Jump to content

MacBookAir M1 $849


Fleer

Recommended Posts

6 minutes ago, Fleer said:

I have a wife, Zo, still do. 

ahahah , more seriously , give us a feedback once you get it , i went with a lenovo for a all in one touich screen based for kids , but imac might be also next option since i need to master Logic for school (was also thinking mac mini)

Link to comment
Share on other sites

Yeah, I’ll only run Silicon native apps: 

AAS

Spire

Surge

Bitwig

Dexed 

Melda

OB-Xd

Hornet

Scaler 2

Valhalla!

Acustica

Synapse

FabFilter!

Pianoteq!

FL Studio

WaveArts

Waveform

Klevgrand

Virtual CZ

AudioThing

SugarBytes

Xferrecords

TwistedWave

WavesFactory

Spectrasonics!

Moog Model 15

Apogee Duet 3

MOTU M2 & M4

MOK Waverazor

  • Great Idea 1
Link to comment
Share on other sites

2 hours ago, Paul Young said:

i can buy a whole lotta desktop for that price, but that eliminates the chance of ever using Logic or Alchemy2 .

Why do I all of a sudden have Zepplin running through my head right now?

Edited by Doug Rintoul
Link to comment
Share on other sites

here are some results of a MacBook Air.  8gb ram.  512gb ssd.  This is just a preliminary test,  but I'm sure  of interest to many, especially since they are on sale so often lately. I took screenshots, as pics or it didn't happen...so for this test, I only am using Kontakt and Logic Pro . logic is running native and Kontakt Rosetta.  9 instances of Noire. 9 of Session Strings 2, and 9 of Grandure.  27 tracks total.  buffer set to 32 and at 44.1.  I looped a 20 bar pattern consisting of sustained notes, as well as short staccato notes.   I attached a pic of the piano roll to demonstrate the notes being looped.  I looped it for 30 minutes.  And I had the project open for over 2 hours.  no drop outs or glitches.  I was also downloading BFD while I was doing this.    The Air did get warm.  I don't have a way to measure the temp,  but it felt on the warmer side.  Maybe like luke warm water.  but not hot. nowhere near as hot  as my prior 2018 15inch MB pro with two fans.  Keep in mind that this is not a real world scenario.  As why have all these instances of the same thing.   Its just a stress test.  And also note that this is with the internal Mac audio.  Im  sure my results will be better when I run the test through my UAD arrow instead.  And I was running this test  off of the battery.  I forgot I wasn't plugged in.  The word amazing comes to mind,  but I'll save the impressions and just present the results.   

Screen Shot 2021-06-23 at 10.22.03 PM.png

Screen Shot 2021-06-23 at 10.23.15 PM.png

Edited by Hugh Mann
Link to comment
Share on other sites

That is pretty impressive for sure. One thing you might want to look at though is the difference between loading nine instances of the same instrument and loading nine different instruments. That is what a real world scenario would look like. There may be sample sharing between the Kontakt instances. I would also be interested to see how it performs with CPU heavy vsts such as anything from Acustica, Diversion with full unison, or Fathom. 

Link to comment
Share on other sites

Here is another version of the same test. This time, as Doug suggested,  I loaded different instruments.  I only have the following installed right now.  But here is a pic of that.  This is with 1 OTS Strawberry, 1 OTS Infinity, 1 Analog Dreams, 1 Hybrid Keys, 1 Butch Big Drums.  And the rest are Noire, Session strings 2, and Grander (as indicated on the track names). So 27 different Kontakt  instances.  Each track playing back 8 notes max,  sustained and staccato.  I also opened up 4 Safari tabs, and I was installing BFD while running the test.  I don't have any Acoustica plug ins.  I have Plug in Alliance,  U-he, Korg, and a few others I haven't installed yet.  More tests to come. 507298107_ScreenShot2021-06-24at6_56_59AM.thumb.png.f867f7574f9d52bfdac99c38d2e27dee.png 

Link to comment
Share on other sites

8 hours ago, Hugh Mann said:

 As why have all these instances of the same thing.

I specifically mentioned this when talking about running stress-tests on a "RAM limited" machine.

If you load many instances of the same library, it completely invalidates the test.

Why?  Because the single library is only buffered once (in RAM).

RAM wise, you may as well only be running a single instance of The Grandeur.

To actually stress-test the machine's limited amount of RAM, you need to load separate/different large libraries.

Edited by Jim Roseberry
  • Like 1
  • Great Idea 1
Link to comment
Share on other sites

1 hour ago, Jim Roseberry said:

I specifically mentioned this when talking about running stress-tests on a "RAM limited" machine.

If you load many instances of the same library, it completely invalidates the test.

Why?  Because the single library is only buffered once (in RAM).

