SAC latency

[Ruben VS]

Hello, how are you? I have some questions about latency. Right now, I'm getting an input-output latency of 10:47. Actually, when I start SAC, I get a latency of 8:47, but it soon goes up.

My question is, aside from the buffer, could latency be lowered by putting the CPU into real time? Is there any way to reduce it by even 2 ms?

10:47 is at the limit for in-ear monitors. Some musicians with good hearing will notice this.

I look forward to your reply, thank you.

[Bob L]

I run all my SAC setups at 2x64 sample buffer settings which provides a very realtime experience with no complaints about latency from anybody. You should be able to achieve this if you follow my windows tweaks guides in the helpfile. Run in ASIO mode... try setting the Force Realtime priority option in the Options menu and save the default prefs after doing so.

If you can not achieve un-glitched audio playback at this buffer setting, then something else is stealing the background thunder from the cpu.

Bob L

[Ruben VS]

Hi Bob, thanks for replying. I'm using exactly this configuration. 2x64 in the Asio buffer.

The PC has all the settings in the SAC step-by-step manual. But when I measure the latency between input and output, I get this data, which I included in my first post.

I wish I could figure out what is delaying the signal, just getting an improvement between 2 and 3 milliseconds would be enough.

[mr_es335]

Ruben,

What OS?
* Version, build....
What is the hardware specs?

[Ruben VS]

My operating system is Windows 10, The PC has an Intel i5 2500 processor, it's quite a few years old but to run SAC alone it's more than enough for me.

I have 16GB of RAM, I'm using 64-bit SAC and I have an SSD hard drive.

[TomyN]

I assume the cause is not related to SAC directly. Because SAC does exactly what it reports and never changes latency. If latency jitters/jumps the problem is your I/O hardware or their driver. Keep in mind, that any I/O device also will produce latency. Some time ago I did some latency measurements using SATlive. It turned out that the old ADA8000 has more a 'roundtrip time' which is more then 1ms above the one archived using the ADA8200 (with exactly the same settings for ASIO and SAC). So take a look on your hardware.

[mr_es335]

Tomy.

Great point.

If possible, test your setup with another interface and see what occurs.

PS: I am very, very, very impressed with the UMC1820 interface. It is absolutely amazing what this unit does employing USB - with the current drivers [V5.57.0, WIN10]

[Ruben VS]

You said: If possible, test your setup with another interface and see what happens.

I say: but I don't want to buy another interface. My system is complete with 48 Yamaha AD8HR preamps, and my interface is a PCI, Digigram LX6464es. I work with a single network cable between the stage and FOH.

I even have gain control from the FOH control through the same cable.

I'm going to try configuring Windows more and putting the processor in real time to see what happens.

[TomyN]

Again: The latency of SAC is fixed. When the performance of the computer is not good enough, then you'll see droped buffers and you might hear some odd sounds. But still SAC will contiune to work with the latency set. So improving the performance of windows will not affect the latency introduced by SAC. By the way, I can control gain also over a single CAT connection :-)

[Ruben VS]

Hi TomyN, thanks for replying. Have you ever measured your system's latency with Dante? What's the delay between input and output?

The problem is that calculating the sum of all the delays applied by the AD and DA converters, Plus the interface latency, it shouldn't be more than 6ms. So I don't know where those extra 4.47ms come from.