I don't have too much trouble with the Droid's touch screen, but I also make extensive use of the pinch zoom.
Well you got a droid when you have pinch to zoom handed to you on a platter. Back in my day, the only zoom we had was getting the phone real close to our face, and we liked it!
Yeah, they just came out with it. Moe flipped when she got the update and realized her phone now has the same function that my 1st Gen iPhone has had all along.
So, that testing is flawed because it's leaving out an important parameter: stroke speed. I was wondering what was going on, because I couldn't get lines on my Droid to appear that jaggy. When I used light pressure and drew slowly, the lines became very jaggy. You know why that happens? Because you can't hold your finger that steady when you're moving that slowly. The touch screen is registering your finger's "drift." Draw lightly and somewhat quickly, and you get perfectly straight lines.
I need to test an iPhone side-by-side with my Droid, but I really don't think the touch screen is the issue. It's the way people are interacting with the touch screen that creates the problems. It's sort of like a MetroCard; you can't swipe it too quickly or too slowly. You have to go at a natural pace. Don't overthink the Droid's screen and it works fine.
So, that testing is flawed because it's leaving out an important parameter: stroke speed. I was wondering what was going on, because I couldn't get lines on my Droid to appear that jaggy. When I used light pressure and drewslowly, the lines became very jaggy. You know why that happens? Because you can't hold your finger that steady when you're moving that slowly. The touch screen is registering your finger's "drift." Draw lightly and somewhat quickly, and you get perfectly straight lines.
I need to test an iPhone side-by-side with my Droid, but I really don't think the touch screen is the issue. It's the way people are interacting with the touch screen that creates the problems. It's sort of like a MetroCard; you can't swipe it too quickly or too slowly. You have to go at a natural pace. Don't overthink the Droid's screen and it works fine.
It's not about the screen itself, or your finger. Pretty much all these phones have the exact same, or very similar, capacative touch screen hardware. The difference is the software algorithms that the phones use to translate the raw input data into the useful input data.
A computer mouse, a touch screen, and an analog stick are all two-axis analog input devices. They constantly send bunches of X,Y coordinates to the CPU. If that raw input data were sent directly to application software, it would never be still, and never move smoothly. Analog sticks and mice have the advantage of friction. It's pretty easy to keep a mouse or stick perfectly still by not moving them. This is why they have options like mouse smoothing, or mouse filtering.
On a capacitive touch screen, it's not so much the position of your finger, but the capacitance between your finger and the screen, measured in farads, that must be translated into those X,Y coordinates. Even if you were a zen master, who could stay perfectly still, there are always going to be fluctuations in the capacitance on a very small scale. Depending on the resolution and sensitivity of the screen, the raw input data will be fluctuating. It's up to the software to turn that raw data into a simple Cartesian coordinate that the application software can understand and use as a useful input.
You can clearly see that the iPhone's touch screen driver has some sort of intelligent magic. They probably spent an insane amount of time in R&D; on that driver until the lines were so smooth. The other guys probably didn't even think to work on that, and probably just use a simple off-the-shelf solution like just taking an average. They probably didn't experiment with all kinds of different sampling rates and filters to find the perfect one. That's where the difference is.
You can clearly see that the iPhone's touch screen driver has some sort of intelligent magic. They probably spent an insane amount of time in R&D; on that driver until the lines were so smooth. The other guys probably didn't even think to work on that, and probably just use a simple off-the-shelf solution like just taking an average. They probably didn't experiment with all kinds of different sampling rates and filters to find the perfect one. That's where the difference is.
Right, I figured that's what the deal is. Though, I'm not sure if it's then that the Droid is less accurate per se, or that it's more sensitive than you're expecting. The iPhone sort of "corrects" your input intelligently to what you were most likely trying to do. The Droid seems to more reflect your raw input with no real "smoothing." So, if you're used to the iPhone making up for your own inaccuracies, the Droid will be a lot less forgiving. If you're, say, trying to pick a single link out of a bunch of crowded links, you need to be dead on, and your finger doesn't always land where you aim.
