Audio Battery Life Testing - Should You Use Bluetooth, the Headphone Jack or the Internal Speakers?

[Fabian Lioner]

Disclaimer:

I made this youtube video about this topic, but since I know people here on the forum also like written content I adapted the script into a post and added the graphs :).

If this post doesnt belong here I'd appreciate if a mod would move it to the appropiate place, and if it doesn't belong on the forum at all, i'd appreciate if you let me know.

That said, if you would like to support me, I would love if you could watch the video, but if you prefer a written review, please enjoy!

 

 

Introduction:

We’re gonna find out what is the best option for battery life when it comes to your listening experience. Should you be using the internal speakers, bluetooth headphones or wired headphones. If you like these ideas and want to support this project, leave a like and subscribe so I know which videos you enjoy the most. That said let’s jump right into the tests.

So in order to do this properly I created a controlled environment where every device will have the display set to max brightness, the battery will be charged to 100%, wifi and mobile data will be disabled and the video will be previously downloaded on the device so it runs locally from a video player.

To know which method consumes more battery life I played a 1 hour long video at 100% volume, with bluetooth headphones, the internal speakers and wired headphones and recorded the results.

Then for more specific results I tested the same devices but with a shorter 20 minute clip and compared the battery life with the volume set to 50, 75 and 100%.

For all of these tests I used a Samsung Galaxy Note 10+ and a Huawei D15 laptop.

 

Results:

Now finally to the results… these were really confusing so I’ll try to make them understandable.

 

1 Hour Long Test (100% volume):

 

Ok so first, lets see overall what consumes more power between bluetooth headphones, the speakers and the headphone jack. After playing a one hour long video on the Note 10+ , while using bluetooth we had 93% of the battery left at the end of the test, then using a usb type c to headphone jack adapter we ended up with 90%, 3% less battery life compared to bluetooth, and finally using the internal speakers we ended up with 89%, 4% worse than bluetooth and 1% worse that using the headphone jack.

 

 

Now to take another look at this test, I played the same one hour long video on my laptop and, I got the same result while using both the internal speakers and the headphone jack, they used up 11% of the battery, leaving us with 89%, in comparison using bluetooth only used up 10% of the battery, leaving us with 90% left.

 

This was quite unexpected , on the Note 10+ using bluetooth gives you between  3 and 4% more battery life compared to using either the speakers or the headphone jack. Then on my laptop, using bluetooth only managed a 1% advantage over both  the speakers and wired headphones, but still, in both cases bluetooth was the more efficient option.

 

I never thought I would  be saying this, but if you wanna save battery, use bluetooth…

 

20 Minute Long Test (50 - 75 and 100% volume):

 

Hey but the video is not over, the previous test was with the volume set to 100% on all cases, we also need to know if lowering the volume will give you more battery life.

Let’s start with bluetooth,and as expected, changing the volume of a bluetooth device will only affect the battery life of the bluetooth device and not that of the phone, so with both the Note 10+ and my laptop, the power consumption of using bluetooth was exactly the same at 100, 75 and 50% volume, it only consumed 3% of battery leaving us with ninety seven percent of the capacity left.

 

 

 

Next, while using the internal speakers at 100% volume, the Note 10+ and the D15 laptop had 95 and 96% of the battery left respectively, then at 75% volume both devices had 96% of the battery left and finally at 50% volume using the speakers once again both devices performed the same leaving us at 97% of the battery capacity. If you're using the speakers, you can save around 2% of your battery if you lower the volume.
It’s not that significant of a difference so you don't have to worry about your speakers consuming all your battery at max volume.

 

Finally, while using the headphone jack, here is where I found the most disruptive results. On the Galaxy Note 10+ I used a usb type C headphone adapter since the phone doesn’t have a headphone jack, and this dongle turned out to be really efficient… it performed exactly the same regardless of the volume. At the end of the 20 minute clip, at all volumes it only consumed 3% of the battery leaving us with 97%.

On the other hand, with the laptop, the integrated headphone jack was really power hungry, and at 100% volume, it consumed 7% of the battery leaving us at 93%, this is the lowest score yet.

Then if we lower the volume to 75 or 50% with the headphone jack in both cases it consumes 4% and leaves us with 96% of the battery.

 

So if you’re using your laptop and don't know how efficient your headphone jack is, you should try using an external dongle, and you’ll also most likely get better audio quality.

 

Conclusion:

To summarize and try to make sense of all of this that I just said I’ll try to wrap it up nicely.