RAM wise, you may as well only be running a single instance of The Grandeur.

To actually stress-test the machine's limited amount of RAM, you need to load separate/different large libraries.

I see you r point,  but you aren't exactly right on that.  One instance of the grandeure, noire, and Session Strings 2 takes up way less ram on this  machine than  9 of each (27 total).  In fact,  27 was the limit before I got drop outs.  So adding more of the same does indeed stress the ram,  after enough are loaded.  Not like you said "RAM wise, you may as well only be running a single instance of The Grandeur.".  I that was the case,  I could load 50 grandures with no problem.  Keep  in mind I'm playing back the 27 separate tracks with midi data.  not just loading it.  I didn't try just loading it,  cause what's the point in that.  Also,  keep in mind its pulling 8 notes from 27 different instances.  So 27 x 8 = 216 simultaneous notes (if my math is wrong). Hardly an invalid test.  I think it shows that even with "just" 8gb ram,  one can do full productions.  This was at 32 buffer setting and on the internal Mac audio.  With 4 safari tabs open (one YouTube video) and also simultaneously installing BFD.  Thats impressive.   But more tests to come.  What Kontakt and sample libraries do you consider sufficient for this type of test?  I only have (as far as large libraries) Kontakt , the stuff in Logic,  ez drummer, and I'm installing BFD drums.  If anyone want s to donate software for me to test,  my DM is always open.  :)

Edited by Hugh Mann
Link to comment
Share on other sites

Yes, the DAW has to allocate resources for many additional I/O channels, etc.

But it doesn't consume the same amount of RAM as using separate/unique sample libraries.

 

Let's say your lean OS install idles at ~3.1GB RAM

Fire up Studio One Pro v5 with an empty project... and you're at 4.3GB

Add a single instance of Kontakt 6 (empty - no libraries loaded)... and you're at 4.5GB

Load The Grandeur into the single Kontakt 6 instance... and you're at 4.8GB

Load a second instance of Kontakt 6 with an instance of The Grandeur.

You've now got two instances of Kontakt 6... each with The Grandeur loaded... RAM use is now 4.9GB.

 

If you load the same two instances of Kontakt 6... load The Grandeur in the first... and The Maverick in the second... RAM use is 5.2GB.

Both piano libraries are a bit over 7GB in total size.

 

 

I'll liken the situation to testing drive speed.

Let's say the drive you want to benchmark has 128MB cache.

  • What happens if you benchmark with a 32MB file?
  • What happens if you benchmark with a 1GB file?

With the 32MB file, you'll be testing the speed of the cache.   😉

With the 1GB file, you'll circumvent the caching... and get the actual speed of the drive.

 

 

Load up 20-30 libraries that are similar in size to The Grandeur (but each a unique separate library).

Report back with the RAM use and RAM Pressure graph.

I pushed the 16GB M1 Mini ridiculously hard... and it did a pretty remarkable job.

The RAM caching, compression, etc allow you to squeeze a lot out of 16GB RAM.

That said, if you really push it... the RAM Pressure went thru the roof... and it took a LONG time for this type of project to fully load, RAM to be compressed/cached/etc.

It's not ideal in a heavy load workstation type scenario... and (to me) not a real substitute for having enough physical RAM for your largest projects.

To be fair, Apple doesn't market the machine for that type of use.

For what it is (small form-factor, inexpensive, silent), it's a nice machine.

  • Like 2
Link to comment
Share on other sites

Here is a different test.  This time in Studio One 5.  I made about a 9 bar midi loop.  8 notes.  sustained and staccato.  I loaded 8 instances of Kontakt, each with a different library. They are listed on the track names in the screenshot.  I also loaded a BFD preset and let it play back a midi drum beat.  I also loaded ezdrummer.  I then loaded 11  different VST instruments.  Also listed on the track names.  That brought me to 21 tracks.  All of this played back fine.  So I copied the 11 instruments again.  This is more of a cpu test with ram also being pushed.  This brought me to 32 tracks.  Played back fine.  So I copied the 11 instruments a couple more times.  This brought me to 57 tracks.  This played back fine.  So I copied the entire 57 tracks.  this brought me to 114 tracks.  

Now all of this is already an unrealistic stress test.  But,  we want to see the limits of what an 8gb M1 can do.  This time,  it took a bit longer to load.  I could see the swap file increasing in size, so that would account for the time it took to copy all of this.   Interestingly,  as you can see from the screenshot,  the ram stayed at around 7.5 gb.  But the swap file kept increasing to 9gb.  This shows what I was saying before of the ssd being used as a sort of ram.  Not faster than actual ram,  of course,  but it works.  The entire project played back fine.  no drop outs.

