Mobile device Screen latency

[MrX8503]

An interesting article on Engadget.
http://www.engadget.com/2012/03/10/microsoft-cuts-touchscreen-lag-to-1ms/

It reminds me of another article on touch screen quality.
http://www.geek.com/articles/mobile...een-accuracy-test-iphone-still-wins-20100325/

I have pretty strict standards when it comes to speed/responsiveness on a touch mobile device. If it lags even the slightest, most likely I won't purchase it.

With all the debate about Android vs iOS on speed/responsiveness, a lot of the focus has been on software...when the hardware plays a small role too. You can do the test yourself by opening up a line drawing app.

I just tried it on my iPhone4 and its about anywhere from 10-1ms latency. This is basically how fast something can track your finger. I never owned any previous iPhone, but I remember my first Android device (EVO 4G) having something that felt like 100ms latency. I'm not sure if that still holds true with current Androids, but give it a whirl with a drawing app and post your findings. 100ms/10ms/1ms.

 

[Dominato3r]

That video is pretty nice. I never knew these things had that kind of touch latency. I wonder, though, if that MS demo is an actual touch screen... A few of the comments on that video say it's a projector/camera setup (Which is a common tech used by MS).

 

[ponyo]

It's because older HTC phones like the Nexus One and Evo had only 2 point multitouch because HTC used cheap crap controllers. Everyone complained about Nexus One faulty multitouch. The Android phones tested on that old test (HTC Droid Eris, Motorola Droid, Google Nexus One) all only had two point multitouch. I think the iPhone had 4. It wasn't til Samsung Galaxy S phones that Android had just as good or better multitouch than the iPhone. Galaxy S was the first with 5 and new Samsung phones have 10 point multitouch.

 

[MrX8503]

That video is pretty nice. I never knew these things had that kind of touch latency. I wonder, though, if that MS demo is an actual touch screen... A few of the comments on that video say it's a projector/camera setup (Which is a common tech used by MS).

What do you mean actual touch screen? Its a touch screen, but a prototype that they do testing on.

It's because older HTC phones like the Nexus One and Evo had only 2 point multitouch because HTC used cheap crap controllers. Everyone complained about Nexus One faulty multitouch. The Android phones tested on that old test (HTC Droid Eris, Motorola Droid, Google Nexus One) all only had two point multitouch. I think the iPhone had 4. It wasn't til Samsung Galaxy S phones that Android had just as good or better multitouch than the iPhone. Galaxy S was the first with 5 and new Samsung phones have 10 point multitouch.

Thanks for the info. Didn't know that, I just know that the EVO touch screen was terrible.

 

[InfinitySquared]

An interesting article on Engadget.
http://www.engadget.com/2012/03/10/microsoft-cuts-touchscreen-lag-to-1ms/

I have pretty strict standards when it comes to speed/responsiveness on a touch mobile device. If it lags even the slightest, most likely I won't purchase it.

I just tried it on my iPhone4 and its about anywhere from 10-1ms latency. This is basically how fast something can track your finger. I never owned any previous iPhone, but I remember my first Android device (EVO 4G) having something that felt like 100ms latency. I'm not sure if that still holds true with current Androids, but give it a whirl with a drawing app and post your findings. 100ms/10ms/1ms.

How did you test and time the response? I'm interested as the article discusses new tech that will eventually allow 1ms response, but you already have it, apparently.

 

[MrX8503]

I didn't time it, but I tested it by moving icons around or playing air hockey. Anything that'll track your finger would work. From their example, my iPhone was anywhere from 10ms or less.

 

[InfinitySquared]

I didn't time it, but I tested it by moving icons around or playing air hockey. Anything that'll track your finger would work. From their example, my iPhone was anywhere from 10ms or less.

If you didn't time it, how do you know it was less than 10ms?

 

[MrX8503]

How would I time 10ms? I did a visual test and the latency looked similar to what they showed. It definitely wasn't 100ms.

 

[InfinitySquared]

How would I time 10ms? I did a visual test and the latency looked similar to what they showed. It definitely wasn't 100ms.

