The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
Anyone interested in my new benchmark to replace the dated Evan's test?
Old 5th September 2018
  #31
Gear Nut
 

Quote:
Originally Posted by TNM View Post
Logic reserves the last core for live VI's and rewire.
I see. Thanks.
Old 18th September 2018
  #32
Im running this on my iMac Pro base config (8 core, 32GB, 1TB) and I can run all the tracks at about 75% cpu at 44.1.

When I switch to 96khz I could run 74 tracks reliably, and sometimes 75 would work but it dropped out after a few bars.

There is one thing this test taught me- that I should start recording at 24/96 just for latency reasons. I have a UAD Apollo 8, and it does indeed reliably run at 1.5ms latency at 96khz with a buffer of 128 samples...

Thanks for putting this together! I will run it on my i9 MacBook Pro at some point but I don't have it really set up for audio yet.
Attached Thumbnails
Anyone interested in my new benchmark to replace the dated Evan's test?-benchmark-96khz.jpg   Anyone interested in my new benchmark to replace the dated Evan's test?-benchmark1-90.jpg   Anyone interested in my new benchmark to replace the dated Evan's test?-benchmark41-130.jpg  
Old 26th September 2018
  #33
Here for the gear
 

Hello All,

Ran the test on both my machines:

MBP Late 2013 i7 2Ghz 8gb Ram - 30 Tracks
MacPro (Trash Can) Quad 3.7ghz 32GB Ram - 49 Tracks

Pretty cool comparison but I'm a newb so I have to ask. I see some posts in here that they are running 100+ tracks with no issues. In the real world, whats a typical project contain and do we have a consensus on what is a good #? (i.e. what is a # where we wont throw our computer out the window while working) I ask bc some of us may be looking to throw the the trash can in the trash and replace with another solution or just upgrade in general. This test most certainly helps get some real world benchmarking. Thanks all!
Old 26th September 2018
  #34
A great record can be recorded with 4-8 tracks. More is always better, but if you can't make something great with 24 tracks it's not the computers fault. The DAW should just get out of the way though, so having the headroom to record 100 tracks with efx just means you don't need to think about it.
Old 26th September 2018
  #35
Lives for gear
Quote:
Originally Posted by S_A_P View Post
A great record can be recorded with 4-8 tracks. More is always better, but if you can't make something great with 24 tracks it's not the computers fault. The DAW should just get out of the way though, so having the headroom to record 100 tracks with efx just means you don't need to think about it.
except film orchestrations of course.. many of those projects can even run up to 500 VI/midi tracks and I am not joking.
Old 27th September 2018
  #36
Quote:
Originally Posted by TNM View Post
except film orchestrations of course.. many of those projects can even run up to 500 VI/midi tracks and I am not joking.
Are you sure about that? I suppose that could be true, and that some people have made records that big, but I would like to know some specific examples. Doesn't Logic ignore any tracks over 256? Protools was a 128 track max until recently, and even now you could only do 512 in VIs. I would love to see some examples, not saying it never happened, but studios recording large orchestras don't individually mic/record each instrument.
Old 27th September 2018
  #37
Quote:
Originally Posted by S_A_P View Post
Are you sure about that? I suppose that could be true, and that some people have made records that big, but I would like to know some specific examples. Doesn't Logic ignore any tracks over 256? Protools was a 128 track max until recently, and even now you could only do 512 in VIs. I would love to see some examples, not saying it never happened, but studios recording large orchestras don't individually mic/record each instrument.

The context is more DAW based composition, rather than live recording. I'm not hugely up to date with it, but even just a few years ago film composers used multiple computer systems to score and compose using high end sample libraries like those by Vienna Pro. Literally on DAW running just extended brass virtual instruments, another running just strings, another just woodwind, etc. People used to use midi over ethernet, but the process is a lot more streamlined now thanks to apps like VE Pro... all controlled by a central composition/scoring system.