If you want to have the best battery life possible, either use  bluetooth, or stick to low volumes with internal speakers and wired headphones, I found that between 50 and 75%  would be the sweet spot in general, but if you absolutely must listen to music or videos at max volume, then bluetooth might just be your best bet.

Also, when you’re using wired headphones, there are some  headphone jacks that are unnecessarily loud and don't sound that great, so I would recommend using an external headphone adapter for both the improved audio quality and power efficiency.

 

So which one were you expecting to be the more efficient option? I for sure did not expect bluetooth to win overall !!

 

 

[The1Dickens]

28 minutes ago, Fabian Lioner said:

So which one were you expecting to be the more efficient option? I for sure did not expect bluetooth to win overall !!

Another interesting analysis would be to set the volume to be the same output "level" for each option and see how an hour of music or whatever affects things. This is very interesting findings, but kind of consistent with use case in my head. After all,  the volume of headphones will be much lower than the internal speakers, because they are closer to your ears. Or, said another way, 100% volume for internal speakers is the typical use case, whereas 50% volume is the typical use case for headphones. But if you could set the db to be "equal" for each use case (i.e. headphones at an inch from the mic, Note10+ at 18-24 inches), that could show a nice "use case" graph.

 

Also, I had figured that Bluetooth would have been the most efficient as it can run at lower power and transmit the data to the headphones to be amplified instead of amplified on the device, as is the case with internal and wired speakers.

[Fabian Lioner]

Just now, The1Dickens said:

Another interesting analysis would be to set the volume to be the same output "level" for each option and see how an hour of music or whatever affects things. This is very interesting findings, but kind of consistent with use case in my head. After all,  the volume of headphones will be much lower than the internal speakers, because they are closer to your ears. Or, said another way, 100% volume for internal speakers is the typical use case, whereas 50% volume is the typical use case for headphones. But if you could set the db to be "equal" for each use case (i.e. headphones at an inch from the mic, Note10+ at 18-24 inches), that could show a nice "use case" graph.

 

Also, I had figured that Bluetooth would have been the most efficient as it can run at lower power and transmit the data to the headphones to be amplified instead of amplified on the device, as is the case with internal and wired speakers.

Thanks for the feedback, I hadn't thought of that scenario, I only did broad comparisons (100 - 75 and 50%) because the testing was already really long at about 12 hours. But your idea is really great. Also, i found that the difference was between 1 and 2% in most cases so we shouldn't worry too much about battery life with audio, and focus more on the display and processor.

[BlueChinchillaEatingDorito]

I like this sort of analysis. For future larger scale analysis, it would be interesting to see the difference across different generations of Bluetooth standards. Like for an older device like my iPhone 6S, could the results be reversed where the headphone jack ends up being more efficient than Bluetooth? Could the onboard AMP and DAC be more efficient than the Bluetooth module? Who knows. 

 

Definitely drop it in the suggestions thread. It's an interesting topic that does affect quite a lot of people as decent sounding Bluetooth headphones are becoming ever more cheaper and thus more popular.

 

[The1Dickens]

Just now, Fabian Lioner said:

Thanks for the feedback, I hadn't thought of that scenario, I only did broad comparisons (100 - 75 and 50%) because the testing was already really long at about 12 hours. But your idea is really great. Also, i found that the difference was between 1 and 2% in most cases so we shouldn't worry too much about battery life with audio, and focus more on the display and processor.

Yeh, agreed. The thing I flubbed in saying was that basically, your conclusion is worded a bit backwards. Listening to internal speakers at low volume or headphones at high volume for same battery consumption is the opposite use cases. You would likely have internal speakers at full volume and headphones at half volume, which would show a larger delta in battery power consumed.

 

But yeh, 12 hours is a lot of testing, so I can see how one would be ready to be done with it by then.

[SpaceGhostC2C]

1 hour ago, Fabian Lioner said:

To know which method consumes more battery life I played a 1 hour long video

Why? Video playback is going to account for most of the battery usage, even more so in the 20-minute runs. For audio testing, why not playing music files with screens off? Sure, it's going to take much longer to make a dent in the battery, but that's the point: if you can't see a significant change in battery, one that allows you to detect differences across devices, in 20 minutes of pure audio, then what you are capturing with the 20-minute video is just margin of error of video playback. If not, then running pure audio, screen off, should be feasible.

 

Quote

This was quite unexpected

On the contrary, this was a foregone conclusion, as the bluetooth headset is powering its own speakers, and therefore running down its own battery, while the other two methods draw directly from the main device's battery. If you would measure total power consumption (i.e., device battery rundown + headset battery rundown) the picture would change.

