• Computers
  • Latency Comparison: MAC vs PC! (p.2)
2015/07/07 08:19:30
dcumpian
 
 
brundlefly
thechurchboi
Buffer Size: 5.0 msec

Input: 11.4 msec, 501 samples
Output: 16.4 msec, 722 samples
Total Roundtrip: 27.7 msec, 1223 samples

 
That's quite a bit of bus and A/D/A conversion latency. Even if you had no buffering at all, your round-trip would be 27.7ms - 2 x 5ms = 17.7ms, which would be a little bothersome with input-monitored instruments, and just tolerable with soft synths that are only subject outbound latency.
 
If the latency is a lot better on your Mac with comparable buffer sizes, it would seem that the USB system and interface drivers are working better there.  SONAR is at the mercy of you Windows hardware performance.
 



Well, that and the ability of the interface driver writers to create a good quality Windows driver. Many driver work better on a Mac because there is only one way to do things. Windows is a very different environment.
 
Regards,
Dan
 
 
2015/07/07 09:03:35
Doktor Avalanche
How is this a bug?
2015/07/07 12:20:03
brundlefly
It's not, and the thread should probably be moved at this point. But it would be nice to get the comparable numbers from Logic with the same buffer size. Beyond that, I'm not sure how you would tease out the individual contributions of bus and driver performance. I'm assuming A/D/A conversion and firmware performance will  be the same for the interface, regardless of which platform it's connected to. And we know that SONAR can deliver competitive latency with the right hardware/driver setup, so it's just a matter of figuring out why the interface is "snappier" on the Mac than on Windows. But at this point, we haven't even confirmed that the buffer sizes are the same.
 
 
2015/07/07 20:12:54
Doktor Avalanche
Why the heck aren't hosts moving these posts? Have they stopped monitoring these forums? Are they even alive?
There are loads of posts around like this as well, it's not like this is an isolated thing... But this is a forum for recipes to reproduce issues and get them validated.
 
Nevermind I gave in ages ago..
2015/07/08 00:56:38
brundlefly
Yeah, I think the novelty of being forum hosts has worn off a little.
2015/07/08 07:40:31
Doktor Avalanche
Well it wore off on me.Looks like this finally got moved. I think fresh blood is called for myself. Anyways moving on...
2015/07/10 11:08:33
johnkeel
With the RME Fireface 802 i get roundtip bellow 10ms even with USB.
 
The problem in maintaining small buffers for that (like 64 buffer sizer) is having a machine that won´t spike with DPC latency, small buffers + spikes = audio crackle and even audio engine stop in extreme spikes.
 
With Apple stuff you get usually acceptable dpc values in a STOCK system (200 to 500 dpc), they have limited hardware so it´s a lot easier for apple to keep the drivers under control. With Windows sometimes a full stock system will have 4000 dpc spikes from some random hardware driver, acpi or whatever and so on...
 
The thing is, with Windows it´s VERY VERY easy to tune and get dpc values between 5 to max 100. I have max 20 dpc :D can´t get that on a mac, i tried, failed. Just be sure to get proper controllers, disable the crap you don´t need (devices, services, etc), lock CPU/Memory/Power values and do a lot of testing.
 
As for drivers, RME blows my mind, the best purchase i ever made! USB or Firewire, both rock the same. 
2015/07/11 03:15:20
robert_e_bone
While I CAN run at an ASIO Buffer Size of 64, I generally run at 128, and 48K sample rate and 24 bit record depth, and I am right around 9.7 ms to 11.3, depending on which of my audio interfaces I am using.  These numbers are for tracking/recording, by the way.  When mixing/mastering, I jack up the ASIO Buffer Size to either 1024 or 2048.  I then switch back and forth between low for tracking and high for mixing, and I never have performance/latency issues.
 
Hope that helps, 
 
Bob Bone
 
12
© 2024 APG vNext Commercial Version 5.1

Use My Existing Forum Account

Use My Social Media Account