or Connect
AppleInsider › Forums › Mobile › iPhone › Robotic testing finds touchscreen inaccuracies at edge of iPhone display
New Posts  All Forums:Forum Nav:

Robotic testing finds touchscreen inaccuracies at edge of iPhone display - Page 3

post #81 of 147
I wonder if there could be an issue with Apple's touch technology, since it's based on some biometrics (heat/electrical signal) that robots aren't 100% able to replicate with a rubber end piece (also explains why styluses suck so badly on iPads but Galaxy Note comes with one built in). I doubt this test is even possible with a living human finger to see what the difference would be.

However, good news for Samsung and bad for Apple if this is holds true.

Side Note:
I think this report may be accurate. I developed a game a while ago where you can put your iPhone on the table between two people and play against each other (aka, each using half of the iPhone screen) and the top/upside down player would always be at a disadvantage because of tap accuracy.

I always thought this had to do with something the OS does in relation to where it "thinks" you want to tap while assuming you are holding the phone right side up instead of upside down.
post #82 of 147
Quote:
Originally Posted by Flaneur View Post


Yep, don't know about "stories," but I sometimes try to make up for a general lack of imagination among a group of people who seem to have spent their lives in front of screens and keyboards. Seems to me they way outnumber the cause-and-effect thinkers, and only trade in memes.
 

Are you nuts? I tried to defend you, and yet you admit to making up stuff?

post #83 of 147

The feature I find most interesting are the images themselves. Based on a picture alone we're to believe fully more than half of the screen on any new iPhone is less than 1mm accurate. I mean clearly we are not just talking about the edges, the entire top half of the screen basically wouldn't work if you had to shift 1mm for every action. Christ what BS. The screen depicted in the images probably would most likely be completely unusable.

turtles all the way up and turtles all the way down... infinite context means infinite possibility
Reply
turtles all the way up and turtles all the way down... infinite context means infinite possibility
Reply
post #84 of 147
Took a bit of digging but Guess who is this companies major partner :Microsoft/Nokia

HPC

http://www.youtube.com/watch?v=JTTwsQUYnoI&feature=youtu.be

Makes sense since Nokia is in Finland

Gotcha
post #85 of 147

I have observed this issue when using the Maps application.  Tapping the End button on the top left corner of the screen almost never responds.  It is extremely frustrating.

post #86 of 147
Quote:
Originally Posted by GeorgeiP5 View Post

Something tell me that their top manufacturer samsung had something to do with it. It's weird how the iPhone 5s works so well but when the amazing 64bit A7 iPhone with ID sensor came the screens sensitivity is bad. I blame samsung.

Probably you are right. We know now, that Sammy pays students to write bogus articles against some competing products/companies, so why shouldn't they bribe a finnish company to fake some test results? To see the 100% perfect Galaxy3 test result besides the two iPhones, seems more than suspicious.
post #87 of 147

I am not sure how to read this report.

 

If you look at the touch map they provided for the phone it not only has an issue at the edges it has issues over most of the phone base on all the red dots. I am thinking something is fundamentally wrong. I think most would agree that Apple would not go from best to worst from one generation of products to the other. I also find it interesting that a whole group of people did not seem to think there was an issue and now that this has been pointed out they are say it explains their typing mistake. Maybe people have always made similar mistakes, but now that can blame it on something else other than themselves, why should people take responsibility for their actions.

 

If we are to believe this error map they provided then most everyone would be experiencing errors all over the display especially when playing games since you tend to use most of the display with games. At this point I am thinking apple changed something with the display that the calibrated finger no longer represents a touch which the digitizer recognized correctly, however, the digitizer still closely works with a real human finger.


Edited by Maestro64 - 10/25/13 at 12:40pm
post #88 of 147
Rofl, the keyboard is better on an S3?????? Have you tried the keyboard, it's such a pos!
post #89 of 147
Quote:
Originally Posted by newbee View Post

I agree ... this "report" doesn't pass the smell test. The first sentence in their "report" :   Out of curiosity, OptoFidelity test patrol engineers wanted to see what differences in performance our untouchable robot systems could find between iPhone 5 S and iPhone C and would that difference guide consumer to pick right version. 


