bravia has a negligible amount of lag, and that slight amount of lag is gone once Game Mode is turned on. Despite this, Killzone 2 still has really bad aiming.

[homerdog]

I still use a CRT for my Xbox gaming, but it's my roommate's. Since no one seems to be able to recommend a lagless TV that isn't Bravia expensive I'll probably get a PC monitor for it when I move.

[Mr Deap]

I made a video of both wireless controllers showing a difference of one frame of delay. It could be the game too...


Soul Calibur IV
720P Component
Same setting on HDTV(all post processing off)


I made sure that both system ran in the same resolution, display & setting.

[grandmaster]

Yes, you'd need to run several tests over different games. What camera did you use? Do you have any way of ascertaining the lag in your display?

While we're at it, can any one recommend a cheapo 720p60 camera? I'm thinking of something in the price range of the Kodak Intro Zi6 (around £100). I'd very much like to add response measurement to the tests we do at Digital Foundry.

[SG79]

Originally Posted by Mr Deap

I made sure that both system ran in the same resolution, display & setting.

Something is a bit off here, since I would've heard countless complaints about this now, particularly that the PS3 version is primarily used in tourneys (more players owning sticks on the console).

Surprising nonetheless.

[Andrew Lauritzen]

Originally Posted by Shifty Geezer

200 ms is crazy lag! Like playing through treacle. And I'm shocked that 50 ms is considered good for a music game. Actually recording music on computer, 10 ms is bad! 50 ms lag between playing a note and hearing the result is unplayable, yet that's the standard of music games?! Having said that, I've played a taster of GH and it felt pretty good. But then I also remember feeling the timing was out.

50ms is completely unacceptable in music games, but the key is that it indeed can be calibrated out. If you're playing by ear (i.e. timing your hits to the sound of the music, not to the bar crossing the thingy on the screen and more-so the visual feedback of it exploding when you hit something), which you should be, then it works perfectly fine.

Of course sections where there is direct feedback (i.e. drum solos) from what you're playing are unplayable, but for that reason they rarely affect gameplay.

Note that in most games though, the concept of "calibration" is not an option as the feedback loop is bidirectional. The reason it works in music games is because it's effectively a unidirectional feedback loop, and if your UI is lagged a bit (with respect to whether you "got it right"), it's not a huge deal. It would be a huge deal in an FPS or similar though.

[paawl]

Originally Posted by SG79

Something is a bit off here, since I would've heard countless complaints about this now, particularly that the PS3 version is primarily used in tourneys (more players owning sticks on the console).

Surprising nonetheless.

If these measurements are valid, and we assume that the PS3 is justly preferred with sticks for tournament play, then we can probably conclude that the extra frame of lag in the PS3 version is due to the PS3 wireless controller using Bluetooth, while MS uses a proprietary wireless communications device. The PS3 sticks wouldn't be affected if they're connected via USB (like my VSHG). I recall that MS claimed that one reason that they went with their own wireless technology instead of Bluetooth was because Bluetooth had unacceptable lag for a gaming controller. Sony retaliated by saying that they had developed a low-latency Bluetooth implementation for the PS3 and its controllers. I was always curious to see if there was any practical difference between the two systems' wireless controller technologies. This is the first such test I've seen. Keep up the good work!

[SG79]

You could be right. Come to think of it, even pad players have the controller connected to the system via USB, so there wouldn't be a detectable difference. The controllers do sync A LOT faster when connected to the system.

Fighters and music games are about the only two genres that get affected with this. There is also the HDMI induced lag with PS/PS2 games on the PS3 that's noticeable again for fighters and music games, but that's another matter.

[Mr Deap]

Originally Posted by SG79

You could be right. Come to think of it, even pad players have the controller connected to the system via USB, so there wouldn't be a detectable difference. The controllers do sync A LOT faster when connected to the system.

Fighters and music games are about the only two genres that get affected with this. There is also the HDMI induced lag with PS/PS2 games on the PS3 that's noticeable again for fighters and music games, but that's another matter.

There are indeed other value, which I didn't demonstrate. The wireless XBOX 360 controller isn't lag free. I presume that this one deliver around 16 or 33ms of lag. I played so much with my wired controller that the joystick was trash & was good for garbage(3 year of hardcore usage).

Also HDTV panel deliver lag too which isn't helpful either shown in the thread.

The lowest you can get on the market is 16ms of lag via HDMI & popular model such as the Sony & Samsung lag of 33ms via HDMI in native resolution & game mode.

33ms indeed is hardly detectable, but most people actually experience 66ms of lag if they play on game mode which not everyone use.

If anyone up to the live experience at home & try an OLD CRT SDTV with the controller plug on the USB, expect quite some tight control.

[Iron Tiger]

I think ever since I conditioned myself against it in the Xbox version of Burnout, I haven't been able to notice input lag unless it's been extreme. The Killzone demo was one of those extreme instances, and last weekend I played the demo of that new pinball game on PSN, and also felt it to be laggy. I play on a Samsung LN46A650, and generally prefer to play with prettier IQ than in the "lag-free" Game mode, but in those rare instances, I have little choice but to give up the pretty.

[SG79]

Originally Posted by Mr Deap

The lowest you can get on the market is 16ms of lag via HDMI & popular model such as the Sony & Samsung lag of 33ms via HDMI in native resolution & game mode.

33ms indeed is hardly detectable, but most people actually experience 66ms of lag if they play on game mode which not everyone use.

Tests of some of the lower end BRAVIA models (W series) ones pegged it at 17-33ms. All manufacturers typically just list 33ms, and I must note that Samsung sets have improved immensely from what they had 4-5 years ago

Surely you meant 66ms if game mode isn't use, right? I use that mode and turn off any other post processing on my nearly 2 year old XBR4. Even TV or movie sources.

But you brought up an interesting point, because I was shocked to learn that there are so many people out there who actually use the 120Hz mode and don't even detect the lag. Not only was the lag awful when I tried it on my set, the motion artifacts were unacceptable.

I guess this will be a thing of the past one day.