Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Apple May Adopt 1704 x 960 Resolution Display for Upcoming iPhone 6

pgiguere1 said:
It’s been a year since I started speculating that the iPhone 6 would have an @3x display, and so far a common reaction is always “But Steve Jobs said pixel-doubling was necessary when introducing the iPhone 4!”.

It’s important to understand why @2x made sense to also understand why @3x does:

1) It was an integer scaling multiplier.

Xcode/Cocoa use a “points” system for every possible coordinate/dimension. Those don’t represent pixel values unless you’re using a non-Retina display (@1x). For everything else, the pixel dimensions of elements are found using points * scaling multiplier. For example, an iPhone 5 is seen as being 568x320 points * @2x = 1136x640 pixels.

If you had a floating point scaling multiplier, say @2.5x, you’d end up with floating point pixel dimensions as well. For example, a 85x45 points UI element * @2.5x = 212.5 x 112.5 pixels. Oops, fractions of pixels! Since the screen can’t render those directly, multisampling will be applied to “blend” fractions of pixels (think of anti-aliasing). What it would look like is this (on the right):

Image

You don’t want this blurry mess. Therefore, you don’t want your UI elements to end up with non-integer pixel dimensions. That’s why you want an integer scaling multiplier. In 2010 with the iPhone 4, @2x was chosen in large part because it was the next integer after 1, not because it’s the double. That means @3x would be an equally good scaling multiplier as long as keeping integer pixel dimensions goes.

2) It didn’t require scaling interpolation on older displays

Steve Jobs explained that one pretty clearly at WWDC.

Image

One pixel would become a 2x2 pixel grid. No fancy math had to be done to “scale” older apps on the iPhone 4. But why does this matter? Because scaling interpolations often reduce quality (not for performance reasons, that’s insignificant). Keep in mind however that unlike with the @1x -> @2x transition we had in 2010, this time we’d only have a 50% enlargement rather than 100%.

The thing is, a 50% enlargement with interpolation doesn’t look worse than a 100% enlargement with pixel-doubling, despite the loss of details due to the interpolation.

See those examples I made with bilinear interpolation (iOS’ default interpolation when scaling images):
Image
Image

As you can see, older non-@3x-optimized apps would actually look better on an @3x iPhone than non-@2x-opitmized apps did on an @2x iPhone. Add to this the fact that the screen's pixel density would be higher this time around, and the perceived image quality difference would be even smaller.
Click to expand...

I would like it if they went for @4 instead. What would the resolution be then?
 
I don't get all the time and effort put into this stuff. Seems like a waste of time mocking up what some thing might look like based on something Apple is rumored to be doing. Basically the only news here is Apple is testing a higher resolution display for the larger iPhone. And is that even news? IMO it would be news if Apple wasn't planning to update the display resolution with a bigger phone.
 
pgiguere1 said:
It’s been a year since I started speculating that the iPhone 6 would have an @3x display, and so far a common reaction is always “But Steve Jobs said pixel-doubling was necessary when introducing the iPhone 4!”.

It’s important to understand why @2x made sense to also understand why @3x does:

1) It was an integer scaling multiplier.
......

Yes, they should. But the reality is that most apps (even those using AutoLayout) wouldn't magically scale well to a bigger display like it was a responsive website. They would require significant tweaking before they can do that, and I can't see Apple forcing AutoLayout everywhere just yet. That would essentially force developers to redesign/test their app in around a month, unless which their app would probably look like crap on the iPhone 6. Since that's kind of an unrealistic timeframe, it's the consumers that would end up paying in the end. I think Apple will go through a smoother transition than that.
Click to expand...

this is probably one of the most professional comments that i have seen in MR for a very very long time! thank you! :D
 
Jsameds said:
Something about lag usually shuts them up lol ;)
Click to expand...

Sorry meant to say 2k earlier

Oh no no no when you mention lag you will hear "I have custom firmware and I can over clock my CPU. My Note III benchmarks are better than my brothers S5"

Then when I say benchmarks don't mean much in real world usage. He literally says something to the effect of "yeah I know it doesn't, but still...."

Hahaha that's when I know it's time to pack up and leave #
 
Nickerbocker said:
Previously "new" iOS resolutions were supported by app developers so fast, I really don't think it matters much if they choose some multiple of an existing resolution or just come up with a new one that makes sense.

Early adopters will have to go a month or two with some pixel stretching. No biggie :cool:.
Click to expand...
I would just launch the ios 8 sdk with tools to create apps at the new screen size a few months ahead of time
 
irfaanfaki said:
Lol one of my friends sent me the 9to5mac article earlier and said
"So Apple iPhone 6 is gonna fail because the great bigger screen isn't big enough since 5" is the minimum these days"

Followed by
"And now we see the screen isn't even 1080p, meanwhile over here at Samsung the note 4 will probably be 4k"

Wasn't sure how to reply to him ��
Click to expand...

