Game mode on: CNET tests TVs for input lag

You just got a new TV, and the picture looks great. Then, when you pop in your favorite shooter, you notice your kills going down, your aim a step slow, and your sterling record taking hits like a noob wandering through the kill zone.

Do you all of a sudden suck at COD? Maybe. But perhaps your newfound lack of feel is caused by input lag introduced by your otherwise awesome new TV. Input lag is the delay, in milliseconds, between a TV receiving a signal and the results of that signal appearing on the screen.

Those milliseconds don't matter for TV shows and movies, and they don't even matter for most games -- the majority of gamers probably wouldn't even notice if their TV was laggy. But if you're an attentive, skilled gamer, especially one who plays "twitch" shooters like Call of Duty or Halo, especially in online multiplayer environments, input lag can mean the difference between virtual life and death.

The good news is that unlike lag introduced by a network, or your own brain, you can actually do something right now about input lag. The easiest course of action is to turn on Game Mode, if your TV has it. That mode typically removes as much video processing as possible, and according to my tests, can reduce lag by half or even more.

If that doesn't work, the only solution is to play on another TV.

TV input lag results compared
If you're buying a TV that will be used to play twitch games, here's some additional good news. CNET TV reviews will now include input lag measurements, so you can tell how much lag a TV introduces.

In fact, here's a chart that includes almost all of the 2013 TVs we've reviewed so far (and a few we haven't), ranked in ascending order of input lag. The numbers reflect the lowest lag number the TV is capable of, typically achieved in Game mode. Update: The results for the two Samsung TVs were achieved with the "trick" of renaming the inputs to "PC." See the video processing sections of the reviews for details.

Television	Type	Input Lag (ms)
Sony KDL-55W802A	LED LCD	16.9
Sony KDL-55W900A	LED LCD	19.7
Vizio E50i-A1	LED LCD	32.2
Toshiba 50L2300U	LED LCD	33.4
BenQ W1070	DLP projector	33.7
Panasonic TC-P50S60	plasma	34.1
Panasonic TC-L55DT60	LED LCD	34.4
Seiki SE50UY04	LED LCD	38.3
Panasonic TC-P60ZT60	plasma	46.2
Panasonic TC-P60VT60	plasma	47.9
Mitsubishi HC7900DW	DLP projector	48.5
Samsung UN55F8000	LED LCD	50.9
Samsung PN60F8500	plasma	53.1
Sharp LC-60LE650	LED LCD	58.0
Panasonic TC-P55ST60	plasma	73.6
JVC DLA-X35	LCoS projector	87.1
Epson 3020	LCD projector	89.3
Epson 5020UB	LCD projector	95.1
TCL LE58FHDE3010	LED LCD	129.8

As you can see these results don't necessarily favor one technology over another, although a few patterns begin to emerge. The most prevalent so far is that TVs heavy on video processing and/or Smart TV features seem to lag behind -- Panasonic's ST60 is much laggier than the minimally smart S60, for example. Of course those "loaded" TVs often have other benefits, not the least of which is better picture quality.

I'll have a better idea of the patterns as I review more TVs throughout the year. Going forward, the Geek Box at the end of every CNET TV review will include a row for Input Lag. It will also include contextual cutoffs like other Geek Box numbers: Lag of less the 40 milliseconds will be Good, between 40 and 70 Average, and more than 70 Poor. As usual these cutoffs are for general qualification only--even the twitchiest gamer can't tell the difference between a Good and Average TV if one is 39.9 and the other 40 milliseconds.

Bodnar: The ins and outs of testing input lag
Until recently, measuring input lag has been a very time-consuming process involving esoteric software, a fast camera, and no small amount of expertise. It's also a performance characteristic that's simply not important to most TV shoppers. Sites like Anandtech and the superb HDTVTest in the U.K. have been at it for awhile, but I've have basically ignored it when reviewing TVs.

Then the Leo Bodnar lag tester came along.

The Bodnar is a little red box with a big yellow button on top, a hole for a photo sensor on bottom, and an HDMI output (plus a USB port used for service only) on the back. Thanks to the box, the input lag measurement that used to take an expert like Anandtech's Chris Heinonen 30-45 minutes or more can be conducted in less than a minute.

And it works great. The repeatability and consistency of the results I've found in the last few months of using the device, as well as subjective corroboration by my twitch-happy colleague Jeff Bakalar and a conversation with Leo Bodnar himself, all lead me to the conclusion that the results are legitimate. I also worked with Heinonen and HDTVTest's David Mackenzie to compare results from different displays and trade notes and thoughts (thanks guys!).