Out of curiosity ???   Give me a break ... and they were just as "curious", I suppose, to "test"it against Galaxy S3 .... how "convenient".  I sense this is more of Samsung's dirty tricks campaign.  Just another paid shill working for the world's most shameful company.  1oyvey.gif

I never thought my thumb was millimeter precise to begin with as a pointing instrument. Or put differently: when I type, I'm not hitting the on-screen keys with any real precision, I'm relying on autocorrect to do most of the heavy lifting, or alternatively, to take the blame for any weird things I end up typing 1wink.gif

"Apple should pull the plug on the iPhone."

John C. Dvorak, 2007
Reply

"Apple should pull the plug on the iPhone."

John C. Dvorak, 2007
Reply
post #90 of 147
Quote:
Originally Posted by StruckPaper View Post
 

I agree with that. But that makes him (or her) arrogant and rude, not wrong.

Noting not experiencing a particular issue does not implicitly dispute the existence of the issue but rather adds to the numerical data of whether the issue is widespread or of a more limited nature. Otherwise the thing should just be locked down after the first post of the incident. My take is a real but of minor occurrence or real and very widespread occurrence of an item is a useful piece of information regarding an issue.

 

Having a different experience, perhaps based on different hardware or circumstances, doesn't attack anyone else's experience as invalid absent one boatload of defensiveness. Hence the acronym; YMMV

post #91 of 147
I have noticed this. Some buttons seem to be closer to the edges in IOS and I thought that was the problem. My bottom right corner is the most noticeable. The P works fine.
post #92 of 147

Is it possible that Apple applies some sort of correction?  Since we are using fingers and thumbs, not precise styluses (as used in this test) to use our devices, that register as blobs on the screen rather than points, perhaps some kind of correction was needed to improve accuracy.

 

And likewise, if it really is a result of inaccuracy, couldn't it be corrected via software?

post #93 of 147
Quote:
Originally Posted by Haggar View Post

Unless someone wants to claim that the robot was purposely set up to test the iPhone differently from the Samsung phone. 1confused.gif  

I think the moral of the story is that you should give robots a Samsung Galaxy S3, because they are capable of precision aim.

When I type on my iPhone 5 screen, I'm flying though it very quickly and there's hardly anything resembling sub-millimeter precision involved. I'm just stabbing at where my muscle memory thinks the buttons are--Apple did say they have predictive algorithms that dynamically expand and contract the target areas of each key based on what you've typed so far. And of course, Apple's excellent autocorrect comes in afterwards to clean up any lingering errors. Anyone who does a lot of rapid typing on iPhones knows what I'm talking about. (I jabbed this message out using my iPhone, for example.)

"Apple should pull the plug on the iPhone."

John C. Dvorak, 2007
Reply

"Apple should pull the plug on the iPhone."

John C. Dvorak, 2007
Reply
post #94 of 147
Quote:
Originally Posted by jfc1138 View Post
 

Noting not experiencing a particular issue does not implicitly dispute the existence of the issue but rather adds to the numerical data of whether the issue is widespread or of a more limited nature. Otherwise the thing should just be locked down after the first post of the incident. My take is a real but of minor occurrence or real and very widespread occurrence of an item is a useful piece of information regarding an issue.

 

Having a different experience, perhaps based on different hardware or circumstances, doesn't attack anyone else's experience as invalid absent one boatload of defensiveness. Hence the acronym; YMMV

So you're counting up all the defensive "no problem here" posts on the internet? Let me know when you are ready to publish your findings. :)

post #95 of 147
Quote:
Originally Posted by StruckPaper View Post
 

So Apple (or you) is 100% sure everyone will hold and tap the device at the angle they predict?

 

No, they cannot. But, they could make a few assumptions and work their way up.


Here is some background to the problem that the Steve Jobs patent tries to solve:

The http://en.wikipedia.org/wiki/Parallax error    (http://en.wikipedia.org/wiki/Parallax#Parallax_in_sights).

 

In order to make the test more believable, I would suggest the following improvements:

 

Each error measured is represented by a vector v=(x,y). That vector is mapped to HSL color space:

  Hue = vector's angle = arctan (y / x),

  Saturation = length of the vector = |v|

  Lightness is fixed at 50%

 

Using that mapping, each tested pixel will get a color, making a map.

 

If that map contains noise (gradients are not smooth), then obviously and undeniably the iPhone 5S is faulty (display and/or OS processing is gravely incorrect).

