• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Apple's new M1 Mac Mini as source (misses out on silent operation)

OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
Yes, the M1 cpu does even better by following what phones have all along. 4 x high power and 4 x low power cores.

Interesting. I am not familiar with silicone computer chips but I assume past chips were 4x/8x high power cord only etc.

So I guess M1 has both mobile efficiency and desktop power.
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
apple could use a bigger and taller fan. Of course a bigger fan means it can operate at lower rpm to move same amount of air. Since rpm affects noise, yes it will be quieter.

Or in the M1’s case I guess they just used a processor that can run cooler. . .
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
I have Question for Mac + Apple Music App (iTunes) > External DAC users.

So after experimenting with the Equalizer on Mac’s Apple Music player with USB output to DAC, I just realised something odd happens.
I can work around it but I suspect it is some kind of bug.

- Whenever starting playback with EQ ”on” there is a delay before EQ actually turns on, for the first 2-3 seconds of playback EQ is disabled.
- When letting playback continue without pause there is no issue, let a playlist play or even change to another track manually and the EQ stays on, everything is fine.
- If you pause playback, and start playing another song it happens again, it takes 2-3 seconds for the EQ to turn on.

Setup: M1 Mac Mini with Apple Music Player (Equalizer on)>USB cable>SMSL SU-8 DAC

Playback method: ALAC files stored on Mac. EQ on.

If you have Mac and Apple Music player, you can test this by turning on the Equalizer and creating a noticeable “manual” EQ by cutting everything above “1khz” by -12dB, or cutting everything by -12dB and hearing the difference in loudness when the EQ turns on after delay of 2-3 seconds.

Can anyone confirm this happens with them as well? Is it a bug?
 
Last edited:
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
Additional Note:

- Leave M1 Mac Mini Bluetooth on

Since Bluetooth was not being used, I thought one time to turn off Bluetooth on the Mac.
Straight after the iOS iTunes Remote App started to have bugs and issues, playlist cover images would randomly stop loading, sometimes I would lose connection and when going to the iTunes Remote start screen the "iTunes Library" would not appear (cannot initiate connection).
I turned the Mac Bluetooth back on, and all the problems went away.
It seems like the iOS iTunes Remote App relies on Bluetooth connection as well as WiFi to work properly.
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
I have been less active here on ASR recently.
But I have had some more experience regarding “Auto MacOS updates” which is a crucial part of the headless M1 Mac Mini setup.

So I have checked on the Mac Mini once before to make sure updates were installing automatically and they were. The Mac would install, restart and I would have the Apple Music Application (iTunes) set to auto open after auto login. Everything is seamless and working as intended.

But I checked just recently just to add a few songs and the update was asking for the Mac password (not Apple ID) before going ahead and installing (this is not “automatic” which is needed for headless operation), this was after the MacOS Big Sur upgrade.

It seems there is a differentiation by Apple between “updates” (smaller patches) and “Upgrades” (big OS changes such as Big Sur) and so “automatic updates” switched on only applies to “updates” and not “Upgrades”.
This meant with “Upgrades” user intervention is needed to install them.
There is no option called “automatic upgrades”.

Luckily these MacOS “Upgrades” are usually once a year or so, so it is not very often. But just thought those who are interested in using MacOS headless should know this.

If anyone thinks this information is incorrect or inaccurate please feel free to correct it.
 