I feel like the iPhone's intelligent averaging is better for end user functionality, but I think it might not be suitable for applications where you would want literal input translation.
I feel like the iPhone's intelligent averaging is better for end user functionality, but I think it might not be suitable for applications where you would want literal input translation.
Almost every time you touch the iPhone, it registers exactly where you wanted it to register. On other phones not so much. What possible application could there be where this isn't the desired functionality?
it registers exactly where you wanted it to register
But that doesn't necessarily correlate to the place where you actually touched.
What possible application could there be where this isn't the desired functionality?
An application that relies very heavily on the accuracy of a given input. Think about a sniping game. If you wanted a true level of challenge, you wouldn't want to even out the input, because that wouldn't be a true test of aim.
For things like button pressing and links, though, you would want that evening out, because allowing for fast and precise input is more important than requiring accurate input in those cases.
It's like everything else that Apple makes: streamlined for user-friendliness. It just so happens that the majority of what you want to do with a smartphone is made a little simpler with a more streamlined interface.
Though, again, it's worth noting that I have no real issues with the Droid's input. Maybe it's because I've never owned an iPhone and thus never got used to it "correcting" my input for me. That forces me to be very very accurate with my screen touching. I'm wondering if a lot of the issues people are having with the Droid are the result of having their input auto-corrected for them by the iPhone.
So, that testing is flawed because it's leaving out an important parameter: stroke speed.
Isn't this issue the very first thing the test addresses? I think the problem is in the raw misplacement of data points, and not any kind of smoothing algorithm. In fact, moving your finger slowly should factor out input smoothing, as it should only happen over the last n-points of data, and not an entire gesture.
The difference in quality is clearly shown in the Android v. Android tests: HTC v. Motorola. Theoretically, if it was all the "fault" of software, we would expect to see similar if not identical results. Yet, HTC provides a good result (despite it's near-axis confusion), while Motorola provides a much lower quality result overall. What we're seeing here is lower quality hardware providing a lower quality result, something that surprises me not in the least when it comes to Motorola (Rachelle owned a RAZRv3, perhaps the worst phone ever made, and I just pawned off my V750, another "winner".)
EDIT: just clearing up my wording a little. Writing code does little for my ability to write in English it would appear, haha.
Isn't this issue the very first thing the test addresses?
Nope. Rewatch the video. Go to 1:05 and watch how long it takes him to draw a line on the iPhone. Then, go to 2:27 and watch how long it takes him to draw a line on the Droid. Nearly twice as long to draw an equally long line. I'm saying that he did not adequately control for speed, and irrespective of that, it's impossible to actually draw a truly straight line with your finger. That's what I'm saying. The Droid isn't fucking up your input; your input is not a straight line. The iPhone averages out the line so it looks like you drew a straight line.
Yet, HTC provides a good result (despite it's near-axis confusion), while Motorola provides a much lower quality result overall.
Except that they used two different drawing programs and two different line thicknesses. Watch the video again. Look at the HTC lines closely, and you'll see a lot of artifacts and bumps. They appear straighter because they're thinner. The Droid lines look messier because they're thicker.
The Droid is more accurately reflecting your real input. The erroneous assumption in these tests is that you're inputting a straight line. You're not. You actually can't. The iPhone just makes it appear like you can.
In fact, moving your finger slowly should factor out input smoothing, as it should only happen over the last n-points of data, and not an entire gesture.
If it updates the last n data points, moving your finger at X speed generates 2n data points. Moving your finger at 2X speed generates n data points. The slower you move, the more input you generate, and you start to see more erratic finger movement.
Whatever the iPhone smoothing algorithms do, they do more of it than the Android algorithms do.
Does anyone know of a really good stylus for a capacitive touch screen.