 

Quote

I never thought I would  be saying this, but if you wanna save battery, use bluetooth…

If you are only concerned about your device's battery, then yes. But if, let's say, you are traveling with a battery bank, your phone, and your headset, how to better handle your overall "power budget" is far less clear.

 

 

[akio123008]

Those are some very small differences right there.

 

Have you done multiple runs and taken an average or was that data obtained from one single run per test? Because if that's the case, the results might be a bit random. 

 

[Fabian Lioner]

Just now, akio123008 said:

Those are some very small differences right there.

 

Have you done multiple runs and taken an average or was that data obtained from one single run per test? Because if that's the case, the results might be a bit random. 

 

I did two runs, and they were exactly the same, so no average was required, the real conclusion is that it doesnt really matter what you use, the display and processor will be the real battery drainers.

 

[Fabian Lioner]

Just now, SpaceGhostC2C said:

Why? Video playback is going to account for most of the battery usage, even more so in the 20-minute runs. For audio testing, why not playing music files with screens off? Sure, it's going to take much longer to make a dent in the battery, but that's the point: if you can't see a significant change in battery, one that allows you to detect differences across devices, in 20 minutes of pure audio, then what you are capturing with the 20-minute video is just margin of error of video playback. If not, then running pure audio, screen off, should be feasible.

 

On the contrary, this was a foregone conclusion, as the bluetooth headset is powering its own speakers, and therefore running down its own battery, while the other two methods draw directly from the main device's battery. If you would measure total power consumption (i.e., device battery rundown + headset battery rundown) the picture would change.

 

If you are only concerned about your device's battery, then yes. But if, let's say, you are traveling with a battery bank, your phone, and your headset, how to better handle your overall "power budget" is far less clear.

 

 

I considered making an audio only test, but for this test round I wanted to address the consumption with the display on since I wanted to know mainly while watching videos, which one was more efficient taking the screen into account. For more in depth tests, and to completely isolate audio you're right, I should make the same test but with the display on.

 

I will probably do a test with the display off later, but this one was more focused on video, and listening to music with the display on.

 

Also, my bet was that the headphone jack was gonna win, since there are some 3.5mm ports that are really efficient, on the phone, using wired headphones was exactly the same as bluetooth, but yeah, since bluetooth doesnt have to worry about volume when it comes to the host device, it makes sense that its the more efficient option.

 

[mariushm]

This is kinda silly.

 

First of all, the pictures, all those bar graphs are kinda using up space and don't provide any help ... when your results is 89,93,90 ... it doesn't help anyone to start from 0 with your graph, just start from something like 80-85.

 

Setting to full brightness ... some phones have ambient sensors and adjust the brightness according to light in room ...

Some phones will have ONE speaker, other phones will have two.   Probably most phones will use a specialized amplifier chip to amplify both headphones and speakers.

The power consumption will depend a lot on the impedance of the headphones... there's gonna be different consumption between 32 ohm speakers and 600 ohm speakers... but it's gonna be obscured by the consumption of the lcd panel backlight which consumes orders of magnitude more.

bluetooth will also vary in power consumption, depending on what codecs the bluetooth headphones supports (do you send lossless audio, do you send aac, ac3, what do you send),  how far from the phone you are also matters because transmission power may increase if you're 5 meters away from phone on a coach/sofa.

 

Ideally you'd use a tool like uCurrent  with a multimeter, connected directly between battery and phone internals : https://www.eevblog.com/projects/ucurrent/

 

This would allow you to log both voltage and current and therefore know exactly how much energy gets consumed by the device.

 

When you have batteries with thousands of mAh , 3% covers a very wide range... and the phone may only drop from 90% to 89% when the internal calculation gets close to 89.5% ... so potentially could be up to 1% of error in that -

 

For example, I could say the original power level was something like 90.4%  (but shown on screen as 90%)  and then drops down to 89% on screen but in reality the level could be 88.7% ... so it's not 1%, it's 1.7% difference . 

[Fabian Lioner]

Just now, mariushm said:

This is kinda silly.

 

First of all, the pictures, all those bar graphs are kinda using up space and don't provide any help ... when your results is 89,93,90 ... it doesn't help anyone to start from 0 with your graph, just start from something like 80-85.

 

Setting to full brightness ... some phones have ambient sensors and adjust the brightness according to light in room ...

Some phones will have ONE speaker, other phones will have two.   Probably most phones will use a specialized amplifier chip to amplify both headphones and speakers.