So it could have been 99ms, 2ms or 67ms, or in fact any level of ms in between and you wouldn't know. My point is you're basically using the human mind, something that is very poor at assessing the passing of time, then passing the findings through a purchase affirmation filter and coming up with a 'fact' that determines that you already have the touch response that you saw in a video of technology that is still in the lab.

It's like a 'butt dyno', but for your hand. There is no science.

 

[MrX8503]

The human mind isn't very good at discerning the passing of time when the change is very small. 100ms and 10ms is not a small change. When the video showed 100ms, 10ms, and 1ms, your mind couldn't tell the difference? Okay....

 

[InfinitySquared]

No need to make this personal, I'm simply suggesting your science is flawed, which it clearly is, so defend that rather than casting aspertions upon my mind.

Unless you can prove you can identify, with accuracy, timings below one tenth of a second which you claim, then you have no argument.

 

[MrX8503]

No need to make this personal, I'm simply suggesting your science is flawed, which it clearly is, so defend that rather than casting aspertions upon my mind.

Unless you can prove you can identify, with accuracy, timings below one tenth of a second which you claim, then you have no argument.

Oh so now its personal when I asked you a question directly if you could discern between 100ms,10ms, or 1ms. I'm guessing the answer is yes. I wasn't the one doing the finger pointing, but when it comes to you answering a question, its now "personal".

There is no argument. This thread is just to post your findings.

I never said that my observation was science and all I was asking from others were observations. If you didn't notice I gave an estimate, its not like I said my phone did 5ms exactly. I'm asking for observations in the realm of 100ms, 10ms, or 1ms, people can time their observations if they want.

 

[InfinitySquared]

Oh so now its personal when I asked you a question directly if you could discern between 100ms,10ms, or 1ms. I'm guessing the answer is yes. I wasn't the one doing the finger pointing, but when it comes to you answering a question, its now "personal".

There is no argument. This thread is just to post your findings.

I never said that my observation was science and all I was asking from others were observations. If you didn't notice I gave an estimate, its not like I said my phone did 5ms exactly. I'm asking for observations in the realm of 100ms, 10ms, or 1ms, people can time their observations if they want.

Well of course it is personal, you asked me directly, with an undertone of suggesting I had a mental deficiency, if I could determine timings below 1 tenth of a second. The laughable part is that you seem to believe you can assess timing down to a 1ms accuracy, which is impossible.

My point is that no one can determine timings below 100ms with any sort of accuracy, not you, not me, not one single person, so your assertions that your device is sub 10ms is bunk and anyone testing their device in the same manner is also 100% incorrect, making the 'test' flawed and pointless. In fact I could say your very own personal smartphone has a response of 120ms, and you have no way of telling me I'm wrong.

Now let's assume the device you own refreshes the screen @60FPS. Each frame is therefor rendered in 16ms. Oops, there's your 10ms or less theory out of the window before we even begin to factor in controller lag, software interpretation and presentation.

 

[MrX8503]

Where did I say I could determine time by 1ms increments? I thought I was pretty clear in asking of observations in the realm of 100ms, 10ms, or 1ms. Using the video, I think most people can make that estimate, if they couldn't....why would MS show the video in the first place?

If I said, my iPhone is 5ms, then you would be right, however I didn't say that.

 

[InfinitySquared]

OK.

 

[Puddle Jumper]

If you had two phones in hand and tested them both I have no doubt most people could say which has less latency but trying to assign that difference a value in ms is pointless.

 

[Bateluer]

Using MrX8503's testing methodology, my Droid 1, Thunderbolt, and Nexus are all sub-10ms.

 

[NikolaeVarius]

Using MrX8503's testing methodology, my Droid 1, Thunderbolt, and Nexus are all sub-10ms.

My Droid x also is under 10ms

 

[Dominato3r]

What do you mean actual touch screen? Its a touch screen, but a prototype that they do testing on.



Thanks for the info. Didn't know that, I just know that the EVO touch screen was terrible.

If you look closely, the "square" that he's moving with is hand, is actually a light/projected image of some sorts. You can see the color of the square on top of his fingernail.