To keep things simple, CNET will only report the results of Game mode -- if you're playing millisecond-dependent games, you should be using Game mode. If the TV isn't equipped with a Game mode, we report the lowest lag the TV is capable of. Check out "How we test TVs" for more on our exact methodology.

Leo Bodnar on Bodnar
The little red box is the brainchild of Leo Bodnar, whose eponymous U.K.-based Web site, a small partnership with two other people, is mainly dedicated to selling devices used in racing simulation, controllers for USB, and other kinds of mechanical interfaces for PCs. The partnership develops these devices themselves.

Intrigued by the unique lag tester, I arranged an interview with Leo. "We developed it primarily for our own use, since we were interested in finding out which displays showed the least lag," he told me. Aside from accuracy, the priority was simplicity and the ability to test many displays quickly -- in contrast to the bulky, lab-based camera method in use previously.

"I'd never designed anything like that before, so it was mainly a proof of concept," he said. "We even took it to a shop and measured every TV they had. The shopkeeper who helped us was very interested in the results."

In about a year he's sold 120-odd testers. He's planning a follow-up Mark 2 version with logging capabilities and the ability to measure other characteristics of the display, perhaps including light levels and color.

How does lag translate to humans?
A bunch of numbers is one thing, but before I could incorporate them into CNET reviews, I needed subjective verification. The main question was how input lag really affected gameplay.

I'm not a twitch gamer, so I enlisted the help of Jeff Bakalar, longtime gaming editor at CNET -- and as experienced and exacting a gamer as I know.

I set up five TVs for him to play on: the Sony KDL-55W802A (with a lag of 16.87ms), the Panasonic TC-P50S60 (34.1), the Panasonic TC-P55ST60 (75.73), the Samsung UN55F8000 (81.43) and the Samsung PN60F8500 (107.5). (Update: The two Samsungs were in Game mode for the test, which isn't their best for lag. The chart above shows updated lag results for each using the "PC trick" as detailed in the reviews). An Xbox 360 spinning Call of Duty: Black Ops 2 was hooked directly to each TV individually, and I asked Jeff to give me his impressions of how laggy they were.

The key is that Jeff had no idea what the input lag numbers were for any of the TVs. I did, however, and I arranged the order in which the TVs were presented -- although he often requested to swap back and forth between two or more to confirm his impressions.

Jeff quickly identified the Sony as his "reference," and when I asked him which one he liked next most, he called out the Panasonic S60. He ranked the other three as worse, and definitely preferred gaming on either the S60 or the Sony. The Samsung PNF8500, the member of the group with the highest measured lag, was his least favorite.

Being a human, Jeff's impressions tended toward shades rather than outright pronouncements. Bouncing between the Sony and the S60 he said it was "very difficult to tell" between them but he did eventually pick the Sony. When I swapped the S60 for the ST60 he immediately noticed a difference, saying "this one feels better." Most of the differences were more difficult to discern, however, and required a few minutes playing to rank in his mind.

Some of his impressions of gaming on the laggy TVs reminded him of playing early builds of games that are missing final refinements. He tried upping the sensitivity of the controller in the game's menu, but it didn't help. He noticed the lag most prominently in fine aiming/fighting sessions. He also said that in many games, like Skyrim or Madden, he might not even notice it, or if he did, it was something he could quickly get used to and compensate for. Nevertheless, he stressed, these conditions would not be optimal for gaming.

Lag and buying advice
Depending on your own sensitivity as a gamer, input lag is potentially a problem with any game that depends on the reaction time between the controller and the TV -- and that's most of them. The principal factors are how you play, how much the game rewards split-second timing, and how much lag is actually introduced.

That said, input lag is primarily an issue for twitch gaming, where every millisecond counts. For that reason, I will not be incorporating its measurement into my consideration of the TV's rating unless noted otherwise.

So if you're an avid twitch gamer, should you buy a TV like the Panasonic TC-PST60, a model I gave 9/10 for picture quality, but which scored a Poor 75.73 on my input lag test? No, you shouldn't. Instead, the S60 would be an excellent alternative. If you're less avid, or play less twitchy games, input lag should be less of a consideration.

I'll continue to test every TV from this point forward for input lag, and if it matters to you -- or doesn't -- please feel free to leave a comment. In the meantime, don't forget to turn on Game mode.