Quote:
Originally Posted by
r0ach 
Since you can hook a mouse up to an android system, and android gaming is becoming larger with things like the Nvidia console, the topic seems relevant here. Digitalversus has long done tests on LCD screens for input lag, and they also started testing tablet touch response lag. The fastest Android tablet they've tested is 48ms input lag touch response for the Samsung AMOLED, and the Apple Ipad air 2 comes in at a whopping 9ms.
When I said that new Apple tablets felt insanely fast for touch response the other day, someone claiming to be an engineer said to me it was "impossible" for the Apple touch response to be less than 32ms or something like that. It looks like they were wrong.
http://www.digitalversus.com/tablet/ipad-air-2-p22017/test.html
If they could test mice, I feel it would work out in a similar way since I've never seen anything with a Linux kernel attached to it that felt as responsive as Windows. At least Apple engineers are doing something with all the money they're getting paid.
What testing has been done on that "review"?
There's nothing suggesting they did actually test the device hands-on, there's no results of any testing, there's no methodology of any testing, there's no graphical evidence of them testing it.
I only see them gluing the technical specs they were handed by Apple, doing free PR.
So, point one, there's no testing done, just a number pulled from thin air as far as that publication goes - there's no indicative of the process they've followed to test it, there's no statistical analysis shown, there's no proof of anything being done hands-on on the tablet, so it's not hard data, nor it is a performance test.
Point two, those devices are capped at 60fps. That gives a target refresh time of 16.6ms, which is almost double of their 9ms speculation - as we already agreed they didn't specify if they even tested it.
Point three, the screen is IPS, which by itself should already increase the input lag more than specialized TN screens. A rather common amount of latency for those screens is about 5ms.
Point four, such stuff - just like the majority of your threads - does not belong to the mouse subforum.
Quote:
Originally Posted by
qsxcv
uh... not possible with vsync from the (at least) double buffered interface.
as far how far the display lags behind the finger when swiping, every ios device i've used (my own, family's, apple store's) has always felt the same: 50ms/3frames latency.
android... lol. it's called lagdroid for a reason
Even if the device is programmed to use triple buffering, remember that one of those framebuffers is busy serving the screen, thus at much it would be 2 frames of latency or ~ 33ms - unless we count for extra processing or a lower fps rate for battery economy reasons.
Won't comment on the Android matter more than saying it's an OS that does what it was designed to do, same as iOS and WP, with all their selling points and design compromises.
Edit:
Quote:
Originally Posted by
r0ach
Because I wanted to make it a short post and not 2 pages long. You either have to accept my statement as fact that the fastest Android device tested was 48ms, or look it up yourself, which is very easy to do and takes all of 60 seconds.
Because Android is competing directly with things such as SteamOS now and you can use a mouse with Android. I see you quoted 5 centuries old "CS" as some kind of joke like it's impossible to be on Android or IOS. I guess you didn't notice Crysis 3 is now on Android:
http://shield.nvidia.com/games/android
Even Nintendo is now developing for Android supposedly:
http://www.theverge.com/2015/3/17/8230477/nintendo-dena-mobile-games-announcement
Desktop Linux failed at gaming and Android leap frogged it.
I would like to see a scientific paper where they do come with a method to measure this delay.
You do realise that Android is an OS built around the concept of expansive compatibility, don't you? It has to make extensive compromises in order to run in a variety of hardware, to the point that it requires a virtual machine in order to run apps. All those compromises do take a toll on performance, yet Android does not run that slow compared to OSs built around a closed hardware ecosystem, which technically should just run over it performance-wise. But they don't.
You do realise that having a Linux Kernel does not mean it's comparable on performance to most desktop Linux distros, don't you? Not to say that most apps on Android run on top of a layer of Java code, which is interpreted (Dalvik) or hybrid interpreted-compiled (ART), instead of how many applications work on desktop space (compiled if the language supports it).
There's still lots of juice to be extracted from the Linux ecosystem, even though it needs to mature for gaming - it was never designed for gaming in mind.