It wouldn't surprise me if people were folding down their multiple sampling computers into smaller numbers thanks to the availability of faster computing. Hans Zimmer for example has routinely mentioned his custom 'sampler'. It would be interesting to know what is means by that. I'm sure it is a series of powerful computers, dedicated to playing certain sounds, all networked together. Those details may well be contained in this article - I haven't read the whole thing... but it does give an insight into how some composers for video work. Hans Zimmer | Steinberg
Old 27th September 2018
  #38
Lives for gear
 

As far as track counts, composers can use key switching for different articulations using the same track. Other composers may prefer to organize the articulations using seperate tracks. Both methods work, although key switching, of course, will result in a lower track count and probably less work for the CPU(s).

Last edited by fastlanephil; 30th September 2018 at 03:11 AM..
Old 30th September 2018
  #39
Lives for gear
 
Fernand's Avatar
It's relevant that both 4,1 -> 5,1 Classic Towers with 2 x 3.46 Ghz Xeons
(mine and @ fastlanephil 's) give exactly the same 116 tracks. I don't
expect the 12 core plastic trashcan 2013 mac pros to do much better.

I expected more from my i7 4790K 4Ghz hack, I got 62 with a lot of
other apps running, I'll have to try it without all that overhead ;-)
But although it absorbs spikes better, and it doesn't waste as much
wattage, it doesn't handle big multitrack projects as well as the old
faithful aluminum Classics.

This benchmark is not checking the "single core spiking" where the fastest
clock speeds shine. What we're getting here is a measure of how well the
rig handles big multitrack sessions.

It's hard to get both on one machine. On the new macBook Pros, if you
run big projects on 6 cores, it's bound to slow down the clock to prevent
overheating.

People who come from 4/4 rock are perfectly content with smaller
systems, and they keep saying that anyone who isn't satisfied
making music on a single laptop with a 16 GB RAM limit doesn't know
what he's doing ;-) But big projects aren't done that way, it would be
silly, just the sync and multi-user requirements alone call for more
than a single machine.

Last edited by Fernand; 30th September 2018 at 02:40 AM..
Old 1st October 2018
  #40
Gear Nut
 
olofd's Avatar
Hackintosh 7900x. 48 gb ram. 4.4 Ghz.
All 128 tracks looping and playing well with a 1650%-cpu load. (2000% max, 20 cores) at 96khz
Old 4th October 2018
  #41
Quote:
Originally Posted by Turbovdub View Post
Hello All,

Ran the test on both my machines:

MBP Late 2013 i7 2Ghz 8gb Ram - 30 Tracks
MacPro (Trash Can) Quad 3.7ghz 32GB Ram - 49 Tracks

Pretty cool comparison but I'm a newb so I have to ask. I see some posts in here that they are running 100+ tracks with no issues. In the real world, whats a typical project contain and do we have a consensus on what is a good #? (i.e. what is a # where we wont throw our computer out the window while working) I ask bc some of us may be looking to throw the the trash can in the trash and replace with another solution or just upgrade in general. This test most certainly helps get some real world benchmarking. Thanks all!
Jacob Collier did projects with a Mac probably not as fast as mine (30 tracks- see below) yet he got hundreds of simple tracks, then started freezing.

People made great albums with way slower and limited machines. There are limits, and then you start compromising slightly ( freezing, bouncing, stems etc) wich can even help commit.

No one can judge what is good enough for someone else, it's also a matter of comfort. I guess Collier now has a MacPro with way less limits ( virtual instruments all over, track count, plugins, etc etc) but his old macbook never stopped him from being a genius, praised by Quincy and the whole world. gigantic high number of tracks and automation all produced in his small room.

Last edited by DownSideUp; 4th October 2018 at 06:53 PM..
Old 4th October 2018
  #42
Macbook Pro late 2013 quad I7 @ 2,6
SSD 500go
16go Ram
128 buffer

High Sierra , Logic 10.3

I'm at around 28 tracks. ( earlier poster with the EXACT same machine has 15 tracks? maybe an older Logic or OS?)


Mac Mini 2012 quad I7 at 2,3[/B] 16Go rame 500 SSD

43 tracks[/B] ! what a beast, on par with some 2013 MacPro.