 I could probably keep going,  but at this point its just academic.   Clearly 8gb is enough for even serious,  full on productions.  Now,  predictably,  some will poo poo this and say this  is only for "lightweight" or "amateur productions".  But that is ,  of course, silly.  Many hit albums and songs have been produced and are being produced now with way less going on than what is on the little test I just concocted.

 The question I had was if an 8gbM1 Air (with no fan!)  could be used for serious productions.  The answer,  as these tests demonstrate, is yes (and then some).  I think this machine can be used in several scenarios.  professionally.  Here are some of the ways I will be using it:

As a host for BFD/Ez drummer.  I play e-drums in live settings and I want to trigger BFD /ezdrummer with as low a latency as possible.  The M1 Air gives me a super powerhouse of a brain for my drum libraries.  Its a small,  lightweight package that is just perfect for bringing live.  With my UAD arrow and this M1,  I can play BFD from e drums at under 2ms latency.  This is no problem at all for the Air.  

Another scenario is as a low latency guitar fx.  It doesn't have to be just guitar.  Any instrument or vocal.  But,  due to the power and efficiency of the m1 air,  I can bring just the laptop and a UAD arrow and have all the fx I would want to use live.  At under 2ms latency.  and at 96khz.  While also sequencing program changes.  

Another scenario is as an Ableton/keys station.  Again,  in this scenario I could go real far without the M1 Air even acting like it cares, lol.  

And another way I will use this is as a mobile recording station.  I can record the audio from the guitar and vocals (arrow only has 2 inputs),  while recording the midi from my e drums and the keyboards.  

Another scenario is as a mixing, production work station.  And again,  this machine will be able to handle it all.  From creation to completion.  

All this is no problem at all for this machine. nowhere near pushing it to do these things.  I had to create some wild and unrealistic stress tests to make it break a sweat. Stuff that is unrealistic in actual,  real world use.  This M1 (any version) is a very useful tool.  No fanboys crap.  Its just a great tool.  And if one wants to save a bit of change,  the 8gb will do very well.  Im sticking with the 8gb cause its doing everything I want to do or can even forsee myself wanting to do.   I admit, it feels weird to have "just" 8gb of ram.  But its handling everything I am throwing at it.  Its juts not X86 architecture.  So the old thinking about ram doesn't apply.  

 

 

1885639159_ScreenShot2021-06-24at10_25_20PM.thumb.png.9eae10a0552294e30eaf73e60993cd7a.png

Edited by Hugh Mann
  • Thanks 1
Link to comment
Share on other sites

In order to answer a few more questions about the M1 with just 8gb of ram,  I ran some more tests.  For this one,  I loaded 20 instances of Kontakt and made sure to load a unique library for each instance.  I created an 8 bar loop playing back 8 notes.  Sustained and staccato.  UAD Arrow set to 96khz and 32 samples. This time,  I wasn't downloading or installing other programs while I ran the tests.  I did have one safari tab open.  And I ran the test off of the battery.  I really love the battery life on this thing.  Its really great to be able to work on real projects,  with no cord.  Truly a portable, powerhouse, workstation.  

The ram did the same as in the other tests.  peaked at around 7.5gb.  Interestingly, it didn't use as large of swap file.  The laptop was not struggling at all.  I ran out of libraries to load ( I only want the ones I installed and used for this test). It all played back fine.  No glitches or drops outs.  My conclusion,  and what a lot of reviewers (that have actually tested the 8gb model ) are also concluding,  is that 8gb on an M1 is not the same as 8gb on an x86 platform.  Apple calls it "unified  memory".  Here is a link to an article that goes into more technical detail. https://www.howtogeek.com/701804/how-unified-memory-speeds-up-apples-m1-arm-macs/

Bottom line...paying apple the high price for the 16gb ram,  may not be worth it,  as it won't necessarily improve the performance.  I've found that to be the case in my tests, so I'm glad I tried it for myself, instead of just assuming 8gb would never be enough.   I could probably keep pushing the machine and get even more impressive results.  But, I'm more interested in real world scenarios vs just pushing it to see its limits. This test gives a really good indication of what can be achieved  with "just" 8gb of ram and an M1  chip.  The libraries I used are: 

Session Strings

 Kontakt Factory Grand

India Sitar

India Ensemble 

 Butch  Vig  Drums

VLS Strings Cello

Noire Pure

Noire Felt

Strummed Acoustic 2 6string

Strum Acoustic 2 12 string

Cloud Supply 

Analog Dreams 2 

Ethereal Earth 2 

Hybrid Keys 2

OTS Evolution

OTS Infinity  

Grandure

 Scarbee Mark 1 

Scarbee Clavinet

 Scarbee A 200.

 

Screen Shot 2021-06-25 at 11.59.09 PM.png

Edited by Hugh Mann
  • Thanks 1
Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...