Jazz

Active Member
Joined
Mar 12, 2021
Messages
136
Likes
76
I've had an M1 for many months as a working Mac. I never hear the fan. It never seems to come on. I am mostly using Apple Mail, Safari (processor hog for some badly coded web sites), Busy Cal, Busy Contacts, Apple News, and music ALAC via Apple Music. Occasionally Piximator Pro. A bunch of small background crap like Healthier, Brightness Slider, Copied, etc.
Note that when you first get it, Spotlight will run for many, many days. The fan might come on. Although, if you are fresh and not moving from Mac to Mac, this might not happen or just a day or so. TimeMachine will also come on and run like mad too long as it clears out stuff. No user control over that. This is Apple being a dick.
The M1 will for sure go to my AV system as a music server when I replace it. The M series is probably going to just get better and better but for music, which is nothing for even old Mac CPUs from the 1990s, none of that matters at all.
The only thing missing from Macs now is a CD drive. I still have about 4 CDs I need to burn on an older Mac and move over to Apple Music collection I have. I am too cheap to buy a portable drive. Plus, I don't want more stuff on my desk…
Note: Unless you are doing HD video editing, just get the lower RAM model. The way the chip works is TOTALLY DIFFERENT from older Macs and Intel. So you do not need piles of RAM.
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
I've had an M1 for many months as a working Mac. I never hear the fan. It never seems to come on. I am mostly using Apple Mail, Safari (processor hog for some badly coded web sites), Busy Cal, Busy Contacts, Apple News, and music ALAC via Apple Music. Occasionally Piximator Pro. A bunch of small background crap like Healthier, Brightness Slider, Copied, etc.
Note that when you first get it, Spotlight will run for many, many days. The fan might come on. Although, if you are fresh and not moving from Mac to Mac, this might not happen or just a day or so. TimeMachine will also come on and run like mad too long as it clears out stuff. No user control over that. This is Apple being a dick.
The M1 will for sure go to my AV system as a music server when I replace it. The M series is probably going to just get better and better but for music, which is nothing for even old Mac CPUs from the 1990s, none of that matters at all.
The only thing missing from Macs now is a CD drive. I still have about 4 CDs I need to burn on an older Mac and move over to Apple Music collection I have. I am too cheap to buy a portable drive. Plus, I don't want more stuff on my desk…
Note: Unless you are doing HD video editing, just get the lower RAM model. The way the chip works is TOTALLY DIFFERENT from older Macs and Intel. So you do not need piles of RAM.
Silent operation has been my experience with the M1 Mac Mini as well.

The way I understand the optimal performance with less RAM though is the chip/OS uses the Solid state memory (SSD) as some sort of frozen "cache" in combination with the RAM (sorry I read up on this awhile ago this is all I remember).
This means more writes to the SSD than normal, and apparently getting more RAM does not stop the M1 chip/OS doing this “SSD caching”, it is a normal part of system operation.
 
Last edited:

elvisizer

Active Member
Joined
Sep 16, 2018
Messages
264
Likes
212
@Jazz is talking about the unified memory model on apple silicon @mononoaware
everything you said is happening on m1's as well, the unified memory architecture is in addition to what you were talking about.
Basically, there's no separate pools of ram for the GPU and CPU, so once something (say a texture) is loaded into RAM it can be accessed by the GPU at that memory address rather than having to then be copied to the GPU's own memory. This to applies all parts of the SoC on the m1s (not just the GPU) so it adds up to using a lot less memory in general compared to other processors
 

Beershaun

Major Contributor
Forum Donor
Joined
Oct 3, 2019
Messages
1,874
Likes
1,921
@Jazz is talking about the unified memory model on apple silicon @mononoaware
everything you said is happening on m1's as well, the unified memory architecture is in addition to what you were talking about.
Basically, there's no separate pools of ram for the GPU and CPU, so once something (say a texture) is loaded into RAM it can be accessed by the GPU at that memory address rather than having to then be copied to the GPU's own memory. This to applies all parts of the SoC on the m1s (not just the GPU) so it adds up to using a lot less memory in general compared to other processors
Yep. Pretty impressive and not needed for audio server or renderer. Raspberry pi is the much better renderer solution in all cases.
 

elvisizer

Active Member
Joined
Sep 16, 2018
Messages
264
Likes
212
Raspberry pi is the much better renderer solution in all cases.
sure- they're cheaper, they use less power, and you've got all the choices for output devices available. I've got a pi with an allo digione in my living room as a roon renderer myself- and sometimes I swap the flash card on it and play some SNES on the tv with it too :). the pi's are great, versatile little computers.
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
@Jazz is talking about the unified memory model on apple silicon @mononoaware
everything you said is happening on m1's as well, the unified memory architecture is in addition to what you were talking about.
Basically, there's no separate pools of ram for the GPU and CPU, so once something (say a texture) is loaded into RAM it can be accessed by the GPU at that memory address rather than having to then be copied to the GPU's own memory. This to applies all parts of the SoC on the m1s (not just the GPU) so it adds up to using a lot less memory in general compared to other processors
Thanks for clarifying that. Ok I understand now "unified memory" is talking about the single centralized memory source.

I guess one down-side to what I was talking about (further optimised and increased use of solid-state/SSD caching) is that it results in more writes to the SSD which means reduced drive longevity.
Some customers were able to confirm this with Apple's own diagnostic tools/Applications. The M1 system does a lot more writes to the SSD even when the system is barely being used (just sitting idle) which is another behaviour they noticed was quite different from previous Mac behavior.