I'm pretty sure temperature inside the macbook, gets the count down. other than that, they are almost the same machines. (Mac mini even has older OS and Logic). My Macbook should even have a faster hard drive bay.

Anyway I'm quite surprise with the difference in performance.

Last edited by DownSideUp; 4th October 2018 at 07:29 PM..
Old 4th October 2018
  #43
and yes I did 44,1, 128 buffer / medium size buffer. integrated sound card . this should be mentioned as a guideline, for a fair comparison. Along with a proper title and short first post (short intro, guideline and link), then it's off to collect data fast

thanks for the work.

Last edited by DownSideUp; 4th October 2018 at 07:31 PM..
Old 8th October 2018
  #44
Gear Head
 

33 Tracks
Attached Thumbnails
Anyone interested in my new benchmark to replace the dated Evan's test?-screenshot_2018-10-09-04.57.48_lvlvaj.jpg  
Old 8th October 2018
  #45
Gear Head
 

Quote:
Originally Posted by S_A_P View Post
Im running this on my iMac Pro base config (8 core, 32GB, 1TB) and I can run all the tracks at about 75% cpu at 44.1.

When I switch to 96khz I could run 74 tracks reliably, and sometimes 75 would work but it dropped out after a few bars.

There is one thing this test taught me- that I should start recording at 24/96 just for latency reasons. I have a UAD Apollo 8, and it does indeed reliably run at 1.5ms latency at 96khz with a buffer of 128 samples...

Thanks for putting this together! I will run it on my i9 MacBook Pro at some point but I don't have it really set up for audio yet.
Can you elaborate on this 96k latency thing? - I don't quite understand what you mean - cheers
Old 10th October 2018
  #46
Quote:
Originally Posted by laughingboy9000 View Post
Can you elaborate on this 96k latency thing? - I don't quite understand what you mean - cheers
Since the latency of the audio interface is in samples, the higher the sample rate the lower the latency for a given sample rate...
Old 11th October 2018
  #47
Gear Head
 

Quote:
Originally Posted by S_A_P View Post
Since the latency of the audio interface is in samples, the higher the sample rate the lower the latency for a given sample rate...
So a few more question if that's cool as I'm about to pull the trigger on the same iMac Pro you got:

1) In your opinion, in ms, what is an unacceptable latency for recording vocals with effects and what is an imperceptible latency?

2) If you were recording vocals with the same sample and buffer rate, will the two macs with different specs give different latency?

3) With your iMacPro, on a scale of 1-10 how awesome is it as far as latency and recording goes?

4) If you are the type of dude that regularly exceeds 100+ tracks and can't set the sample rate to 96, is it impossible to get imperceptible latency at 44.1 or 48 without engaging LL mode?

Thanks in advance
Old 11th October 2018
  #48
Quote:
Originally Posted by laughingboy9000 View Post
So a few more question if that's cool as I'm about to pull the trigger on the same iMac Pro you got:

1) In your opinion, in ms, what is an unacceptable latency for recording vocals with effects and what is an imperceptible latency?

2) If you were recording vocals with the same sample and buffer rate, will the two macs with different specs give different latency?

3) With your iMacPro, on a scale of 1-10 how awesome is it as far as latency and recording goes?

4) If you are the type of dude that regularly exceeds 100+ tracks and can't set the sample rate to 96, is it impossible to get imperceptible latency at 44.1 or 48 without engaging LL mode?

Thanks in advance
1- 100% depends on what you are doing. There are times that I’ve had 1000ms of plugin delay running while tracking and it was fine. However if I am tracking the initial groove and the part is intricate or otherwise busy I want the latency as low as possible. Getting the right feel is definitely affected by latency. However in my experience I can just adjust after things are built up and I can vibe with the track even if latency is high. So general rule for me is acoustic/electric guitar, live or finger drumming need lower latency especially when I don’t want to quantize or there is a shuffle. I just don’t like logic/reason/flstudios shuffle. I prefer to get the feel in a live mode.