 

If, however, the there is no noise, and  gradients are smooth, then there is a consistency in the way touches are processed. And then, we have the Steve Jobs patent in action.

 

Instead, what we get is a gravely filtered data set. The direction of those error vectors is not preserved, and a threshold is set to map the magnitude of each vector to a binary (RED/GREEN) value. Both "techniques" strip meaningful information and introduce a huge error in the process of measuring those errors.

 

So, forgive me for being a bit sceptical on that topic, but what I see is pseudo science in action.

 

EDIT: I stand corrected! The research is not completely bogus. There is actually a meaningful image, the third one (http://cdn1.appleinsider.com/1025-touchaccuracy-3.jpg). It is an alternative representation of the map I described above. And it shows a significant consistency in the way touches are detected and processed. That image, however, goes directly against the conclusions of the research. 


Edited by capasicum - 10/25/13 at 1:17pm
post #96 of 147

This test was done by idiots - or someone on the take from Samsung.

 

The edges have less sensitivity _by design_ so you don't accidentally press anything when holding the phone. This is software+hardware integration in action.

 

The fact that the Galaxy doesn't account for this and is showing even sensitivity across the surface just shows Samsung doesn't have a clue - or care - about how their users use their products.

 

I've never had any problem pressing Q, O or P.

post #97 of 147

Poster 'fenevadka' @ Cnet comment board posts this:

"Sometimes it helps to read the actual report rather than the article or other people’s comments. 

If you look at page 4 of the report (http://www.optofidelity.com/wp-content/uploads/2013/10/OF_iPhone5C_vs_iPhone5C.pdf), there is a rather distinct and obvious pattern to the discrepancy that is absolutely consistent with the notion that it is a parallax correction at the OS level and not a sensor problem. It looks suspiciously like the correction you would expect for a device held primarily in the left hand and operated by the right hand.

Now it may be that the compensation amounts are wrong or too much or too little—and the user reports seem to indicate that too much correction at the edges may be the case here—but the idea of correcting for parallax is not a problem per se. Not correcting for parallax could be a bigger problem.

Assuming the results are valid, it doesn't tell us where the issue arises. My guess, again looking at the report and not just other people’s comments, is that the issue arises at a low level in the OS as it is processing the raw input stream from the screen. That is good news because if the correction amounts are wrong an OS update could easily correct them. If the problem was just a bad sensor, then nothing could fix it. (Well, theoretically, a software correction/calibration could do it, but random error distribution wouldn't make that easy.)"

post #98 of 147
Bet u this is sponsored by Samsung or Google
post #99 of 147

One hundred posts in, and not one quip about sharpened fingers.  You guys are losing your touch.

Quality isn't expensive... it's priceless.

Reply

Quality isn't expensive... it's priceless.

Reply
post #100 of 147

Synthetic testing like this rarely converts to real life application. I have to admit it does make me laugh when I read member making it sound like Samsung was the sponsor of this test. I doubt those same people had a problem with the testing methodology when the iPhone tested as having one of the most accurate touch screens in 2010.

 

Lets remember screen dynamics changed from 2010. In any case unless you are Lady Gaga that has her nails shaved to a 1mm point I doubt anyone is going to have any real problems. Everyone in my family has a 5s and the only problem I have is my fat finger hits the space bar.

 

Samsung makes very good hardware which is why Apple has used them for decades and will continue to use them. The problem with Samsung smartphones isn't the hardware it's the fact the put Android on it.

post #101 of 147

I'm finding it hard to try and grab the video-slider on my iPhone 5 now that I upgraded to iOS7. It tends to jump to a completely different position than where my finger is. This never happened with iOS 6, I wonder if this is an OS issue or maybe the slider is way too small to grab.

 

Btw, I'm not a fan of the iPhone's new video-player UI, I find it too opaque and intrusive.


Edited by bloggerblog - 10/25/13 at 1:43pm
bb
Reply
bb
Reply
post #102 of 147
Originally Posted by PScooter63 View Post
One hundred posts in, and not one quip about sharpened fingers.  You guys are losing your touch.

 

This is funnier than the point you’re trying to make.

