M1 battery life is terrible, Rosetta to blame?

[Zazoh]

I'll check that today and see what happens from a full drain. I'll leave it running. Any apps to monitor wattage drain?

You can probably get through a terminal command. I don't like installing monitor apps, because they just take whatever info is avail anyway.

but if you start with Activity Monitor on Energy you can focus on what takes the most power and prevents sleep etc.

I need to run to an appointment, I'll check on wattage pull when I return.

 

[darkmatter343]

welp... the movie just finished. Movie was 9.5GB x264 .MKV playing in VLC full screen, with brightness at 4 bars, NO external display. Movie runtime was 1hr 38mins, battery % after movie completion is 94%. ?‍♂️ Battery Health at 100%. So weird.
I'm going to retest Netflix for 1hr and see what happens, also I will retest watching a movie on an external TV through HDMI tonight, using VLC as I'm also wondering how much of a battery drain feeding an external display is.

 

[acidfast7_redux]

welp... the movie just finished. Movie was 9.5GB x264 .MKV playing in VLC full screen, with brightness at 4 bars, NO external display. Movie runtime was 1hr 38mins, battery % after movie completion is 94%. ?‍♂️ Battery Health at 100%. So weird.
I'm going to retest Netflix for 1hr and see what happens, also I will retest watching a movie on an external TV through HDMI tonight, using VLC as I'm also wondering how much of a battery drain feeding an external display is.

I usually see around 5-7%/h when watching PBS, YT, Netflix, Vimeo or some other streaming service. I see approximately the same battery usage with VLC.

I expect that I just ran through HouseMD/DrHouse episodes on loop on VLC, I'd get around 15h of playback on mid brightness.

Your data looks correct to me, to be honest.

 

[darkmatter343]

I usually see around 5-7%/h when watching PBS, YT, Netflix, Vimeo or some other streaming service. I see approximately the same battery usage with VLC.

I expect that I just ran through HouseMD/DrHouse episodes on loop on VLC, I'd get around 15h of playback on mid brightness.

Your data looks correct to me, to be honest.

Yeah, it actually exceeded my expectations. I was expecting to replicate my battery drain, but maybe it was the external display. Loosing 5% for a 1hr 40min movie is something I'll take, but that being said after the movie was done, I close VLC down, left the laptop alone for 1hr with the screen staying on and just the screen saver going. Lost 6% in that 1hr doing absolutely nothing. So I don't know what is going on. 5% for a full screen VLC, and 6% loss for just having the screen saver going for an hour.

 

[inc0gnito]

Yeah, it actually exceeded my expectations. I was expecting to replicate my battery drain, but maybe it was the external display. Loosing 5% for a 1hr 40min movie is something I'll take, but that being said after the movie was done, I close VLC down, left the laptop alone for 1hr with the screen staying on and just the screen saver going. Lost 6% in that 1hr doing absolutely nothing. So I don't know what is going on. 5% for a full screen VLC, and 6% loss for just having the screen saver going for an hour.

I am not sure how reliable the first reading is. I find that first percent from 100% takes around 30-40 minutes to move down to 99%. From there it's a regular drain.

It may be more reliable to do this test from 90% downwards.

 

[darkmatter343]

Ok, another update... I hooked the Air up to an external monitor last night through HDMI to watch a movie, and had both the laptop display and monitor set in mirror mode. The battery was at 82% when I hooked up the external monitor and started the movie, same type of video file, 5GB file, x264 MKV, and playing in VLC and 1hr 30min approx. When movie completed battery was at 70%.
So 12% for a 1.5hr long movie while displaying on both the external display and laptop display ins't bad. Maybe this laptop is just acting normal to prove me wrong ? because watching netflix the other day, and watching a movie on the same external display in mirror mode drained this battery super quick, and nothing else was running.

 

[matrix07]

If I have to guess I would guess it’s a video CODEC. Some CODEC just incompatible with M1 and needs CPU to decode, then the machine will get hot and the battery life suffers.

 

[Krevnik]

If I have to guess I would guess it’s a video CODEC. Some CODEC just incompatible with M1 and needs CPU to decode, then the machine will get hot and the battery life suffers.

Close. You have to use Apple’s codecs to do hardware decode, specifically through either the AVFoundation or VideoToolbox frameworks. Using any other codec will use the CPU to decode.