The power consumption will depend a lot on the impedance of the headphones... there's gonna be different consumption between 32 ohm speakers and 600 ohm speakers... but it's gonna be obscured by the consumption of the lcd panel backlight which consumes orders of magnitude more.

bluetooth will also vary in power consumption, depending on what codecs the bluetooth headphones supports (do you send lossless audio, do you send aac, ac3, what do you send),  how far from the phone you are also matters because transmission power may increase if you're 5 meters away from phone on a coach/sofa.

 

Ideally you'd use a tool like uCurrent  with a multimeter, connected directly between battery and phone internals : https://www.eevblog.com/projects/ucurrent/

 

This would allow you to log both voltage and current and therefore know exactly how much energy gets consumed by the device.

 

When you have batteries with thousands of mAh , 3% covers a very wide range... and the phone may only drop from 90% to 89% when the internal calculation gets close to 89.5% ... so potentially could be up to 1% of error in that -

 

For example, I could say the original power level was something like 90.4%  (but shown on screen as 90%)  and then drops down to 89% on screen but in reality the level could be 88.7% ... so it's not 1%, it's 1.7% difference . 

For the bluetooth testing I used the original apple airpods, using the AAC codec at 16bit 44.1khz, for the wired headphones, i used the standard apple earpods because of how common they are, and for the speakers the phone uses a main speaker and an amplified top speaker, the laptop uses dual downward facing speakers. I understand the level of specificity that you tell me but I figured that for just the daily life use this could clear out some surface level doubts.

 

I didn't claim the testing was neither universal nor perfect, I used % rather than mah because that way is more understandable for non tech heads like us and shows how insignificant the consumption is. The graphs start at 0 for this same reason, at the end of the day, the difference bewteen all these options was minimal, not quite error, but minimal, so if the graphs start at 0 you will barely see the difference, and its the same in real life, I barely notice a difference between bluetooth and the headphone jack for example. Also each % both from the laptop and the phone equals around 40-50mah. Yes I could have achieved more exact results while measuring the exact mah consumption, but the results were already so close that increasing the level of detail one more step would yield the same results.

 

I did all these tests well aware that both the LCD backlight and panel, and the OLED screen of my phone were the main components draining battery, but I made this test to fit my personal scenario of watching videos. This tests show in rough terms the power consumption of the 3 audio options on top of the display and processor, since that is how most people watch videos or music visualizers. For another test I would certainly dive deeper, and do it with the display off, but for the time being this was just a quick look at how efficient most audio options are compared to each other.

[akio123008]

16 minutes ago, mariushm said:

it doesn't help anyone to start from 0 with your graph, just start from something like 80-85.

Sure the differences are so small you can't see them properly at full range, however, that's actually a good thing because it reveals there's little difference. I don't like dodgy bar charts that don't cover the whole range. It sort of ruins the whole purpose of the chart because the whole scale is neglected. I wouldn't do it that way.

 

Other than that, I agree with your whole post. These test results are very questionable because there are just too many variables involved.

 

31 minutes ago, mariushm said:

This would allow you to log both voltage and current and therefore know exactly how much energy gets consumed by the device.

I suspect if you were to monitor the power consumption of a phone with a multimeter, the general fluctuations in power use may already be bigger than the power difference from audio playback. It's sort of like trying to measure seismic activity from a vechicle moving on a dirt road. It's extremely hard to test this as far as I can tell. 

 

 

[Fabian Lioner]

Just now, akio123008 said:

Sure the differences are so small you can't see them properly at full range, however, that's actually a good thing because it reveals there's little difference. I don't like dodgy bar charts that don't cover the whole range. It sort of ruins the whole purpose of the chart because the whole scale is neglected. I wouldn't do it that way.

 

Other than that, I agree with your whole post. These test results are very questionable because there are just too many variables involved.

 

I suspect if you were to monitor the power consumption of a phone with a multimeter, the general fluctuations in power use may already be bigger than the power difference from audio playback. It's sort of like trying to measure seismic activity from a vechicle moving on a dirt road. It's extremely hard to test this as far as I can tell. 

 

 

Thank you for the note, I'll use graphs that show the difference more dramatically.

 

And a way I found that could make the testing more exact, would be sticking to %, but doing the test with the display off and instead of 1 hour, doing a 6-8 hour test with 4 passes instead of 2, so that the difference is drawn out in a longer period and hopefully it will be shown more clearly. But yeah, without diving into more extreme testing or specialized equipment, I cant imagine more ways to make it more exact than the ones suggested above