Originally posted by Relic

...those little naked weirdos are going to get me investigated.
Reply

Originally posted by Relic

...those little naked weirdos are going to get me investigated.
Reply
post #103 of 147

I dislike the slider in the music app too. ugh I guess no one ever scrolls through their music? Why did Apple's designers make it even smaller and basically useless? I mean what's the point of sliding your finger upwards to change the shuttle speed when you can even make contact with the slider? It's the most awkward gesture in iOS7 IMO.

 

Quote:

Originally Posted by bloggerblog View Post
 

I'm finding hard try and grab the video-slider on my iPhone 5 now that I upgraded to OS7. It tends to jump to a completely different position than where my finger is. This never happened with iOS 6, I wonder if this is an OS issue rather than a hardware/sensor issue.

 

Btw, I'm not a fan of the iPhone's new video-player UI, I find it too opaque and intrusive.

turtles all the way up and turtles all the way down... infinite context means infinite possibility
Reply
turtles all the way up and turtles all the way down... infinite context means infinite possibility
Reply
post #104 of 147
Quote:
Originally Posted by Gatorguy View Post

"While the latest test suggests the outside of Apple's iPhone displays may experience inaccuracies, the iPhone still leads the way when it comes to responsiveness. One study published last month by Agawi found the iPhone 5 display to be twice as fast as competing Android-based handsets."


Hmm. . .
So faster mistakes then according to AI. What an odd way to attach a positive to a negative finding (if true). 1bugeye.gif

Or you could say Android devices are slower and less accurate than iPhones. Or do you have evidence to the contrary?
Enjoying the new Mac Pro ... it's smokin'
Been using Apple since Apple ][ - Long on AAPL so biased
nMac Pro 6 Core, MacBookPro i7, MacBookPro i5, iPhones 5 and 5s, iPad Air, 2013 Mac mini.
Reply
Enjoying the new Mac Pro ... it's smokin'
Been using Apple since Apple ][ - Long on AAPL so biased
nMac Pro 6 Core, MacBookPro i7, MacBookPro i5, iPhones 5 and 5s, iPad Air, 2013 Mac mini.
Reply
post #105 of 147
Quote:
Originally Posted by rrobr View Post

Just tried on my 5S. Zero problems with any key in any orientation. Don't recall it being a problem on my 4S or 3GS, either. And I have fat fingers.

Good thing I'm not a robot!

Touch-gate for Robots! LOL.

Of course the same test would be interesting on an array of Android devices.
Enjoying the new Mac Pro ... it's smokin'
Been using Apple since Apple ][ - Long on AAPL so biased
nMac Pro 6 Core, MacBookPro i7, MacBookPro i5, iPhones 5 and 5s, iPad Air, 2013 Mac mini.
Reply
Enjoying the new Mac Pro ... it's smokin'
Been using Apple since Apple ][ - Long on AAPL so biased
nMac Pro 6 Core, MacBookPro i7, MacBookPro i5, iPhones 5 and 5s, iPad Air, 2013 Mac mini.
Reply
post #106 of 147
Quote:
Originally Posted by spliff monkey View Post
 

The feature I find most interesting are the images themselves. Based on a picture alone we're to believe fully more than half of the screen on any new iPhone is less than 1mm accurate. I mean clearly we are not just talking about the edges, the entire top half of the screen basically wouldn't work if you had to shift 1mm for every action. Christ what BS. The screen depicted in the images probably would most likely be completely unusable.

 

Excellent point.  Those pictures seem to imply that only a small section of the screens work well, and that's clearly not the case.

post #107 of 147

That would explain the problem I'm having with one app. This app displays a list of items with checkboxes along the outer left edge. Often, and repeatedly, I try to check the boxes and the app instead responds by bringing up information on the item instead. 

 

As far as keyboard accuracy, I've always had problems with 'M' regardless of phone or iPad. When I want to hit the space bar, m appears instead. Something like "I'mmcomingmhomemformdinnermamlittlemlate."

post #108 of 147
Quote:
Originally Posted by waldobushman View Post

That would explain the problem I'm having with one app. This app displays a list of items with checkboxes along the outer left edge. Often, and repeatedly, I try to check the boxes and the app instead responds by bringing up information on the item instead. 

As far as keyboard accuracy, I've always had problems with 'M' regardless of phone or iPad. When I want to hit the space bar, m appears instead. Something like "I'mmcomingmhomemformdinnermamlittlemlate."

