UR interfaces audio drop outs (follow-up thread)

Hi Hippo,

Thanks for this useful info.

  • Setting “Power Management” to “Prefer Maximum Performance” in the NVIDIA Control Panel (3D Settings) appears to have addressed the audio dropouts in Cubase.
  • Setting “Advanced - ISO Extra Buffer” in Yamaha’s BETA driver to “2.000ms” has also significantly reduced the audio dropouts in other Windows applications. A lower value would still cause infrequent dropouts.

More testing is needed. I’ll update this message in due course.

I have updated a relevant NVIDIA support ticket and still waiting for NVIDIA’s response.

installed LatencyMon, launched and began to work, 10 minutes later appeared the following message (see screenshot). Tell me more details what should I do in windows 7 that all worked consistently?

Useful utility. Within 1 minute, Wdf01000.sys (Kernel mode Driver Framework Runtim, Microsoft Corporation) DPC routine reportedly took 6212μsec to return control of the CPU.

Within less than 10 minutes, another delay was presented by dxgkrnl.sys (DirectX Graphics Kernel, Microsoft Corporation).

These are with NVIDIA 3D settings at “Prefer Maximum Performance”.

Which goes back to my earlier comment that this is an issue that Yamaha, NVIDIA and Microsoft (and potentially Intel, as it appears to be more prevalent on X99 systems) have to collaborate on, to address.


Your conclusion sounds unreasonable. LatencyMon is not a utility that’s somehow tailored to Yamaha USB-devices. It’s a general-purpose utility that’ll check your OS for latency problems. In your case, LatencyMon identified that two Microsoft OS components, one relating to the OS kernel, the other relating to DirectX, are causing latency problems. These problems have nothing to do with Yamaha’s USB driver. In fact, your screenshot proves that latency problems are caused by something that isn’t remotely associated with Yamaha, so I’m puzzled why you’d want Yamaha to help Microsoft or nVidia debug problems in Microsoft or nVidia products.

Furthermore, judging from the results of LatencyMon it’s reasonable to assume that your latency problems would manifest with any other USB-connected audio device, too, e.g. from Roland or Behringer.

I’m new to this forum, and I’m still puzzled why so many folks here seem intent to pin-point general OS latency problems to Steinberg’s UR devices, when even a very cursory search will clearly prove that this has nothing to do with Yamaha or Steinberg but can and will happen in absence of any Steinberg UR devices:

https://answers.microsoft.com/en-us/windows/forum/windows_10-performance/extremely-high-dpc-latency-from-directx-graphics/c9cb07bd-83b6-446d-a898-48a67819472d?page=2&auth=1

https://www.google.de/search?q=x99+latencymon

That said, I can actually help you find the real culprit of your latency issues:

  1. Your machine is built on the X99 chipset. I take it that at some point you were talking to a half-knowledgeable computer salesperson who claimed that “for professional use you should get a X99 mainboard because Intel’s X99 chipset is for professional users”. This is not true. The exact opposite is true. The X99 chipset is for “enthusiasts”, which is Intel’s marketing code word for “kids with rich parents who want an I Am Rich computer”. Mainboards with X99 chipsets are not expensive because they are better than the mainstream chipsets - they are expensive because of much lower volume. High volume means high scale means low cost. Low volume means low scale means high cost.

X99/X299 chipsets are meant for a tiny niche of enthusiasts, mostly gamers, who only care about PCIe lanes so they can cram as many GPUs as possible into a single computer. They are not optimized for real-time, they are not as tested as the mainstream chipsets, and mainboard vendors do not spend as much time validating their mainboard with various combinations of RAM, CPU, PCIe devices other than GPUs - or USB interfaces.

If you want a tested, proven and professional-level platform, chose the Q chipset line, e.g. Q87, Q97, Q170, Q270. This chipset is almost identical to the Z87/Z97/Z170/Z270/Z370 line from a performance perspective, but validated for mission-critical tasks, and actively supported in enterprise environments. Intel spends most of their validation efforts on the Q chipsets to make sure that they run as smooth as possible for the millions of business customers who depend on them.