Aparently, Hawk from applegeeks rates the Pogo pen as quite good - I'm not an artist myself, so I can't give a personal recommendation, but it's the best one I've heard of.
The Droid is more accurately reflecting your real input. The erroneous assumption in these tests is that you're inputting a straight line. You're not. You actually can't. The iPhone [ed: and HTC, apparently!] just makes it appear like you can.
Interesting ... I could see why this would happen if the Droid had a particularly high polling rate in comparison to the others, and applied no filtering to the input. The article seems to think it would be the opposite, however:
Instead, the lines look jagged or zig-zag, no matter how slowly you go, because the sensor size is too big, the touch-sampling rate is too low, and/or the algorithms that convert gestures into images are too non-linear to faithfully represent user inputs.
Even if this is the case, however, I argue that there are few cases where it would be useful to use 100% raw input. So few, in fact, that its exclusion seems negligent, not insightful.
I think this is more the issue. I also disagree with many of the article's conclusions. I'm very doubtful that the jagginess has to do with sensor size or a low touch-sampling rate; if anything, I would expect it to be higher in Android phones. I'm betting it has to do with the conversion algorithms. That makes a hell of a lot more sense to me, especially considering that Apple is very big on making its software as user-friendly as possible. Again, the flaw in the test is the assumption that drawing your finger in a straight line actually creates a straight line. Try grabbing a piece of white paper and some food coloring or dye or something, and trace your finger in a "straight" line across the paper. It'll be jaggy, I guarantee.
Also, I don't really think the HTC draws cleaner lines. Given what I saw for speed control in the video, there's no guarantee that all of the lines in the photos there were drawn in exactly the same way. In the video, the lines on the HTC are noticeably thinner, which makes the image appear cleaner. Close inspection shows a ton of "bumps" that would be far more exaggerated with a thicker line.
Even if this is the case, however, I argue that there are few cases where it would be useful to use 100% raw input. So few, in fact, that its exclusion seems negligent, not insightful.
For most applications for most phone users, I agree. I contend that the Droid's touch screen and Android's algorithms actually result in more accurate interpretations of input than the iPhone. The problem is that the user is not good at giving precise input, so the Droid shows you how inconsistent your input is. This is not necessarily a good thing in all applications.
The iPhone makes your life easier by cleaning up your input. This makes it very easy to use it, and use it in a consistent manner. I'm actually saying that I think the iPhone's interface is easier to use, and thus better. The Droid's interface takes a lot of getting used to, and it's unforgiving. The iPhone is very forgiving. I think that's the big difference between the two.
I'm just indignant because the test doesn't necessarily lead to the conclusions that were drawn; I don't care about the conclusions, I care about the process that generated them. I demand rigor in my testing! More controls! There are too many uncontrolled variables running around!
I demand rigor in my testing! More controls! There are too many uncontrolled variables running around!
I absolutely agree, their tests were slipshod at best. I tried to include video of my test, a full-resolution dump of the results, and used two line weights for just this reason! Despite that, I concede that besides using some kind of mechanical rig we might not get to the end of this debate.
Pete, I think you're making a false assumption here. You seem to think that some level of increased skill on the part of the user will allow them to be more precise with the screens of the Droid vs. the iPhone. That's simply not the case.
Even if a zen master is holding his finger perfectly still on the screen, there will be fluctuations in capacitance. That's just how this shit works. The device must account for that to be usable at all. I'm sure you've used old Atari paddles. That is a perfect example of an analog input device with absolutely no correction whatsoever. Remember how the pong paddle would just blip all over the place uncontrollably? No amount of user skill can make up for that. Of course, a major part of the problems with those paddles was hardware, but not all of it.
You seem to be thinking that with some amount of skill a user could be more precise using the droid screen vs. the iPhone. That's just not true. Even a zen master holding their finger perfectly still will have changes in capacitance that cause the inputs to fluctuate wildly. The Apple screens and software are a sharp knife, and the others are dull knives by comparison. They're just plain worse.
