## How much phone battery can you save by switching to dark wallpapers and themes?

UPDATE 30/01/2011: I've had to correct the maths a little. Fortunately, the numbers that come out are almost identical.

Ever since the invention of flat panel monitors and screens, it didn't matter what colour your pixels were, you'd still consume the same amount of power. That's because these screens have a big flat light at the back that's always on. If a pixel wants to show you white it lets all the light through. If it wants to show you black it blocks all the light. Either way, the light behind is still on and still consuming the same power.

On a mobile phone that means the colour of your wallpaper doesn't affect battery life. But all that changed with the introduction of OLED screens (or more commonly AMOLED and Super AMOLED) and a lot of new phones have these. If your phone has an OLED screen keep reading, you might be able to improve the life of you battery.

# Why is OLED different?

OLED screens don't have a back light. Instead each pixel is an Organic Light Emitting Diode that makes its own light. That means reducing how many pixels you have on during regular use of your phone can have a big impact on battery life. That doesn't just mean changing your wallpaper. You could also look at alternative apps that have dark themes. For example, I spend a lot of time reading news on my phone and I've switched to an app call NewsRob that has the option of light text on a dark background. Here's NewsRob:

# An unanswered question

There seems to be a lot of speculation about just how much battery you can save by changing things like your wallpaper but not much in the way of hard evidence. So I wanted to be a bit more systematic about it and come up with some robust answers. It's not quite scientifically rigorous but I'll tell you my methods and you can judge for yourself. In case you're not interested in the method, the conclusions come first...

# Conclusions

Here are some caveats

• These tests were performed on a Nexus S. I can only assume other Super AMOLED phones will have similar results, but it is just that, an assumption.
• The figures for maximum brightness and minimum brightness are exact. For brightnesses in between I calculate power consumption by assuming that it varies linearly with brightness. This may very well not be true.

So how much more power does a white pixel at full brightness use than a black pixel? First of all, a black pixel does use some small amount of power. I'm not sure why. Perhaps it's to keep the pixel in a state of readiness or something like that.

A full brightness white pixel uses 5.8 times more juice than a black pixel:

That doesn't mean you could ever increase your battery life by 5.8 times. That's for 2 reasons. The first is that no one spends all day staring at a fully white screen and because no one would every want to switch to a fully black screen! The second is that the screen isn't the only thing using power.

But what we have got is the extremes and we can use those to calculate the in betweens. For example, if half of your battery was used up by the screen and you were able to, on average, reduce how "on" your pixels were from a half to a quarter you'd reduce battery consumption by:

$How much phone battery can you save by switching to dark wallpapers and themes?$

That's quite a complicated looking equation but it's not really. If you're interested in where it came from you can have a look at what I'm glamorously referring to as Appendix A. But if you're just wondering where the 4.8 comes from, it's the 5.8 from earlier minus 1!

So this shows an 18% saving in this imagined scenario. Not bad! But is it realistic?

To answer that question we need to look at two things:

• How much of the battery is used by the screen.
• By how much you can realistically reduce how bright your pixels are.

The first question depends on how you use your phone. If you spend a lot of time on calls that's what will mostly be using your battery. But if it's email and Angry Birds it'll be your screen that tops the charts. The good news is, you can find out for yourself just how hungry your screen is. Or at least you can on Android. I don't know about iPhone but iPhone screens are backlit anyway. Same for Blackberry.

From the home screen, press menu button, then Settings, then Applications, then Battery use.

Here's mine:

So 41% of my battery was used by the screen in the hour and three quarters prior to this screenshot. This varies quite a bit so I kept track of it for a while and it averaged out at about 40%.

But by how much could I reasonably reduce the average pixel brightness on my phone? By analysing screenshots and monitoring brightness settings we should be able to work it out. Let's look first at changing wallpaper. I've got no idea what the brightness of the average wallpaper is so I'm just going to go with a 50% grey wallpaper for the purposes of this comparison. Here's a screenshot with the grey wallpaper next to another with black wallpaper:

When the grey wallpaper is used the average pixel brightness is 47%. For the black wallpaper it's just 15%.