  1. You use a software RAID. RAID configurations were never meant to support real-time tasks. RAID is for average high-throughput (with “average” being the keyword here) and data safety, and especially software RAID configurations are notorious for huge latency spikes because of internal reconciliations or maintenance tasks. Your LatencyMon screenshots shows that spikes occur on the kernel level, i.e. the level where Microsoft’s software RAID is running. It is very probable that your RAID is causing latency issues.

@humbugger

  1. Steinberg has not informed us that its products won’t work well with X99 systems,
  2. The problem predates deployment of RAID and it was confirmed even without it.
    and
  3. If Steinberg doesn’t care to collaborate with NVIDIA, Microsoft and Intel, to solve an issue that affects its products and its customers, then who will?

Furthermore, I’m not a kid, as a matter of fact I am a middle aged professional computer scientist who happens to have written Windows kernel device drivers in the past. Currently, I need the PCIe lanes and the 4x GPUs for deep learning applications. This is my workstation and I also use it for Cubase. Hence, I think that asking for a solution is the reasonable thing to do in my case.

Trolling me here isn’t really helping me or other Steinberg customers overcome the technical problem.

Kind regards

I have a General problem of the appearance of sharp noise from the speakers, maybe my UR824 is faulty? me it is very annoying that I constantly write about it, and they get no specific answers. Motherboard ASUS Z-97K, Windows 7 Professional 64 bit, i-7 4790, 16 gb of RAM SOMEONE PLEASE ADVISE THAT CAN HELP ME? Sorry for my English, is not my native language

@Nikolaos As it happens, I’m also a professional software engineer, having developed a number of commercial a/v applications. Hence me trying to help you out with issues that clearly stem from your particular choice of hardware and software. Your situation is not uncommon, although what is uncommon, very uncommon actually, is your claim that you are a professional computer scientist (your public CV does not mention any device driver development, by the way), and yet you are not able to investigate the matter properly. Anyway, I now realize you don’t want help. You just want to complain and point fingers at Yamaha/Steinberg engineers.

Normally I wouldn’t mind, but I get the impression that your persistent complaining here is taking away valuable development time from the fine folks at Yamaha/Steinberg which might be better invested in fixing real bugs instead of trying to replicate issues that originate in users like you who use a system that, regardless of the USB audio device attached, will never be suited to real-time audio processing in any sense.

Who is Nikolaos?

In any case, in my humble opinion it does seem that the few users that have issues are the minority and it is more to do with PC’s built on the inexpensive end or owned by those that do not have the knowledge to effectively troubleshoot the issues. That being said, I am surprised this is still an issue for any user of Steinberg products. It is really a tough call to place blame on any particular manufacturer or user. It just sucks that it is even an issue to begin with.

I suggest none of us get frustrated with each other. We will not move forward by calling out others. Let’s just try to move forward as a community so we can hopefully find a fix that works for everyone. :slight_smile:

Cheers!

I tested with my system on this setting:

  • Nidia control panel
    Set Power management mode to Prefer maximum performance
    Set Threaded optimization to Off
  • skip initial sample (v1.9.9)

It works pretty good so far. Havent got an drop out yet.

My system in detail: https://pcpartpicker.com/user/phil_wc/saved/Ptt8dC
Win8.1

I just wanted to mention that on my new machine (a Dell XPS 15 core i7-7700 quad core) I have only gotten very rare momentary dropouts in Windows Media Player and the Edge browser. It is as good or better as I have experienced with my previous audio interfaces (a Focusrite Scarlett and an M-Audio Fast Track Pro) on previous machines. I don’t think it is realistic to expect to NEVER get an audio dropout. Every now and then, Windows is going to do something that interrupts the audio stream. In my previous experiences, sometimes, Windows Media Player would just stop and the only way to get it playing again was to close it and restart it. The same thing occasionally happened in internet browsers while watching YouTube videos or streaming audio. Of course, it would be ideal if I never got an audio dropout but I can listen to Media Player and Browse for hours without any dropouts. In Cubase, I have yet to have any sort of dropout that wasn’t clearly due to just having the buffer size set too small. So, for my setup, I would say the ASIO part is pretty solid. The Windows driver part is very good (as good as I have had with other interfaces and computers) but not perfect.