VLC is built on top of libavcodec, a library of various codecs, which does support VideoToolbox as a decoder. However, it’s been a while since I’ve last worked on libavcodec so I’m not super familiar with how it decides to use VideoToolbox vs one of the software decoders.

It must be something Apple has decoder blocks for, like H.264 or H.265, but there’s potentially other, possibly buggy, logic that decides if the video stream is suitable for VideoToolbox. So if VLC/libavcodec decides not to use VideoToolbox (even if supported), or your video is using a format without a hardware decoder available, decoding will be done on the CPU, which can be quite expensive for 4K content.

 

[shady825]

My M1 MBA battery life has been great!

Same here. I did a full battery cycle a week or two ago just for the heck of it. It really took a lot to kill this battery. I had every app open, screen brightness 100% and multiple 4k YouTube videos playing simultaneously. It still took almost half the day to get the battery to full drain. I'm SUPER impressed with this battery.

 

[darkmatter343]

If I have to guess I would guess it’s a video CODEC. Some CODEC just incompatible with M1 and needs CPU to decode, then the machine will get hot and the battery life suffers.

Close. You have to use Apple’s codecs to do hardware decode, specifically through either the AVFoundation or VideoToolbox frameworks. Using any other codec will use the CPU to decode.

VLC is built on top of libavcodec, a library of various codecs, which does support VideoToolbox as a decoder. However, it’s been a while since I’ve last worked on libavcodec so I’m not super familiar with how it decides to use VideoToolbox vs one of the software decoders.

It must be something Apple has decoder blocks for, like H.264 or H.265, but there’s potentially other, possibly buggy, logic that decides if the video stream is suitable for VideoToolbox. So if VLC/libavcodec decides not to use VideoToolbox (even if supported), or your video is using a format without a hardware decoder available, decoding will be done on the CPU, which can be quite expensive for 4K content.

Thanks for the replies. I'm going to assume it's the codec causing the issue because I'm typing this reply on the M1 Air with the same charge as I started the tested with days ago, and battery is at 60% still.

So basically so far from 100% charge I've watched a 1hr 40min movie full screen on VLC (intel Binary), then hooking the laptop up to a big external monitor through HDMI and watching another 1hr 30min movie mirrored on both displays full screen, not to mention downloading a large amount of files yesterday, some web surfing for about an 1hr, and I'm still at 60%. Man I feel like an idiot but maybe I found the problem in the codec's, because I wasn't aware if it was software based decoding then it'll chew through the battery.

I'm fully expecting things to get much much better in the coming months as more and more apps are ported to be native for Apple Silicon.

 

[Krevnik]

To some extent, less Rosetta will help.

However, I took a peek at what video encoders my M1 Mini is reporting via VideoToolbox. This reports both software and hardware codecs, but includes a flag telling you if it is hardware or not, so I dumped out just the formats

• H.264
• HEVC (H.265)
• JPEG
• Apple Muxed Alpha

Apple Muxed Alpha is a new one to me, looks like a raw format potentially used for video editing.

So, based on the above, H.264 and H.265 are the only video formats that can benefit from hardware decode. Apple uses these two pretty exclusively these days, so it's expected.

So in certain cases, even with everything being native, there will be software decoders used. VP-9 is used by YouTube for 4K, and will wind up using more power for example.

 

[Le0M]

I usually see around 5-7%/h when watching PBS, YT, Netflix, Vimeo or some other streaming service. I see approximately the same battery usage with VLC.

I expect that I just ran through HouseMD/DrHouse episodes on loop on VLC, I'd get around 15h of playback on mid brightness.

Your data looks correct to me, to be honest.

Do you also use an external monitor?

 

[Le0M]

Yeah, it actually exceeded my expectations. I was expecting to replicate my battery drain, but maybe it was the external display. Loosing 5% for a 1hr 40min movie is something I'll take, but that being said after the movie was done, I close VLC down, left the laptop alone for 1hr with the screen staying on and just the screen saver going. Lost 6% in that 1hr doing absolutely nothing. So I don't know what is going on. 5% for a full screen VLC, and 6% loss for just having the screen saver going for an hour.

Keep in mind that the from 100% to 90% the battery drains way slower than - say - from 50% to 40%.

 

[ikir]

Since receiving my M1 Air, I’ve notice the battery drain is pretty fast and it’s not just in one app. I’m wondering if Rosetta is to blame or if it’ll just take more optimization of the Apple apps as well.