2- if you had 2 macs, and one was a 2 core and one was an 8 core you can get the ~ same latency up to a point with the same interface hardware. The bigger machine will let you run lower latency longer, and it’s possible that if you want 1-3 ms latency the 8core box will always win. However as I alluded to in point 1, it’s more critical to me to get the feel right, and then I can back off the latency as the track builds or the cpu runs out of headroom.
3- it’s definitely the most powerful computer I’ve owned. I can get 1.5 ms latency with my Apollo 8 at 96 khz. I normally run it at 44.1 and get ~10-12 ms I think. It’s not something I think about really as the computer just gets out of the way. Honestly that is what I really wanted. I don’t want to have cpu anxiety, I just want to record/produce/mix. I had the 2008 Mac Pro octo core and felt the same about it. It gets out of the way.
4- I may have gotten to 50-100 tracks but the music I make or have recorded is way sparser than that question implies. I really think that the 96khz recording just helped me get the right feeling down at the early stages without having to time the latency or think about it.
Old 19th October 2018
  #49
Here for the gear
 

Thank you for the updated benchmark -

For any of those curious - Heres my results with the 2018 MBP i9 with 32GB RAM 1TB internal SSD and maxed out video card -


71 tracks @ 44.1 128 Medium Buffer
22 tracks @ 96k 128 Medium Buffer - Clean Playback
Frequency was clocking 3.3-3.6 Ghz occasional 3.8ghz.

I was hitting the wall on a session I was producing and mixing yesterday on this machine but after the benchmark it does seem to be right in line.

The session was 28 tracks @ 96k with 3 Vi's - a logic stock drum kit, the logic b3 and 1 instance of Kontakt action strikes (shaker)

Full automation written across most all of the tracks, LOTS OF UAD across the tracks from Lexicon, Manley, and the Studer. A few LA-2a's. It did also have quite a few tracks with flex time enabled (bass, a few of the guitar tracks, etc) and 3 sends with verb - UAD Lexicon, logic space designer, and sound toys little plate. Nothing on the stereo bus or the separate send mix bus.


I was disappointed to be hitting the wall at that low of track count on a $4k+ brand new Mac but I do have to remember:

1. its 96k its obviously going to be a hog.
2. there are many things I wasn't doing to help alleviate the strain- bouncing in place my flextimed tracks, utilizing track freezing, etc.
3. While I really would love the performance of the iMac Pro (I just upgraded from a rock solid iMac for my studio computer) - I cant stand the 27" display your stuck with, and the size of it is near impossible to take with you anywhere. One thing to love about the Mac Pro 2013, you can take it with you around the world to studios youll be working out of for an extended period of time!

Im hoping the new Mac Pro is similar in that it is either modular or relatively easy to ship/transport. One thing I DO really love about the 2018 mbp.

Im looking forward to putting the benchmark through its pace on other machines.
Old 20th October 2018
  #50
Well @ 44khz, that session would hit almost 100 tracks? that machine is quite a beast.
Tchad Blake converts all sessions down to 48hz anyway, I 've decided to do the same.

To lose two third of power, using 96khz: this change should be at least two third better sounding and obvious, and that's not. That's just my humble view.

I guess you can also export the Logic instruments as audio file, and let the machine focus on mixing?

I agree it's not a 18 core, but for a mobile machine @ 48k hz it seems quite a monster !
Old 20th October 2018
  #51
Gear Head
 

Quote:
Originally Posted by S_A_P View Post
1- 100% depends on what you are doing. There are times that I’ve had 1000ms of plugin delay running while tracking and it was fine. However if I am tracking the initial groove and the part is intricate or otherwise busy I want the latency as low as possible. Getting the right feel is definitely affected by latency. However in my experience I can just adjust after things are built up and I can vibe with the track even if latency is high. So general rule for me is acoustic/electric guitar, live or finger drumming need lower latency especially when I don’t want to quantize or there is a shuffle. I just don’t like logic/reason/flstudios shuffle. I prefer to get the feel in a live mode.

