Hi all-- I’ve been banging my head on the table for a while on this, but I’m hoping someone might have some suggestions. I’m running C4 on a NUC (i5, 1.6 gHz, 16 GB RAM, SSD). I have consistently high time load with a number of plugins, which leads to pretty consistent audio glitches. I have gone through Brad’s book and done all of the steps in there, and LatencyMon shows that wdf01000.sys is the main culprit. CPU load says pretty load, but it’s just the time load that goes through the roof.
As an example, with C4 running but not playing any notes, PianoTeq alone puts me at about 15%. Turning on VB3-II puts me at 30%. Any third plugin (Velvet, Xpand, Kontakt, EQ, others) puts me at 60%. Analog Lab, Amplitube or Tracks alone will put me at 100%, way over if I have another instrument running. Actually actively playing with any of these instruments (just one or two-- not all) in real-world means I’m living in the 75-150% range.
Since WDF01000 is getting flagged by LatencyMon, it looks like it’s a driver issue, but I tried to update all my drivers to no avail. Everything I can find about wdf01000 is just “keep trying to fix your drivers!” Any other suggestions out there?
I think that wdf01000 is getting flagged but that you interface driver is really causing the time load problem. What is your hardware & ASIO driver you use?
Also I don’t think 1.6 GHz (2 or 4 core?) is not enough to run the layering’s you are trying. I have found that a minimum of 2.5 GHz with 4 cores would be needed for some of the layering’s you described.
Thanks, Dave-- I’m using an XR18, with the X-AIR ASIO driver (current version, 5.12). I’m definitely mindful that I might just be overloading the processor, but I’ve used this NUC with Pianoteq and VB3 for a couple of years now, and I don’t remember running into this problem in the past.
Thanks Jeff, I also use the X18 so the driver probably isn’t it, the file you referred to is tied to the Net Framework package so possibly a Windows update of that introduced the problem. Have you tried going back to an earlier version of Cantabile that was stable for you? Also are you using the latest X18 drivers and if so have you tried rolling them back to an earlier version? I am trying to figure out when things changed as far as the performance degradation …
A Windows update was the best I could come up with as well. I started having issues independent of any Cantabile update (and I rolled up from C3 to C4 last weekend in hopes that might help solve the problem), and there haven’t been any hardware changes.
I recently posted almost exactly the same question. I’ve been really busy with bass guitar so not on the keyboard as much the past 5 months. During this time period I was hearing some noise when playing piano using Mrs. Mills and figured it could be a loose cable or because I added some new software, some of which leaves services running (NI for example).
When I finally looked into it more closely last week I observed that when playing many simultaneous notes I was getting time loads of 115% and that’s when the noise was happening. I noticed at least one person mention that they were seeing just slightly higher cpu usage with V4 so I loaded up V3 and the issue is still there but reduced enough that you might not notice it if not looking for it.
I think we are probably using the same driver from Behringer - that is another change I made a month ago. I upgraded from 4.38 to 5.12.
I also upgraded Windows 10 to 21H2. Yes - too many changes without testing after each
I think Jeff and I are experiencing the same issue.
Yeah, thanks Doug-- I also remember people talking about slightly higher CPU load with 4 than 3, but for me is the Time Load that’s more problematic than the CPU stats. I can live with the CPU living in the 20s and 30s (which it does). And the two don’t correspond, which makes fine sense-- but the Time Load spikes don’t necessarily correspond to any CPU spikes.
I had unchecked the 5.12 driver’s safe mode checkbox - checking it again allows more notes before the issue occurs. I don’t know what it’s doing when you check it buy I notice it indicates you’re actually increasing the buffer size. As with any interface, increasing the buffer size will lesson distortion - when I mix in Mixbus32C I put it at 2048 and it can then run quite a few plugins without problem. I prefer 256 when playing keyboards. For my edrum kit I use 64 with the same interface.
The crux of this thread is that something changed - the same system was handling Mrs. Mills before. Going forward I’m going to log any change with the date. I wish I knew how this problem relates to the various changes I made
I just loaded everything onto a MB Pro running Boot Camp. Just running Amplitube gets up into 80% and glitchville. While it doesn’t discount everything else about the NUC setup, is Amplitube really just that much of a hog?
Probably worth loading in a utility to see the actual cpu frequency at any point in time to confirm. Here’s a free one: https://openhardwaremonitor.org/
Also, some bios provide a way to prevent throttling down and force run at max clock all the time (that’s what I do with my desktop workstations).