For instance, watching Netflix last night through Safari brought the battery down from 100% to 80% which seems steep. In another case, playing a 3hr movie in VLC (which I assume is using Rosetta) and displaying on my monitor brought the battery down from 70% to 20%. Now obviously in this last example VLC is being emulated but to loose 50% for a 3hr movie even using Rosetta seems extreme.

I’ve also notice just using Safari for web browsing I can loose 10%(ish) per hour.

Anyone else experiencing this? Thanks

My M1 battery life is INCREDIBLE. 10% if you have plug in, heavy websites, high brightness could be ok. Check battery settings, to understand which app is draining battery

 

[ikir]

Ok, battery at 100%. I'm runnng a 1.5hr movie (9.5GB in size) on VLC full screen right now, since this is the Intel binary it'll use Rosetta, so once it's done I'll check the battery drain and post back. I have Safari also in the background with 3 tabss open, but according to battery/power monitor it's using next to no drain. I set the brightness to 4bars, whatever that equates to... 20-30%. Will post results back after.

Use iina not VLC

IINA - The modern media player for macOS

iina.io

 

[Le0M]

I don't know... I stopped using an external display several days ago, and yet I don't see a great battery life improvement. I'm barely browsing the web, no other apps running, and with 70% battery remaining it gives me an estimate 7 hours top. Brightness two steps below max.

 

[Zazoh]

I don't know... I stopped using an external display several days ago, and yet I don't see a great battery life improvement. I'm barely browsing the web, no other apps running, and with 70% battery remaining it gives me an estimate 7 hours top. Brightness two steps below max.

What does Activity Monitor show for your top energy user?

Browsing is also not a good test. Some websites consume loads of energy as they are mining data or otherwise trying to act like a TV commercial. And if you are using Chrome, it just does stuff on a schedule whlie it is running. Look for Keystone process.

 

[Manthila]

Activity Monitor hasn't been useful for monitoring energy consumption. Example: Viewing a static PDF page on Edge browser has noticeable power drain, 10% over an hour or two. The same PDF on Safari drains 1-2% battery in the same amount of time. Yet, the Activity Monitor does not view Edge as a problem.

Edge and Safari are both ARM native yet there is a sizeable difference in performance relative to the energy each app consumes. ARM native isn't synonymous with ARM optimised.

Brightness also makes a big difference. Something as minor as 8 pips (from the bottom) to 12 pips is enough to cause noticeable battery drain.

 

[PortoMavericks]

Movie playback is a very CPU intensive workload and that's the reason we have hardware decoders to play videos so that we can offload the workload from CPU. If VLC is not using hardware acceleration, such battery usage is quite normal.

On the contrary, if you’re playing a video from a compatible format with AVFoundation it’ll use very little power from the SoC.

The thing in this case is probably the Wi-Fi that keeps constantly buffering the video, so naturally will lead to more power drain.

 

[Zazoh]

Here is my experience and it is good enough. Using Zoom and VM Ware Horizon, I can sign into work at 7:30 ish and make it to 4:30 EASILY on battery and still have enough to spare for using the computer in the evening for other things. Like the iPad, the M1 battery is an all day battery. And, that is all it has to be.

 

[darkmatter343]

Just a follow up, the battery life has been nothing less than stellar, I'm getting days out of a single charge. Not sure what was initially draining it when I first made this thread but this machine rocks.

 

[Apple_Robert]

Just a follow up, the battery life has been nothing less than stellar, I'm getting days out of a single charge. Not sure what was initially draining it when I first made this thread but this machine rocks.

I think we are getting spoiled being able to routinely get 15 - 18 hrs.

 

[Johnnyangel]

Has anyone mentioned that there's now an M1 version of VLC?

 

[russell_314]

Just a follow up, the battery life has been nothing less than stellar, I'm getting days out of a single charge. Not sure what was initially draining it when I first made this thread but this machine rocks.

Do you think since it was a new laptop it was due to some type of indexing or other initial setup? I know after every macOS and iOS update you will get dozens of people upset about the battery life but it seems to fade away till the next update.

 

[kofski]

I don't think it's normal to have 50 % less battery because of 3 hours of playing movie... I watched Suits (40minutes approx) on the train and I had 96 % battery (from 100 %).

BTW even watching series during traveling via train (7 hours total) didn't drain my battery completely.

Of course, VLC Media with m1 support was used.