Screenshots don't take the brightness setting on your phone into account (a white pixel will come out as white in a screenshot even if your brightness is set low). So we need to factor that in as well. I have my brightness set to auto and this gives me on a typical day an average of about 50% full brightness. If you use your phone a lot outside, this value will be higher.

So plugging all this into our equation means that if I spent all day on the home screen I'd save:

$How much phone battery can you save by switching to dark wallpapers and themes?$

The 0.4 comes from the 40% of battery used by the screen. The 0.5 comes from the brightness setting on my phone. The 0.47 and 0.15 come from the average pixel brightness of the screenshots.

Of course I don't spend all my phone time just looking at the home screen. I do spend a lot of time reading news. Here's a comparison between reading an article on Google's official Reader app for Android and reading an article on NewsRob set to night mode:

The average pixel brightness for Goolge's official app is 89%. For NewsRob it's 21%. So while I'm using NewsRob I'm saving:

$How much phone battery can you save by switching to dark wallpapers and themes?$

That's quite a bit of battery!

Another app that could save your battery is Handcent, a replacement for the default SMS app. If you spend a lot of time sending text messages this might be a good idea. You'll need to customise it to get a dark theme but that's pretty easy. Here's Handcent:

Know of any other apps with dark themes? Let us know in the comments.

If you want to work out how much you could save, read though the Method section below where I talk about how I calculated the average brightness.

# Method

## Calculating white pixel battery use vs. black

To compare white pixel battery use to black pixel use. I just filled the screen with white by looking at a full white image in the galley app full screen and left it like that for 2 hours. I did the same with a full black image.

Some extra steps and precautions:

• Put the phone in airplane mode
• Reboot phone before starting each test
• Prevent the screen from timing out
• Set screen to full brightness

The first two points just standardise the experiments by making sure that other things that might use the battery are minimal. Though this isn't so important because we account for other uses in a later step.

At the end of the two hours I measured the battery level and looked inside the battery app to see what percentage of the battery was used by the screen. Here are the results of that:

 percentage drop proportion of that due to screen Black 14 0.87 White 74 0.96

It's a simple case of multiplying the columns together to get the actual percentage drop due to the screen. That comes out as:

White pixels: 71.0%

Black pixels: 12.2%

So we know that a full brightness white pixel uses:

$How much phone battery can you save by switching to dark wallpapers and themes?$

times more battery than a black pixel.

## Analysing average brightness

First of all you need to take screenshots of your Android. Here's how to do that. Second you need to work out average pixel brightness. To do that I opened the screenshots in PhotoShop, applied an Average Blur filter then used the eye dropper tool to get the colour. Look at the B value of that colour to find the percentage brightness.

# Some final tips

All colours are not equal. I read that blue pixels consume more power. I've not tested this but I might for a future post. But if you do want a colourful wallpaper maybe go for a red and green theme.

When I switched from grey to black wallpaper I didn't change the icons at all. But if you install ADW Launcher as your home screen you can install icon themes. Other apps enable you to do this too. So you could save even more by switching to a dark theme with dark icons. Here's ADW Launcher:

There's a great app called Screen Filter that lets you dim the screen even further than Android allows in the settings. Here's Screen Filter:

Any more tips for saving battery? Let us know in the comments.

# Appendix A - Formula derivation

Here's how I got to the formula I used to calculate battery saving.

So even when the screen is showing all black pixels some power is used. Lets call that power c. The value of c isn't important because it drops out later.

Max brightness full white screen uses 5.8 times this power or 5.8c.

Assuming power goes linearly with brightness then:

$How much phone battery can you save by switching to dark wallpapers and themes?$

where P is power and L is brightness on a scale of 0 to 1 (substitute those values in to verify).