It's from the same site Scott linked, the link right above the DIY touchscreen test - It was pure chance I looked at the sidebar and found it, Scott should get the credit for linking the site originally.
You seem to be thinking that with some amount of skill a user could be more precise using the droid screen vs. the iPhone.
Did I say that even once? No. In fact, I said that the iPhone's interface is better for a smartphone. I can think of applications which could benefit from very raw input like the Droid has, but those are specialized circumstances. With practice, you can use the Droid just as easily as an iPhone, but you can use an iPhone that easily out of the box with no practice.
Slightly off-topic but I only got around to listen to the episode today and it was discussed in it: Soul Eater isn't a bad series and it has the big advantage on other Shounen fighting shows that it is a self-contained 50 episode series which means nothing gets drawn out to the point of boredom and it all stays nice and clean and fits together.
However, if you're going to watch a show in that manner and don't want to do it just because it's popular right now, I'd definitely go for Hajime no Ippo before that. Hikaru no Go is also a series similar in the manner of production (self-contained, no stretching, stays interesting) though it doesn't really involve fighting per so. However, neither series have really the fantasy world element that Soul Eater has. Only Hikaru no Go touches it slightly with Hikaru being possessed by Sai, but that is really the only supernatural thing in the entire series.
Anyways, I've never heard you guys talk about Hajime no Ippo and that is in my opinion the best shounen fighting TV series created so far exactly because it doesn't suffer from the tactic of "never stop throwing episodes out" like most other shows do.
Comments
I need to test an iPhone side-by-side with my Droid, but I really don't think the touch screen is the issue. It's the way people are interacting with the touch screen that creates the problems. It's sort of like a MetroCard; you can't swipe it too quickly or too slowly. You have to go at a natural pace. Don't overthink the Droid's screen and it works fine.
A computer mouse, a touch screen, and an analog stick are all two-axis analog input devices. They constantly send bunches of X,Y coordinates to the CPU. If that raw input data were sent directly to application software, it would never be still, and never move smoothly. Analog sticks and mice have the advantage of friction. It's pretty easy to keep a mouse or stick perfectly still by not moving them. This is why they have options like mouse smoothing, or mouse filtering.
On a capacitive touch screen, it's not so much the position of your finger, but the capacitance between your finger and the screen, measured in farads, that must be translated into those X,Y coordinates. Even if you were a zen master, who could stay perfectly still, there are always going to be fluctuations in the capacitance on a very small scale. Depending on the resolution and sensitivity of the screen, the raw input data will be fluctuating. It's up to the software to turn that raw data into a simple Cartesian coordinate that the application software can understand and use as a useful input.
You can clearly see that the iPhone's touch screen driver has some sort of intelligent magic. They probably spent an insane amount of time in R&D; on that driver until the lines were so smooth. The other guys probably didn't even think to work on that, and probably just use a simple off-the-shelf solution like just taking an average. They probably didn't experiment with all kinds of different sampling rates and filters to find the perfect one. That's where the difference is.
I feel like the iPhone's intelligent averaging is better for end user functionality, but I think it might not be suitable for applications where you would want literal input translation.
For things like button pressing and links, though, you would want that evening out, because allowing for fast and precise input is more important than requiring accurate input in those cases.
It's like everything else that Apple makes: streamlined for user-friendliness. It just so happens that the majority of what you want to do with a smartphone is made a little simpler with a more streamlined interface.
Though, again, it's worth noting that I have no real issues with the Droid's input. Maybe it's because I've never owned an iPhone and thus never got used to it "correcting" my input for me. That forces me to be very very accurate with my screen touching. I'm wondering if a lot of the issues people are having with the Droid are the result of having their input auto-corrected for them by the iPhone.