2- if you had 2 macs, and one was a 2 core and one was an 8 core you can get the ~ same latency up to a point with the same interface hardware. The bigger machine will let you run lower latency longer, and it’s possible that if you want 1-3 ms latency the 8core box will always win. However as I alluded to in point 1, it’s more critical to me to get the feel right, and then I can back off the latency as the track builds or the cpu runs out of headroom.
3- it’s definitely the most powerful computer I’ve owned. I can get 1.5 ms latency with my Apollo 8 at 96 khz. I normally run it at 44.1 and get ~10-12 ms I think. It’s not something I think about really as the computer just gets out of the way. Honestly that is what I really wanted. I don’t want to have cpu anxiety, I just want to record/produce/mix. I had the 2008 Mac Pro octo core and felt the same about it. It gets out of the way.
4- I may have gotten to 50-100 tracks but the music I make or have recorded is way sparser than that question implies. I really think that the 96khz recording just helped me get the right feeling down at the early stages without having to time the latency or think about it.

Really appreciate the detailed response. Will help me towards making the right decision. Kind regards.
Old 20th October 2018
  #52
Gear Maniac
Hi nick. Thank u very much for taking the time to post the results

Am I correct to assume that ur results are based without any interface connected? If u have the time can u please redo the test with a interface and post ur results?

I did the 96 test with a Motu 1248 interface and I could play only 1 track without glitches, so ur result is 22 times better. Which is kinda very impressive to me hehehe

Thanks and congratulations for ur system


Quote:
Originally Posted by Nick.margiotta View Post
Thank you for the updated benchmark -

For any of those curious - Heres my results with the 2018 MBP i9 with 32GB RAM 1TB internal SSD and maxed out video card -


71 tracks @ 44.1 128 Medium Buffer
22 tracks @ 96k 128 Medium Buffer - Clean Playback
Frequency was clocking 3.3-3.6 Ghz occasional 3.8ghz.

I was hitting the wall on a session I was producing and mixing yesterday on this machine but after the benchmark it does seem to be right in line.

The session was 28 tracks @ 96k with 3 Vi's - a logic stock drum kit, the logic b3 and 1 instance of Kontakt action strikes (shaker)

Full automation written across most all of the tracks, LOTS OF UAD across the tracks from Lexicon, Manley, and the Studer. A few LA-2a's. It did also have quite a few tracks with flex time enabled (bass, a few of the guitar tracks, etc) and 3 sends with verb - UAD Lexicon, logic space designer, and sound toys little plate. Nothing on the stereo bus or the separate send mix bus.


I was disappointed to be hitting the wall at that low of track count on a $4k+ brand new Mac but I do have to remember:

1. its 96k its obviously going to be a hog.
2. there are many things I wasn't doing to help alleviate the strain- bouncing in place my flextimed tracks, utilizing track freezing, etc.
3. While I really would love the performance of the iMac Pro (I just upgraded from a rock solid iMac for my studio computer) - I cant stand the 27" display your stuck with, and the size of it is near impossible to take with you anywhere. One thing to love about the Mac Pro 2013, you can take it with you around the world to studios youll be working out of for an extended period of time!

Im hoping the new Mac Pro is similar in that it is either modular or relatively easy to ship/transport. One thing I DO really love about the 2018 mbp.

Im looking forward to putting the benchmark through its pace on other machines.
Old 23rd October 2018
  #53
Lives for gear
 

2018 2.2 i7 6 core MBP 16GB, 1TB SSD Mojave 10.14 - 76 tracks as downloaded. Fan makes a noise at that level
Old 23rd October 2018
  #54
Gear Head
 

Quote:
Originally Posted by dr_Jezz View Post
2018 2.2 i7 6 core MBP 16GB, 1TB SSD Mojave 10.14 - 76 tracks as downloaded. Fan makes a noise at that level
Buffer?

24bit?

44.1, 48, or 96?

Either way, that's pretty decent - love the portability of laptops but thinking of going desktop as my 2013 Quad i7 MBP 16GB ram 2GB Video is choking at about 50tracks with lots of plugins - I need 100 tracks with lots of plugins
Old 23rd October 2018
  #55
Lives for gear
 

Quote:
Originally Posted by laughingboy9000 View Post
Buffer?

