LG G3 review: A great phone with too many pixels | Hacker News Reader

LG G3 review: A great phone with too many pixels

arstechnica.com

38 pointssizzle43 comments

That's kind of crazy. The thing that gets me with all these things is the energy density in your pocket. 3Ah is a lot of energy. When a LiPoly goes off, it goes off in style. Even the 2Ah one in my Moto G is scary.

And in this 3Ah behemoth, the battery life sucks!?!?

Battery life on the G3 is better than most other smartphones.

I'll be purchasing a G3 as battery life is becoming a nuisance on my Samsung Galaxy 4.

From a rough estimate, a 2x increase in resolution reduces the battery life by 30 percent overall, if everything else is equal. But they usually increase battery size and get a more powerful (therefore more efficient with high resolutions) GPU, to mask that. That's how a 1440p phone can get the same, or almost the same battery life as a phone with 720p screen and a significantly smaller battery.

Now the question that most OEMs don't seem to ask themselves is - how much better would the battery life be, if they kept the old "standard" resolution, and increase the battery life and used a new GPU, too? It could be up to 50 percent longer.

Not to mention that focusing on resolution, and making the screen more expensive that way, takes away from them focusing on other aspects of the screen, such as viewing angles (there was a comparison around between LG G3 and Galaxy S5 viewing angles, which I can't find now, but LG G3 viewing angles were pretty bad).

Focusing on needlessly expensive higher resolution screens also takes component money away from investing in even better cameras, more storage, and so on. For example, instead of paying $70 for a 1440p display, they could be paying $50 for a 1080p one (which are pretty abundant now), and get a $40 camera sensor instead of a $20 one. I, for one, would appreciate such a component choice much more.

The answer to your question is here: http://www.anandtech.com/show/8031/huawei-ascend-mate-2-revi...

iPhone does some of that: resolution constant, GPU up, process shrink: battery life constant.

It's pretty scary what that amount of power can do when released quickly. On the other hand realize that a watt hour is slightly less than a single food calorie. That battery has 10 calories. A bar of chocolate has 150-250.

For that reason alone it's a small wonder that we haven't seen miniature fuel cell technology take off in recent years. If only those membranes wouldn't foul up.

It keeps getting close to being accepted and then vanishes again, for instance:

http://www.iec.ch/etech/2013/etech_0113/tech-1.htm

Hopefully the fuel cells won't run on chocolate bars, I prefer to use them as my own fuel :)

Interesting comparison with food calories, though I don't know how much intuitive help we can get from that - how do I imagine a bar of chocolate "going off" (exploding)?

If we put the bar into a person (to use as food calories), a person can punch someone violently and repeatedly for a while or do other extremely damaging things with very few calories, so again, not much help intuitively. You have to do a lot of work to burn 200 calories.

that's some trigger happy downvoting. So how do we imagine how much power is in 250 food calories?

250 calories'-worth of wheat flour dispersed in a cubic metre of air is enough to explode quite nicely. So there's that.

well okay, but I don't see how that's better intuitively than saying "2.5 ounces of wheat flour dispersed in a cubic metre of air is enough to explode quite nicely". [1]

uh...okay?

In terms of some kind of intuitive help, I don't know how much we can get out of considering food calories, I was hoping other poster could tell us more about how/why to consider food calories, for visualizing stored energy.

[1] based on convert in first lazy link - this could be wrong. http://www.fatsecret.com/calories-nutrition/generic/flour-wh...)

I welcome our chocolate battery overlords.

Unless I have miscalculated, a phone battery has the energy equivalence of less than a milliliter of gasoline.

>A bar of chocolate has 150-250.

So does a hand grenade (around 600kJ, or 167Wh).

Running a game on this thing means that the gpu has to work that much harder to render those extra pixels. I'd rather have a higher framerate than more pixels.

> Android OEMs have shown time and time again that they design, not for a great user experience, but for the showroom, spec sheet, and marketability.

You see this with a lot of their ardent fans, they talk up the specs, as if we are back in the old days.

> You see this with a lot of their ardent fans, they talk up the specs, as if we are back in the old days.

Depends upon which fans. Ardent Android fans are, well, fragmented into at least two groups:

Those who are all about the specs of the latest flagship phone from the various high-end phone manufacturers (Samsung, HTC, LG, etc) and those who wouldn't touch a non-Nexus Android phone with a ten foot pole. The second group generally doesn't care about specs as long as they are reasonably decent for a modern phone.

Very high resolution panels will be needed for Virtual Reality headsets to reach human eye level fidelity. When looking at even this LCD panel from just an inch away from your eyes and through lenses, it's actually quite a low effective resolution per eye. We still have to go A LONG way to go with respect to screen technology.