The difference in quality is clearly shown in the Android v. Android tests: HTC v. Motorola. Theoretically, if it was all the "fault" of software, we would expect to see similar if not identical results. Yet, HTC provides a good result (despite it's near-axis confusion), while Motorola provides a much lower quality result overall. What we're seeing here is lower quality hardware providing a lower quality result, something that surprises me not in the least when it comes to Motorola (Rachelle owned a RAZRv3, perhaps the worst phone ever made, and I just pawned off my V750, another "winner".)
EDIT: just clearing up my wording a little. Writing code does little for my ability to write in English it would appear, haha.
The Droid is more accurately reflecting your real input. The erroneous assumption in these tests is that you're inputting a straight line. You're not. You actually can't. The iPhone just makes it appear like you can. If it updates the last n data points, moving your finger at X speed generates 2n data points. Moving your finger at 2X speed generates n data points. The slower you move, the more input you generate, and you start to see more erratic finger movement.
Whatever the iPhone smoothing algorithms do, they do more of it than the Android algorithms do.
Here's my own version of the test, done in Simply Draw:
Video of me DRAW'RIN.
Image of the result
Not a bad result, if I say so!
Also, I don't really think the HTC draws cleaner lines. Given what I saw for speed control in the video, there's no guarantee that all of the lines in the photos there were drawn in exactly the same way. In the video, the lines on the HTC are noticeably thinner, which makes the image appear cleaner. Close inspection shows a ton of "bumps" that would be far more exaggerated with a thicker line. For most applications for most phone users, I agree. I contend that the Droid's touch screen and Android's algorithms actually result in more accurate interpretations of input than the iPhone. The problem is that the user is not good at giving precise input, so the Droid shows you how inconsistent your input is. This is not necessarily a good thing in all applications.
The iPhone makes your life easier by cleaning up your input. This makes it very easy to use it, and use it in a consistent manner. I'm actually saying that I think the iPhone's interface is easier to use, and thus better. The Droid's interface takes a lot of getting used to, and it's unforgiving. The iPhone is very forgiving. I think that's the big difference between the two.
I'm just indignant because the test doesn't necessarily lead to the conclusions that were drawn; I don't care about the conclusions, I care about the process that generated them. I demand rigor in my testing! More controls! There are too many uncontrolled variables running around!
Even if a zen master is holding his finger perfectly still on the screen, there will be fluctuations in capacitance. That's just how this shit works. The device must account for that to be usable at all. I'm sure you've used old Atari paddles. That is a perfect example of an analog input device with absolutely no correction whatsoever. Remember how the pong paddle would just blip all over the place uncontrollably? No amount of user skill can make up for that. Of course, a major part of the problems with those paddles was hardware, but not all of it.
You seem to be thinking that with some amount of skill a user could be more precise using the droid screen vs. the iPhone. That's just not true. Even a zen master holding their finger perfectly still will have changes in capacitance that cause the inputs to fluctuate wildly. The Apple screens and software are a sharp knife, and the others are dull knives by comparison. They're just plain worse.
Soul Eater isn't a bad series and it has the big advantage on other Shounen fighting shows that it is a self-contained 50 episode series which means nothing gets drawn out to the point of boredom and it all stays nice and clean and fits together.
However, if you're going to watch a show in that manner and don't want to do it just because it's popular right now, I'd definitely go for Hajime no Ippo before that. Hikaru no Go is also a series similar in the manner of production (self-contained, no stretching, stays interesting) though it doesn't really involve fighting per so. However, neither series have really the fantasy world element that Soul Eater has. Only Hikaru no Go touches it slightly with Hikaru being possessed by Sai, but that is really the only supernatural thing in the entire series.
Anyways, I've never heard you guys talk about Hajime no Ippo and that is in my opinion the best shounen fighting TV series created so far exactly because it doesn't suffer from the tactic of "never stop throwing episodes out" like most other shows do.
That test is really shocking. I want to send that to Motorolla and be all "WTF guys?"
Now we can find out if it's a hardware or software difference.