24bit?

44.1, 48, or 96?

Either way, that's pretty decent - love the portability of laptops but thinking of going desktop as my 2013 Quad i7 MBP 16GB ram 2GB Video is choking at about 50tracks with lots of plugins - I need 100 tracks with lots of plugins
As downloaded and straight out of the box (see front of thread) - 44, 128

Last edited by dr_Jezz; 25th October 2018 at 07:24 PM.. Reason: got rid of the thumbs up - people don't read and don't respond ;)
Old 24th October 2018
  #56
Lives for gear
 
syntonica's Avatar
MacBook Pro (Retina, 13-inch, Mid 2014) 2.6Ghz, 8GB RAM, i5 500GB SSD... 14 tracks... *cry* ... I can almost squeeze 16 with a larger buffer... If there were only a button I could press to pre-render the audio. Even on a per-track basis!

Edit: The best improvement came from using the Large buffer. Changing the latency up or down had little to no effect.
Old 25th October 2018
  #57
Lives for gear
 

Quote:
Originally Posted by syntonica View Post
MacBook Pro (Retina, 13-inch, Mid 2014) 2.6Ghz, 8GB RAM, i5 500GB SSD... 14 tracks... *cry* ... I can almost squeeze 16 with a larger buffer... If there were only a button I could press to pre-render the audio. Even on a per-track basis!

Edit: The best improvement came from using the Large buffer. Changing the latency up or down had little to no effect.
Don't worry Syntonica, I was making do with my 2008 MBP prior. Up until a few years back that still did me for what I was doing in terms of composing (I woodshedded for a while) - it's still a useable machine live, just not for the next projects. This MBP will see me for years yet (hence the 1TB drive). That's the only reason I was willing to spring for this obscene price as a non-professional. I now have at least 6 Macs sitting around from my plus through 3 laptops. This had better last ... and hopefully without all the repairs I needed for the last one
Old 25th October 2018
  #58
Lives for gear
 
syntonica's Avatar
Quote:
Originally Posted by dr_Jezz View Post
Don't worry Syntonica, I was making do with my 2008 MBP prior. Up until a few years back that still did me for what I was doing in terms of composing (I woodshedded for a while) - it's still a useable machine live, just not for the next projects. This MBP will see me for years yet (hence the 1TB drive). That's the only reason I was willing to spring for this obscene price as a non-professional. I now have at least 6 Macs sitting around from my plus through 3 laptops. This had better last ... and hopefully without all the repairs I needed for the last one
It actually does me just fine EXCEPT if I whip out an analog VST (Diva, Oxium, et.al.) They all tend to eat up 20-50% of my CPU, depending on the patch. For some reason, bell sounds really spike the CPU, regardless of VST. The FM implementation, maybe?

Unfortunately, I can't afford the latest and greatest and I won't buy another Mac until I can get a significant increase in CPU power for under $500. Which will happen the 2nd of Never!
Old 3rd November 2018
  #59
Lives for gear
 
Krubbadoo's Avatar
 

Macpro 5,1 3.46 12 cores

As some other folks here I get a reliable 116 tracks @ 128 buffer 44.1khz on my 5,1 Macpro.
I did the test again but this time with my old metric halo uln-n fw400 device @ same buffer and sample rate and could get a reliable 110 tracks. I thought it would perform worse than that! Both tests were done with the last track being record armed. Really anxious to see how the 2018 Mini 3.2 will perform on this test.
Old 8th November 2018
  #60
Lives for gear
 
Chevron's Avatar
 

2010 Mac Pro 5,1 / 2.93 ghz 12 core with 28 gb RAM

@ 44.1k I can run 94 tracks and loop, but any more and it stalls. The buffer size didn't seem to make much difference

@ 96k I can run no more than 34 tracks, and even then it is a bit sketchy

Nice test!

I am also keen to find out what a 3.2 ghz Mac Mini can do

Last edited by Chevron; 8th November 2018 at 09:28 PM..
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump