(Look at the description first!) Latency Testing: Odin 2 Mini vs Odin 2 Pro (and 4 other handhelds)

[@tommythuyen4063]

This is a true test. I own GPD Winmax 2 and AYN ODIN 2. When playing "Prince Of Persia: The Lost Crown" on Yuzu, the delay between the button press and the movement of the character Sagon on ODIN 2 is very clear. Especially when needing to stop in a dangerous area, the character always tends to move later. Meanwhile, this situation does not happen on my GPD Winmax 2. When you need to stop, the character will immediately stop.

[@tonyrosario1940]

Love input lag tests! Thanks for looking into this! 🍻🤘🏼🔥

[@Shigarui]

Love the empirical data. Thanks

[@MisterDrumsNoob]

That's interesting. I was expecting at least 2 frames of difference (~33ms) at 60hz between the odin2 and the rest... This is the result I found by the way when I compared my odin2 and my retroid pocket 2s running Super Mario World in retroarch without runahead or preemptive. Odin2 had an average response time of 130ms, while the retroid pocket 2s had an average response time of 100ms Thank you for this test. Subscribed! ❤

[@chemicalburrito782]

I used Gamepad Tester mainly because I was curious to see what the response time was independent of any lag introduced by the various emulators - and also "green light goes on" is pretty easy to see on my camera along with the button press. I would say that there might be something else going on between retroarch and the Odin 2 hardware (drivers?), but the Odin 2 Mini is more responsive than the original Odin 2 and they're nominally the same hardware with a different screen for the Mini. (Like some of the people on Reddit, I "felt" that the Mini was more responsive than the original 2 ... but wanted to see some actual data.)

[@dgevert]

@@chemicalburrito782 I've seen people speculating that the LCD panel on the Odin 2 has higher latency, which is a plausible theory when a different display tech on the Odin 2 Mini seems to be outperforming it. Now we just need some kind of comprehensive testing for input lag in the different emulators. I know Yuzu just does not feel great on my Odin 2. It might be game-dependent. I can't play Spelunky 2 through Yuzu on my Odin 2 at all, it feels terrible. I can literally play the same game through remote play on other devices and it feels better than it does emulating locally on my Odin 2.

[@MisterDrumsNoob]

@@chemicalburrito782 hey, you know what a interesting test would be? Stock gpu drivers vs turnip drivers. I don't know if this is placebo or not, but Hollow Knight felt terrible on odin2 yuzu using qualcomm drivers, but after I switched to turnip, it became more tolerable. Anyway, thanks for all the testing you've been doing. I would definetely get a odin2 mini if I could, instead of my regular odin2, because latency is a big deal to me.

[@chemicalburrito782]

That sounds about in line with my updated results - on the gamepad tester I got 20-ish ms slower response from the original Odin 2 than the other Android devices (except the Edge which was the speed demon in terms of response time - especially if you kick in the 144Hz screen refresh). I do think the margin of error here is going to be somewhat high just because of my testing setup being very low-budget.

[@chemicalburrito782]

@@MisterDrumsNoob I think the general consensus is "use the turnip drivers unless they just don't work". You think the difference in feel you're seeing between the two drivers is a frame pacing issue, maybe?

[@DSolo59]

Thank you for these tests! I would recommend using a external controller via bluetooth and usb as well. Also, making use of the Odins performance mode would be very interesting to see during these tests. Again, thank you for all your work here!

[@chemicalburrito782]

@@DSolo59 High performance mode was on for the Odin twins. I usually use that when testing new PS2 games so I just leave it that way.

[@TheFirehands150]

For those of you who dont know: Smash Melee on the Gamecube is right around the Odin 2 input lag level, with Melee averaging 58.83 ms. Smash Ultimate averages input lag of 98.31 ms. Do with the information what you will.

[@golivaguitar]

Love this! Tks so much!

[@rob8311]

How do you account for travel distance of the buttons

[@chemicalburrito782]

To try to keep things consistent among the devices, I am pushing the button as fast as I can and starting the frame count when the button has been completely pressed. This is, of course, not perfect ... but I think it is more reproducible than starting the count at the beginning of the press because - as you say - the buttons have different travel distances.

[@TheFalco708]

was about to buy the odin 2 mini but after seeing this I'm worried to a 12 years old ps vita would be better at this point

[@impheris]

OMG 50ms -.-"