Steam Deck Input Latency Test

[RoyalMetalKnights]

I saw an old reddit post on input latency using the frame rate limiter and screen refresh slider. DF didn't test it from what I watched, so I decided to test it using "Is it Snappy" app, iPhone with 240fps slo-mo, and a tiny wireless keyboard with a led that lights up when a button is pressed. I used Counter-Strike with a pistol and timed from when the led first lights up to the first visible frame of muzzle flash.

Results:
60Hz/Uncapped = 45.8ms
60Hz/60fps cap = 62.1ms
60Hz/30fps cap = 120.4ms
60Hz/In-game vsync = 66ms
-
48Hz/Uncapped = 50ms, 45.8ms
48Hz/48fps cap = 66.7ms
48Hz/In-game vsync = 79.2ms
-
40Hz/Uncapped = 49.6ms
40Hz/40fps cap = 87.5ms
40Hz/In-game vsync = 104.2ms, 112.5ms
-
MangoHUD
40Hz/40fps cap = 58.3ms, 62.5ms
48Hz/48fps cap = 41.7ms, 41.3ms
60Hz/60fps = 33.3ms, 37.5ms, 37.5ms

[ReBoot]

Have you accounted for potential additional latency from that wireless keyboard?

[RoyalMetalKnights]

ReBoot eredeti hozzászólása:

Have you accounted for potential additional latency from that wireless keyboard?

I considered it, but when comparing to other Reddit post that relied on judging button presses by eye, with higher latency numbers across the board, I thought it reasonable compromise.

The alternative would be to use a wired controller or built-in controls with a led wired to the buttons. Judging the press timing by eye seemed inaccurate.

This is a worst-case scenario, so to say, but the difference between modes should be similar.

*edit* This is with a RF Wireless keyboard, not bluetooth.

[ReBoot]

Bluetooth is RF. But let's roll with "proprietary dongle".
Regarding this as a comparison of worst cases, not benchmarks with real-world numbers though, I retract my complaint.

[RoyalMetalKnights]

ReBoot eredeti hozzászólása:

Bluetooth is RF. But let's roll with "proprietary dongle".
Regarding this as a comparison of worst cases, not benchmarks with real-world numbers though, I retract my complaint.

Feedback is appreciated.

[RoyalMetalKnights]

Using Ratchet and Clank Rift Apart PS5 results from Digital Foundry:
Fidelity Mode: Target->30fps/33.3ms | Display Mode->60Hz | Input Lag->117.5ms
Fidelity Mode: Target->40fps/25ms | Display Mode->120Hz | Input Lag->80.8ms
Performance RT Mode: Target->60fps/16.7ms | Display Mode->60Hz | Input Lag->75ms
Performance RT Mode: Target->60fps/16.7ms | Display Mode->120Hz | Input Lag->60.8ms

[okcomputing]

Googling around, I stumbled onto this. Thanks for testing. We can argue methodology, but there's a clear and consistent pattern that the limiter adds _significant_ latency. You don't even need an elaborate setup, because the extra latency large enough you can feel the difference when playing a game and switching the limiter on/off.

It's been a good while now. We've gotten lots of great updates. But when is Valve gonna do something about this? The 30 fps cap is especially unpleasant to use.

As times goes on and new titles become more demanding for the Deck, the 30 fps cap will become increasingly important. It's already key for games like Elden Ring, and they'd feel much better to play with decent response times.

Edit: more feedback, suggestions

https://old.reddit.com/r/SteamDeck/comments/vwckdi/input_latency_on_the_steam_deck_what_its_like_now/