Also unfortunately this technology is LCD (not OLED), so not suitable for VR applications due to lack of low-persistance.

Can't any LCD be made into a low-persistence display just by pulsing the backlight? I don't see how OLED is necessarily any better.

The persistence is just 1 factor... LCD response time is very slow compared to OLED (~5ms vs ~0.01ms). Every millisecond is important when it comes to VR.

So, the persistence isn't really a factor at all, just the response time.

Well persistence is a factor too. Say your display is updating at 120hz without low persistence... the image stays on screen for 8ms, so you have, say, 1ms of accurate image plus 7ms displaying an inaccurate image. For whatever reason our brains prefer to have no image rather than an inaccurate image, so it helps to turn off the display for that 7/8 of a frame.

While you can do that with LCD, you have to wait 5ms for the pixels to change before you can strobe, so your frame is an extra 5ms out of date. That's my understanding at least, I'm not an expert by any means.

I wouldn't describe that as persistence being a problem, since obviously the LCD can be illuminated for arbitrarily short periods of time. The response time penalty applies regardless of whether you're trying to run the LCD in a low-persistence mode. An OLED's response time advantage over LCDs is across-the-board and completely orthogonal to persistence (where some kinds of backlights could enable LCDs to strobe even faster than OLEDs), and the impact of 5ms extra lag for VR applications will be about the same with or without strobing.

My reasoning is this: if 4K is a new standard resolution for media (TVs, desktop monitors, laptop screens), phones should not be excluded, so there wouldn't have to be any upscaling/downscaling/converting which increases energy consumption and results it lower quality. LG is taking the right path and this screen should be just a test run for an upcoming 4K one.

Won't the high res screen use more power?

It would. But screens are getting more efficient in power usage, do they not? I think that most companies aim for a battery that would sustain a moderate usage throughout the day, which is perfectly reasonable in my opinion.

Most "4K" is really 3840x2160 which is exactly twice 1920x1080. So scaling 4K resolution to a 1080p screen is easy, just average every four pixels into one. This screen though has 1/3 more pixels in both dimensions so scaling is actually harder.

Since a Nexus 5 is already at 445PPI with a 1080p screen I don't see any advantage of going 4K on this form factor. On a large tablet perhaps, but not on a phone.

Maybe I was just assuming a similar point of view and I realize this maybe far fetched, but in my opinion and in very rough terms - the natural progression of the screen concept should be to shrink the pixel to the size of a molecule which would then produce an image indistinguishable from the real matter. Such arguments as "human eye can't make a difference" don't really apply here, wouldn't you agree?

>Such arguments as "human eye can't make a difference" don't really apply here, wouldn't you agree?

"Can you see a difference?" is literally the only thing to care about in a screen. If you can't see a difference it's useless to go any further.

Generally, yes. But even just in resolution choice, you have mainstream vs edge use cases that affect the "can you see it" test, and other technical factors (e.g., try to find a panel with an odd-number width or height)

Then there are other factors like power consumption, durability, thickness, weight, cost to manufacture, environmental impact of life cycle, cost of supporting chipsets, ...

Looks cool except for the customized android part with a 1-2 year half-life. I'm baffled at how companies can't figure out that less software customization mixed with long-term OS upgrades is consistently more valuable to consumers.

He is criticizing the phone maker for making the wrong tradeoff, but really, that's for the consumer to decide. Some people may value the (possibly) marginal improvement in sharpness. Why resist technological improvements?

You realize this is a review right? He's not proposing we ban the phone, or prevent its sale, or in any way interfere with the market. How is he resisting technological improvements, and how is he preventing consumers from deciding?

He is a consumer who has decided the tradeoff being made is wrong, and expressed his opinion publicly.

How long does the battery last on average for their typical usage?

I carry around an external battery in my pocket along with my phone. You can buy one for pretty much any phone. Whenever I'm low on battery, I just plug in the external and continue using it, even if I'm walking. It's really convenient.

I think I'd prefer the big screen size even if the battery dies after only a few hours.

Battery life on the second page.

7 hours 20 minutes browsing on Wifi.

Not bad at all.

If the battery will last long enough for me to get from my 3 metre charge cable in the kitchen to the charger in the car to the charger at work, it'll do.

Meanwhile, those of us who "get out more" would appreciate any improvements that can be made.

oh my, I'm old I remember Nokia's 3310 battery lasting for a whole week :)

Turn off data, wifi, gps and all background apps. You'll be surprised how long your battery lasts.

I have a couple year old Android I use with the SIM card from my home country when traveling to receive emergency calls from work. It usually lasts just over 2 weeks with roaming data turned off.

The screen itself draws a large amount of power; making heavy use of the kindle app used to threaten to exhaust my phone.

Reply on news.ycombinator.com