I've had problems with I and O. When I watch for it while typing real fast I can hit the correct letter. If I'm typing real fast and not thinking about it I always hit the other vowel.
post #109 of 147

Before people accuse me of being anti-Apple, or crazy, (or both) let me provide my bona-fides. Mac owner since 1993. I've had every iPhone from the original to the 5s (no 5c.) I'm an Apple fan, a Mac guy, ad an iPhone owner. Search for my previous posts here.

I think this has to do with iOS 7. (I don't have an iOS 6 device to compare it against.)

 

I did this on my 5s, once with my finger, and once with a tablet touch pen that I use when testing application on the Android devices at work.

 

1)First take off any case you have on your iPhone.
2)Go to the home screen of your springboard.
3)Swipe down to reveal the Search Bar (,this will bring the keyboard out too.)

 

Here's where I can see the issue:
Attempt to click the "Cancel" button by tapping the right side of the word directly in the middle and baseline, like on the "cel" letters in "Cancel."

 

Result for me is that the "Cancel" button does not register a press. 

I can get it to register a press by clicking in this way:
1)Click on the capital "C" in "Cancel"
2)Clicking the "cel" of "Cancel" just below the letter's baseline but still within the blue bar.

 

I repeated this same test with a known "good" iPhone 5 and it had the same issue. (It was my old iPhone 5 I sold to a co-worker which had it's touch screen replace in late July.)

Now do this (with the search bar and keyboard still present):

 

1)Click on the right side of the P button on the keyboard. Aim as far right as you can.

Result:The P registers properly and is shown in the search bar. Type a few more letters for the next step.

 

2)Click the delete button on the keyboard (which is just above the Search key at the very bottom) and aim for it's outer right edge.

Result for me: It worked sometimes, but not other times.

 

3)Click on the left side of the "Q" button on the keyboard.

 

Result: The Q registers properly and is shown in the search bar.

 

4)Click on the "123" button at the bottom left of the keyboard. Aim as far left as you can by the "1"

Result: The button is triggered sometimes, but not all of the time.  (Make sure you have some letters in the search are before performing the next step or the button will not be active.)

 

5)Try to click the very right edge of the "Search" button, on the "arch" or "rch" letters.

Result: The "Search" button is not triggered most of the time. You can trigger it by clicking on the "S" or to the right of the "S" (but still to the left of the space bar.)

 

My theory is: I think the software behavior that has the phone attempt to disregard "accidental touches" is suppressing the button pushes when they are near the top and bottom outer edges of the device.

 

This, combined with iOS 7's new layout styles of placing "buttons" to the very edges is leading to some frustration for some users. The buttons I originally posted about being non-responsive in other apps , such as Music or Nike+ are 

Just my thoughts.

post #110 of 147
I notice this on the 5s playing Candy Crush Saga (Please don't judge me!). I get significantly more mistaske moves and fails to detect fingers on the edge than I has with my 4s.
post #111 of 147
Quote:
Originally Posted by tomhayes View Post

Before people accuse me of being anti-Apple, or crazy, (or both) let me provide my bona-fides...

And this is as far as anyone is going to read.

When will you people learn that claiming to own a ton of Apple stuff doesn't make your arguments more valid nor does it make your opinion more credible?
post #112 of 147

I just tested with my 5s by repeatedly tapping each key on the virtual keyboard and didn't see any difference wherever i tapped.   I also have a galaxy S3 and didn't see any difference there either.  One thing i've noticed is that everything is larger on the S3 so you don't have to be that accurate.

 

Btw, why would the testing use the S3 instead of S4?? Could it be because S4 is worse than S3.   Is this company sponsored by Samsung by any chance?

post #113 of 147
Quote:
Originally Posted by tomhayes View Post
 

Before people accuse me of being anti-Apple, or crazy, (or both) let me provide my bona-fides. Mac owner since 1993. I've had every iPhone from the original to the 5s (no 5c.) I'm an Apple fan, a Mac guy, ad an iPhone owner. Search for my previous posts here.

I think this has to do with iOS 7. (I don't have an iOS 6 device to compare it against.)

 

I did this on my 5s, once with my finger, and once with a tablet touch pen that I use when testing application on the Android devices at work.

 

1)First take off any case you have on your iPhone.
2)Go to the home screen of your springboard.
3)Swipe down to reveal the Search Bar (,this will bring the keyboard out too.)