Another interesting piece of information I came across: While idle, the M1 chip consumes slightly more power than the previous Intel chip equivalent. While under load the M1 chip consumes considerably a lot less power.
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,769
Likes
8,139
Thanks for clarifying that. Ok I understand now "unified memory" is talking about the single centralized memory source.

I guess one down-side to what I was talking about (further optimised and increased use of solid-state/SSD caching) is that it results in more writes to the SSD which means reduced drive longevity.
Some customers were able to confirm this with Apple's own diagnostic tools/Applications. The M1 system does a lot more writes to the SSD even when the system is barely being used (just sitting idle) which is another behaviour they noticed was quite different from previous Mac behavior.

Another interesting piece of information I came across: While idle, the M1 chip consumes slightly more power than the previous Intel chip equivalent. While under load the M1 chip consumes considerably a lot less power.

The excessive writing to SSD is a bug/issue involving certain Intel (non-M1 native) apps running through the Rosetta2 emulation. I believe this issue did not affect a large percentage of M1 users, and has either been resolved or else is naturally resolving itself as the offending apps are being updated to M1-native versions. The M1 systems do not shorten SSD life in normal use conditions.

As for idle power draw, the tests I've seen from earlier this year cite about 7W at idle for M1 and about 19-20W at idle for Intel. I'm sure Intel's idle power draw varies for i3 vs i5 vs i7 vs i9, but to my knowledge M1 chips do not consume more power than Intel chips at idle (or at peak of course, as you note).

Finally, it should be noted that unless you're using a playback app that's doing massive upsampling, a music server is a relatively non-intensive use scenario for any modern computer, in terms of both processor demand and RAM usage/SSD caching.
 

elvisizer

Active Member
Joined
Sep 16, 2018
Messages
264
Likes
212
Thanks for clarifying that. Ok I understand now "unified memory" is talking about the single centralized memory source.

I guess one down-side to what I was talking about (further optimised and increased use of solid-state/SSD caching) is that it results in more writes to the SSD which means reduced drive longevity.
Some customers were able to confirm this with Apple's own diagnostic tools/Applications. The M1 system does a lot more writes to the SSD even when the system is barely being used (just sitting idle) which is another behaviour they noticed was quite different from previous Mac behavior.

Another interesting piece of information I came across: While idle, the M1 chip consumes slightly more power than the previous Intel chip equivalent. While under load the M1 chip consumes considerably a lot less power.
the SSD thing is not actually an issue, the amount of writes going on is normal for macOS and isn't decreasing drive longevity- this was blown up in the media due to people not understanding the write stat reporting in macOS.
the intel idle power draw thing is just wrong- where are you reading this? i would stop trusting that source heh
the i9-8950HK in my old MBP idles at around 26-30 watts.
 
Last edited:
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
As for idle power draw, the tests I've seen from earlier this year cite about 7W at idle for M1 and about 19-20W at idle for Intel. I'm sure Intel's idle power draw varies for i3 vs i5 vs i7 vs i9, but to my knowledge M1 chips do not consume more power than Intel chips at idle (or at peak of course, as you note).
Hmmm I am not quite sure of the exact numbers, but at idle power consumption on the article I read stated something like M1 chip - X watts idle, and the equivalent Intel chip was 2 or 3 watts less at idle. Sorry I cannot link the source article it was awhile ago and I cannot even remember what website.
When it came to peak/under load the M1 basically blew the Intel out of the water with much much less power consumption (something like 1/4 or less consumption).

I also think the article was referring to overall system power consumption (measured from the wall socket) and not the actual chip power consumption.
 
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
the SSD thing is not actually an issue, the amount of writes going on is normal for macOS and isn't decreasing drive longevity- this was blown up in the media due to people not understanding the write stat reporting in macOS.
Ok thanks.
I just remember reading the article, then seeing the Apple diagnostic tool stats and just confirming it to myself by what I saw on the M1 Mac Mini I just purchased and setup. To be fair this was a few days after setup so the system could have been doing a lot of background tasks even though I only use it to playback music.
the intel idle power draw thing is just wrong- where are you reading this? i would stop trusting that source heh
the i9-8950HK in my old MBP idles at around 26-30 watts.
Sorry I cannot link to the source article, it was quite awhile ago and I cannot even remember which website.
I think they were referring to overall system power consumption (measured at the wall-socket), and the idle consumption numbers were really close. The equivalent Intel chip was only consuming about 2 or 3 watts less than the M1 at idle.
When both were given heavy load demanding tasks, the M1 ended up consuming much much less power. From memory something like 1/4 or 1/5 of the Intel power consumption.
 