We can calculate L by multiplying the level of brightness derived from the screenshots, l, and the brightness setting of the phone,b:

$How much phone battery can you save by switching to dark wallpapers and themes?$

So we can calculate the change in power consumption between to screenshot derived brightness levels like this:

$How much phone battery can you save by switching to dark wallpapers and themes?$

Which means the fractional change in power consumption is:

$How much phone battery can you save by switching to dark wallpapers and themes?$

To calculate by how much total battery use is affected just multiply this value by the fraction of the battery used by the screen. In my case that was 40% or 0.4 but lets call it $How much phone battery can you save by switching to dark wallpapers and themes?$ for a giggle. Then the fractional decrease in battery drain would be:

$How much phone battery can you save by switching to dark wallpapers and themes?$

which is the equation I've been using in the post.

http://lifehac.kr/i7YZ2s
• Matt

Are your tests for percentage reduction based on recharging it to some fixed point (I assume fully charged) in between or not?

Otherwise you have an assumption based on the linearity of the battery meter, this is better with lithium ion (than say Ni-MH) because the battery normally contains some decent tech to know it’s own state (to stop it blowing itself up when you charge it) but I doubt it’s truly linear. Also the phone mitgh be sensible enough to reduce usage (at the cost of performance) as it got closer to it’s limits.

Also the age of your phone’s screen (in both actual age, hours of emmision and lumens emitted) may change this if the screen has circuitry to compensate for this degradation. Plasmas for instance up the voltage (or current I’m not sure) as they age to deal with the reduction in performance of the phosfors.
Things like this: http://www.freepatentsonline.com/7812795.html seem to imply it

As to linearity of the current to lumens output I can’t find anything simple on it but based on the materials I would expect it to be non linear towards the low end (which this abstract also implies: http://www.computer.org/portal/web/csdl/doi/10.1109/ISQED.2007.20)

Fancy a temperature controlled test harness for your phone that takes it to full battery loss, then an external circuit kicks in, charges it up, reduces the luminance by x% and repeats

• http://stevemould.com Steve Mould

I always did it from 100% charge. And I wondered about the non linear charge drop off too. But looking at the battery graph it’s pretty much a straight line.

As for linear relationship between brightness and current, that really is a big assumption I guess.

If you want to make the test harness I’ll run the tests!

• http://www.alohatony.com Hawaii real estate

Love it. Wow. Amazing post. I’m changing my G2 to dark background immediately!

hey you made lifehacker today!

• Cvrx

Does’nt G2 have a TFT screen? This only applies to OLED screens.

• http://www.alohatony.com Hawaii real estate

I thought it was amoled but now I’m not so sure. Can’t seem to find an
authority on it. where should I look?

• Mike
• BDO

The Kindle app offers a black background when reading books.

• http://twitter.com/dflite Dave

Steve,
Not dark-pixel related, but can I recommend OSMonitor for keeping on eye on when bad apps might be killing your battery:
https://market.android.com/details?id=com.eolwral.osmonitor
It gives you a readout similar to “top” on *NIX, but, more useful than that, it puts a CPU usage icon in your notification bar, so you can see if a naughty app is doing something that’s hammering your CPU 100% when the phone *should* be idle. After 1 week of having my phone, the battery life fell to 5 hours, using this I was able to work out which app was going mental and draining it so quickly.

Oh, and the “top” display has a dark background, so this was a relevant comment after all. Let’s just imagine that that was my intention all along..

–Dave

• Joaquinllacer

I did that. But at the end what you want is a beautifull screen
And see how awesome it is!

• Andrewmcmahon1

Also Try Black Google (bgoog)

• Spencer

Great article! Exactly what I was looking for. Cheers!

• http://wallpapersphone.info/ Eloktavian

Find you at goolge stevemould.
Thx for article.

Nice one & Great Post.

• Sherin

you have done a great job ! actually i am doing a seminar on green computing. actually i am searching for this concept and you give me the scientific explanation too.
thankz!!!!!!!

• http://stevemould.com Steve Mould

Great! Glad it was useful.

• http://twitter.com/nOizGR John noiz G.

Great post, well done!