Easy Answer: Don't care what other people think. Unless you are paid to work in Apple Marketing or PR, all that really matters is what YOU think.

Typical Answer (when specs aren't in Apple's favor): Try to change the subject. A often used one against such facts: "but which single phone model makes the most profit?" (as if that somehow is very important to us consumers).

Typical Answer #2: bash all other phones and OSs slinging words like "abomination" and "stupid". If anyone challenges the bash, call them "trolls". Sprinkle in some "Die Samsung Die" and similar.

But then- should Apple adopt what was formerly bashed- Apple's incarnation of "abomination" will be "best ________ ever" and "shut up and take my money" and anything different than that will become the new "abomination" and "stupid".
 
I'm happy they increase the PPI, because the PPI on the iPhone 5/5s/5c is clearly too low. This became noticable with iOS 7, because thin lines seem often jaggy (or the signal circles in the top bar).
 
pgiguere1 said:
It’s been a year since I started speculating that the iPhone 6 would have an @3x display, and so far a common reaction is always “But Steve Jobs said pixel-doubling was necessary when introducing the iPhone 4!”.

It’s important to understand why @2x made sense to also understand why @3x does:

1) It was an integer scaling multiplier.

Xcode/Cocoa use a “points” system for every possible coordinate/dimension. Those don’t represent pixel values unless you’re using a non-Retina display (@1x). For everything else, the pixel dimensions of elements are found using points * scaling multiplier. For example, an iPhone 5 is seen as being 568x320 points * @2x = 1136x640 pixels.

If you had a floating point scaling multiplier, say @2.5x, you’d end up with floating point pixel dimensions as well. For example, a 85x45 points UI element * @2.5x = 212.5 x 112.5 pixels. Oops, fractions of pixels! Since the screen can’t render those directly, multisampling will be applied to “blend” fractions of pixels (think of anti-aliasing). What it would look like is this (on the right):

Image

You don’t want this blurry mess. Therefore, you don’t want your UI elements to end up with non-integer pixel dimensions. That’s why you want an integer scaling multiplier. In 2010 with the iPhone 4, @2x was chosen in large part because it was the next integer after 1, not because it’s the double. That means @3x would be an equally good scaling multiplier as long as keeping integer pixel dimensions goes.

2) It didn’t require scaling interpolation on older displays

Steve Jobs explained that one pretty clearly at WWDC.

Image

One pixel would become a 2x2 pixel grid. No fancy math had to be done to “scale” older apps on the iPhone 4. But why does this matter? Because scaling interpolations often reduce quality (not for performance reasons, that’s insignificant). Keep in mind however that unlike with the @1x -> @2x transition we had in 2010, this time we’d only have a 50% enlargement rather than 100%.

The thing is, a 50% enlargement with interpolation doesn’t look worse than a 100% enlargement with pixel-doubling, despite the loss of details due to the interpolation.

See those examples I made with bilinear interpolation (iOS’ default interpolation when scaling images):
Image
Image

As you can see, older non-@3x-optimized apps would actually look better on an @3x iPhone than non-@2x-opitmized apps did on an @2x iPhone. Add to this the fact that the screen's pixel density would be higher this time around, and the perceived image quality difference would be even smaller.

Another common comment has been "But shouldn't developers be using AutoLayout constraints, which make apps pretty much resolution-independent? Therefore, who cares about the exact resolution?"

Yes, they should. But the reality is that most apps (even those using AutoLayout) wouldn't magically scale well to a bigger display like it was a responsive website. They would require significant tweaking before they can do that, and I can't see Apple forcing AutoLayout everywhere just yet. That would essentially force developers to redesign/test their app in around a month, unless which their app would probably look like crap on the iPhone 6. Since that's kind of an unrealistic timeframe, it's the consumers that would end up paying in the end. I think Apple will go through a smoother transition than that.
Click to expand...


But what happens if you have an app made for iPhone 5/5S. And you have a UI Element that is say, 999pixels. Then you scale that by 1.5x and you have 1498.5. Then, multisampling will have to be used like you said to blur pixels. Right? Which defeats the purpose of integer scaling...
 
Mattsasa said:
But what happens if you have an app made for iPhone 5/5S. And you have a UI Element that is say, 999pixels. Then you scale that by 1.5x and you have 1498.5. Then, multisampling will have to be used like you said to blur pixels. Right? Which defeats the purpose of integer scaling...
Click to expand...

It's already impossible to have a 999 pixel @2x UI element. You need even number dimensions when designing an @2x asset. Apple thought of it all. Similarly, @3x elements would probably need to have pixel dimensions that are multiples of 3.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back