Steam Deck OLED doesn't just look better than LCD, it plays better too

With its remarkable OLED HDR display, dramatically improved battery life and numerous quality of life fixes, the new Steam Deck OLED is an exceptional PC handheld - but there was something else I noticed during testing that I wanted to check out. The system felt snappier and more responsive - and it is. Backers of the DF Supporter Program were also expecting response time improvements from the new screen which further motivated my testing and yes, it turns out that Steam Deck OLED doesn't just last longer and look better, depending on your sensitivity to input lag, it plays better too.

At a basic level you can define input lag as the time taken between user input - like pressing a button - then the resulting action happening on-screen. There's always going to be processing time within the game engine itself, which is the time taken to process the input from the player through the game logic and then through the renderer. This will vary quite dramatic in of itself from game to game. And then there's extra lag from the display itself.

In receiving an impressive upgrade to its display, Steam Deck OLED attacks input lag seemingly on two fronts. First of all, the response time of the panel itself in like-for-like display refresh rates seems to be faster on my OLED Deck up against my launch LCD model. So, comparing the LCD panel at 60Hz with a game running at 60fps up against OLED in the same scenario, the OLED is typically 10ms faster to respond in my testing. But there's more than that - the OLED can top out at 90Hz, opening the door to even faster response. You may not be playing too many games at 90fps on Steam Deck, but the 90Hz panel can play a key role when the Deck's frame-rate limiting options are engaged - and you'll be using those a lot.

I wanted to put some numbers to the experience, so I tested input lag using Nvidia's LDAT system. This takes the form of a sensor that attaches to the display, tuned to trigger when the sensor 'sees' sudden increases in luminance - most commonly, the muzzle flash of a weapon fired in a first person shooter. The sensor connects to a customised mouse, which in turn connects to the Steam Deck via a USB hub. The LDAT sensor itself has a USB connection to a second PC that essentially measures the time taken between mouse button press and the sudden increase in luminance on-screen. LDAT is typically used on much larger screens, but as I discovered, it works just fine on handhelds too.

For my testing I concentrated on two games - Doom Eternal and Crysis 3 Remastered - and I chose two titles for a very specific reason. You see, the input lag of any given game can be very very different, even if they are running at the same frame-rate. For both titles, I tested using their tutorial areas, which are very simple in nature and allow us to hit maximum frame-rates on all tested displays. From there, I can use the frame-rate limiter option within SteamOS to test input lag at various performance levels.

The new Steam Deck update features a single slider that allows you to choose your frame-rate cap, then automatically adjusts the frequency or refresh rate of the display to provide the most consistent fit. Limit your Steam Deck to 30fps and the LCD model still runs its display at 60Hz, refreshing the screen twice with the same image. However, Deck OLED ratchets that up to 90Hz, with the same frame refreshed three times instead. However, some target frame-rates - like 60fps, for example - will see both displays run at 60Hz. There's no other clean divider for the 90Hz screen. Let's start by taking a look at my testing with Doom Eternal, where I acquired my metrics from the tutorial level - which easily runs at 90fps on Deck OLED at 800p on medium settings.

In measuring input latency, each result is an average of 20 different LDAT captures in identical conditions, but there is a certain degree of variance - even in comparing different runs of 20 capture. I'd say it's in the region to 2ms to 3ms. With that in mind, these Doom Eternal results throw up some interesting findings.

First of all, even with both displays running at the same refresh rate, OLED is faster to respond, meaning that you're looking at a circa 10ms improvement to input latency. This may not sound like much, but bearing in mind the relatively low frame-rates we're dealing with here, it's a big deal. However, what is more impactful is the use of frame-rate caps when the OLED model is able to tap into a higher refresh rate, where the gap increases to circa 20ms. This is where the Deck definitely feels a lot snappier in gameplay.

By the way, I included 25fps testing here for a couple of reasons. First of all, it's the lowest frame-rate I would consider capping to on the handheld - and it does have utility in slower-paced games and also increases battery life. The GPU has much less work to do when each frame persists for 50ms up against the 33.3ms of 30fps. Next up, I repeated these tests with Crysis 3 Remastered - a much 'laggier' game, even when running at the same frame-rate.

There are some commonalities with the Doom Eternal results, but the spread definitely differs in some crucial aspects. The gap between the two panels running at the same refresh rate increases in favour of the Steam Deck OLED, and while the same doubling of that advantage is seen with the 45fps result (and the doubling of refresh rate on OLED), it's interesting to note that the 30fps and 25fps results see some kind of late comeback for the LCD screen, which I can't really explain. At 25fps, they're essentially the same, within margin of error.

An interesting question to ask is why the OLED screen refreshes at 50Hz at all when frame-rate limited to 25fps. The screen should be at 75Hz, but curiously, there's a gap between 73Hz to 76Hz on the OLED panel. With no clean divider for 25fps, the Deck OLED defaults to 50Hz instead.

Even so, Deck OLED still offers superior input lag result in this second test - and in actual fact, in my first tests, the divide was actually much wider. Implausibly wide, in fact. I consulted Valve about this who recommended updating the LCD firmware to the preview firmware, which should bring parity in software between the two units. This changed the results to what you saw above, but either something was wrong with my Deck, or else all LCD owners are also going to get a noticeable improvement with the new firmware, as you can see from the tables below.

Valve did share with me that they're looking to improve input lag, and assuming these older firmware results aren't a problem of my LCD Steam Deck, it seems that owners of the existing machines should be seeing a nice improvement to response, particularly when those all-important frame-rate caps are used.

But let me just stress again that the results in the top two tables are comparing like-for-like system firmware, and even with the improvement seen on LCD, the OLED model is still noticeably faster to respond. And yes, when you're talking about 10ms and especially 20ms reductions in input lag, I am fine in saying that there's a tangible improvement to the gameplay experience, though the extent to which the user may notice may vary depending on their own experiences with input lag - and the improvements may change on a game-by-game basis.

And that's all before we factor in that Steam Deck can now run titles at 90 frames per second - and as you can see from the results in the first two tables, that presents an even bigger improvement to input lag, it's just that the range of games that you'll be able to run at that frame-rate is going go be quite low. That said, Ori and Will of the Wisps managesit, while also delivering one of the most stunning examples of the Deck OLED's display at work - so that would be my go-to game for those buying the new handheld.