elvisizer

Active Member
Joined
Sep 16, 2018
Messages
264
Likes
212
Yeah it's possible there was a comparison to an older i5 or something, in the mini's those were just below the m1's idle consumption- the 2014 model used an i5 that idled at 6 watts, the m1 is 6.8
Wouldn't be the first time someone's done a comparison like that to make intel look better (for example the comparisons between the m1max and the not-yet-shipping alder lake CPUs :)). If you want to make intel's performance look better, you use a CPU that isn't shipping yet. If you want to minimize the power consumption differences, just compare it to a low power chip that's a few generations old. Easy peasy.
the macalope did a good job summing up the silliness
 
Last edited:
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
Wouldn't be the first time someone's done a comparison like that to make intel look better
Ah I see. Thanks for explaining that. Yeah I try to believe instead of "ill will" it was more a case of making the comparison with that they had available (not everyone has access to anything and everything)
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,769
Likes
8,139
Some Intel/AMD systems are most definitely better performers than M1 Macs, when those Intel/AMD systems have high-quality dedicated graphics cards.

(For the M1 pro and Max it's not yet clear, as most configurations of the Pro and Max appear to run circles around the regular M1 when it comes to multiprocessor and especially graphics-intensive performance, suggesting that these new Apple systems have reduced or even eliminated the performance advantage that has until now been held by many dedicated-graphics-based Intel/AMD setups.)

I mention this to clarify that the M1 systems are not magical and they don't outperform every competitor in every situation. (For single-processor tasks, though, the M1 is virtually impossible to beat with an Intel/AMD system unless you spend 2-3x the money.)

However, the M1 (as opposed to the M1 Pro and Max) has always explicitly been designed and marketed as Apple's entry level/lower-end chip. And it's simply a fact that it consumes far less power than comparable Intel/AMD offerings; that it provides far more performance per watt; and that as an entry-level chip it generally outperforms not only entry-level Intel and AMD chips but also mid-level Intel and AMD chips. This is confirmed by numerous benchmarks - both synthetic and real-world timed tests. It's not a matter of opinion.

IMHO, there's not exactly "ill will" or bad faith in some of these dubious 3rd-party comparisons (although Intel's attempts to market against the M1 are indeed quite shady and bad-faith). But these comparisons are in some cases rigged because (again, IMHO) many in the non-Apple computer computer community, especially those focused on performance and benchmarking, have long been suspicious of Apple's performance claims and disdainful of Apple's overall consumer-based, non-techie aesthetic and approach/language. So they have a preconception, conscious or not, that Apple must be overstating things and that Apple fans must be idealizing and worshipping the products, and therefore the reality must be less impressive than the claims. In the case of M1, though - and so far it seems like M1 Pro and Max too - the reality is pretty much just as good as Apple marketing says it is.
 
Last edited:
OP
mononoaware

mononoaware

Addicted to Fun and Learning
Joined
Apr 6, 2021
Messages
816
Likes
669
However, the M1 (as opposed to the M1 Pro and Max) has always explicitly been designed and marketed as Apple's entry level/lower-end chip.
I understood it to be the most "power efficient" chip, therefore ideal for use in their iPad Pro's (with limited battery capacity) as well as Mac's etc (made sense to me to have one of their first proper chips be "compatible" with a wide range of products.
To my understanding more processing speed usually results in more power consumption so M1 Max likely consumes more power than the M1, but I figure if they are sticking it in their battery-powered Macbooks then it should still be relatively power efficient (I have read some great things about the M1 Pro/Max Macbook battery-life).
The speed at which laptop batteries degrade has been a sore spot for laptop owners for quite awhile. I helped decide on a "Macbook" style Intel laptop for a family member and of course 1.5-2 years later the battery can hardly last 2.5-3 hours on a full charge. I think the M1 chips will bring a fresh look at laptop battery-life as I have seen reviews state they are pretty impressive.

It was a no brainer for me to go for the M1 Mac Mini, since all I wanted was power efficiency since at least for the first few years of ownership I will be using it strictly to wake in and out of sleep just to play music files. And if there comes a time my current ancient PC dies, I will start using the Mac Mini only for basic word processing and internet browsing tasks. So in my use case I like to think it was quite a good investment (I went with the 512GB model).
 
Last edited:
Top Bottom