Lastly, I also thought it may be helpful to mention that I actually use the Thunderbolt Dock (Dell TB16) with the UR44 plugged into the dock. I was concerned about how well it would work when I ordered it but I have a second monitor plugged into the dock as well as my Axiom Pro keyboard and it has not been any problem at all. It performs as well as when they are plugged directly into the computer.

Video drivers that hold interrupts or hog IO lines for way longer than they should has been a decades-old problem in the pro audio community. Like, literally, 20 years of complaints about video drivers impacting audio drivers. Buy the oldest and slowest video card you can for your audio PC and you’ll never have a problem.

Buy the oldest and slowest video card you can for your audio PC and you’ll never have a problem.

Where’s the fun in that? It’s ALL about the toys!!
:smiling_imp: :laughing: :smiley:

lol, spend it on plugins instead!

I can also confirm, that the problem seems to be solved, when “Power Management” ist set to “Prefer Maximum Performance” in the Nvidia Control Panel. So as many have said before, the issues has something todo with the videocard/drivers.

But this solution is defenitely not optimal, since the GPU is clocking higher when idle with these settings, which causes a higher power consumption. It’s a workaround but should only be a temporary solution.

I find the new Beta 2 drivers work very well on my system. previous serious of drivers completely useless. Regardless of the NVIDIA driver setting.

Why don’t Steinberg just introduce a buffer in order to avoid all these latency problems? I bought the Steinberg UR22mk2, but it is useless due to all the gaps and also lots of noise in my speakers (Yamaha HS7). I might have to return all this. The purpose of me buying the Steinberg was to record sound while playing piano using a pair of SE5 microphones.

I will probably return the Steinberg, and try another soundcard. Amazing that they haven’t solved this issue. When googling the problem, I see that it has been here since at least 2012.

Hi Ed,
I used the UR22 on an old Win7 machine without any issues until my laptop broke up.
With my new HP laptop (Intel Celeron CPU N3060 @ 1,6 GHz, 8 GB RAM, Windows 10 64bit Home) I´m getting these dropouts continously. I installed the Beta driver - set the buffer time, extra buffer and base period to maximum, but this doesn´t change anything.

With every dropout this error is shown in the Log:
0:25.57 : 0 3 11
Maybe this might help for solving the issue.
Best regards,
Mirco

Using Cubase and audio/MIDI projects with this kind of processor I would honestly expect dropout to occur every now and then. Sorry to sound harsh but these kind of processors are not really up to the task for running real time audio with effects and maybe even VST Instruments. What kind of processor did the Windows 7 system have?

We have released the Yamaha Steinberg USB driver 1.10.0 officially now. Please have a look at the news just released:
https://www.steinberg.net/en/newsandevents/news/newsdetail/article/yamaha-steinberg-usb-driver-v1100-and-ur44ur242-firmware-updates-available-4580.html

The additional “Error Info” and “Advanced” tab settings are hidden in the release driver now but can still be accessed by pressing “Shift + Ctrl + A” while the control panel is open.
If the driver is uninstalled or updated, any modifications made here need to set accordingly again.

The update went smoothly but I will mention that just before I updated, I opened the dspMix App (in Windows) and on the “Information” tab where it says I have firmware versions 2.11 for Boot and Main, I clicked on “Check for Update” and it said I had the most current version. Of course, with 2.12 firmware available, I just downloaded it from the link provided and it went smoothly from there. Maybe the place where the app looks for the current version needs to be updated?