 

Here's where I can see the issue:
Attempt to click the "Cancel" button by tapping the right side of the word directly in the middle and baseline, like on the "cel" letters in "Cancel."

 

Result for me is that the "Cancel" button does not register a press. 

I can get it to register a press by clicking in this way:
1)Click on the capital "C" in "Cancel"
2)Clicking the "cel" of "Cancel" just below the letter's baseline but still within the blue bar.

 

I repeated this same test with a known "good" iPhone 5 and it had the same issue. (It was my old iPhone 5 I sold to a co-worker which had it's touch screen replace in late July.)

Now do this (with the search bar and keyboard still present):

 

1)Click on the right side of the P button on the keyboard. Aim as far right as you can.

Result:The P registers properly and is shown in the search bar. Type a few more letters for the next step.

 

2)Click the delete button on the keyboard (which is just above the Search key at the very bottom) and aim for it's outer right edge.

Result for me: It worked sometimes, but not other times.

 

3)Click on the left side of the "Q" button on the keyboard.

 

Result: The Q registers properly and is shown in the search bar.

 

4)Click on the "123" button at the bottom left of the keyboard. Aim as far left as you can by the "1"

Result: The button is triggered sometimes, but not all of the time.  (Make sure you have some letters in the search are before performing the next step or the button will not be active.)

 

5)Try to click the very right edge of the "Search" button, on the "arch" or "rch" letters.

Result: The "Search" button is not triggered most of the time. You can trigger it by clicking on the "S" or to the right of the "S" (but still to the left of the space bar.)

 

My theory is: I think the software behavior that has the phone attempt to disregard "accidental touches" is suppressing the button pushes when they are near the top and bottom outer edges of the device.

 

This, combined with iOS 7's new layout styles of placing "buttons" to the very edges is leading to some frustration for some users. The buttons I originally posted about being non-responsive in other apps , such as Music or Nike+ are 

Just my thoughts.

 

All worked fine, nothing odd on my end. I'm wondering if this is actually an issue with some of the sensors.

bb
Reply
bb
Reply
post #114 of 147
Quote:
Originally Posted by iaeen View Post


And this is as far as anyone is going to read.

When will you people learn that claiming to own a ton of Apple stuff doesn't make your arguments more valid nor does it make your opinion more credible?

Why don't you read the post and decide for yourself??

post #115 of 147

Quote:

Originally Posted by SurferBoi View Post

[...] you might want to bias the hit points in a different direction to compensate for the fact that your thumb is actually contorting to hit other areas, and different parts of your thumb are useful depending on where it touches.

 

I wonder if there's a way to find out whether or not Apple really does apply weighting to compensate for thumb parallax? If so, it would explain why I get mis-types on keys I'm POSITIVE I hit dead on. Maybe the "hot zone" is not actually within the boundaries of the screen image of the key.

post #116 of 147
Quote:
Originally Posted by tomhayes View Post

Why don't you read the post and decide for yourself??

Good post but you don't need to state you've owned Apple products since 1993.

It seems reasonable. A bit too long for my ADD.
post #117 of 147
Quote:
Originally Posted by StruckPaper View Post
 

Are you nuts? I tried to defend you, and yet you admit to making up stuff?

 

Humor, yes?

 

If not, please go back and read what (s)he wrote: "...make up FOR a general lack of reading comprehension..."

 

I thought it was one of the better posts in this thread, and made Flaneur seem a much more sympathetic character!

post #118 of 147
When did SadScum buy OptoInfidelity?

Clearly my iPhone was never designed to be used with a pointy stylus, unlike other brands that aren't clever enough to work out where my finger is touching.
post #119 of 147
💤💤💤💤💤💤

It doesn't mean anything and no big deal but great for those who love to nit pick.
post #120 of 147
Forgive me if I missed someone else say this. Could this be due to the software for dealing with accidental/stray touches on the edges? That the robot isn't subtle enough?

Just spitballing.

Plus, the S3 image is a perfect frame. The iPhone looks alive.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: iPhone
  • Robotic testing finds touchscreen inaccuracies at edge of iPhone display
AppleInsider › Forums › Mobile › iPhone › Robotic testing finds touchscreen inaccuracies at edge of iPhone display