Recently have been noticing brief dropouts in my audio that last about a half second and seem to be occurring with increasing frequenc. I might go a few minute with no dropout and then have two or three per minute in a totally random fashion.
 
I have a desktop system and laptop system. I have recently posted about some issues with using my Kontrol s61 keyboard with the laptop, but this issue is completely unrelated to the laptop or the Kontrol controller.
 
My desktop system is definitely getting old and maybe tired. I am trying to figure out if this issue is likely related to the computer or something ancillary.
 
The desktop is prob about 7 years old. Dell XPS, i7 920 processor, only 6gB, Win 10, two internal sata hd. Dual monitors w original nVidea card. The interface for this system is Roland Octa-capture. Monitors are JBL 4328 with matched 4312 sub. 
I use a Kurz PC3X and Knotrol s61. The prob is the same with either.
 
In past I would get occasional (rare) pop click when using NI sampled pianos (Akoustik, Berlin, New York, Maverick, etc). The current prob seems different. It is not a pop or click but rather a dropout. It is not a sharply defined dropout like you might get with an intermittent cable contact error. It is as if there is a slight distortion, almost like system is trying to "catch up" with the output stream. Though I first noticed it with sampled pianos, I also get the problem using Lounge Lizard and other non-sample based vsts. Prob occurs in stand alone and in Sonar (in both record and playback).
 
 
I have disconnected and disabled wireless with no effect.
I have tried to watch system monitors in real time and have not seen any cpu spikes, network issues (even before I disconnected network), or issues with hard drive access. Chkdsk says there is no problem with either disk
Roland driver is auto updated by Win10 and is current. 
 
Any thoughts? I have a scarlet 2i2 that I use with laptop. Will hook that to desktop today and assume if problem goes away the Octa might be the culprit. If it persists, it seems the culprit is either the computer or monitors (very unlikely).
 
FWIW, I installed Win10 just a couple of months ago. The problem seems like it may be temporally related to Win10 install, but I'm not sure of ghat and it seems like a pretty random Win10 issue (unless it has to do with the Roland driver)
 
Any suggestions appreciated!
 
Brian
 
 
 
 
 
 
 

Hi Brian,
 
Have you measured the machine's DPC Latency with LatencyMon?
If not, I'd recommend doing that...
High DPC Latency would cause the symptoms you're seeing.

some ideas ...
 
... could be a MIDI loop going in the background that fills up MIDI buffers and then causes hickups (check/disable MIDI connections / unplug devices for a test)
 
... try latencyMon to track down what causes the thing to start after a minute (task manager may not be enough to spot a windows service popping up briefly)

Thanks to you both - that was quick! This occurs even if the only midi is the live midi in feed to a vst so I don`t think there could be a loop situation.
 
I will run latency mon and see what happens. Despite have read about DPC latency and latency mon in innumerable threads over the years, I really don`t know what DPC latency is or what latency mon does. I assume latency mon runs in background and looks for activity that cause latency (DPC latency?) That is over and above the usual round trip latency??
I will prob need some help interpreting results. I have seen some reports (i think posted from latency mon) that look like a memory dump to me!

Is Latency Mon the software produced by Resplendence? I see that as a $120 download on their site or free download on cnet and other servers. Same thing?

There is a free version of Latencymon on the Resplendence site. They are in the downloads on the left side. Latencymon 6 has always been free. You may be looking at a pro version of it, but you only need the free version.

DPC Latency and Audio Latency are two completely separate/different things.
 
DPC stands for Deferred Procedure Call.
Tasks in Windows are scheduled via deferred procedure call.
ie: If a driver monopolizes CPU time, it causes high DPC Latency.
Many WiFi drivers are notorious for causing high DPC Latency (monopolizing the CPU).
 
At a 64-sample ASIO buffer size/44.1k, the CPU has 1.5ms to mix the next audio buffer.
If the CPU is monopolized with another task and doesn't get the next audio buffer mixed in time, at the least you'll hear a glitch... and at worst it'll cause a drop-out.
 
Thus, you need low/consistent DPC Latency in order to effectively work at low audio latency.   
 
You just need the free version of LatencyMon.

Thanks for the explanation! Work has been getting in my way this weekend. I will dl the free latencymon as soon as i get the chance.

I am running Latencymon. For first 2.5 hrs it said my system had no issues with real time audio. I was working the system using plugs that have been dropping out, and during that time I had no dropouts. Left the system on and when I came back an hour later the banner had turned red stating that system may be having trouble with real time audio and other tasks. The highest reported DPC routine execution time was 2134 microsec attrib to ataport.SYS ATAPI Driver Extension, Microsoft. 
Latencymon is recommending checking for BIOS updates and disable cpu throttling.
Curiously, Latencymon reports that my OS is Windows 8, which has never been on this machine.
 
Any insights?
 
Brian

HD controller is causing DPC Latency spikes.
Check for an updated driver.
If an updated driver doesn't resolve the issue (or one isn't available),  try changing the driver mode in the BIOS.
ie: If it's AHCI, try IDE
 
Beyond that, there's not a lot you can do...