Topic: Thoughts on RME drivers with NVIDIA

Hi all!

David again, this time with some report.

I use a Dell Precision 3560. The laptop includes an NVIDIA T500 display driver, in addition to the Intel integrated graphics. I recently updated the NVIDIA to the latest driver (Version: 31.0.15.1654, Date: 6/22/2022). However, I noticed that playing audio, even under Windows with the Babyface Pro, is not stable. In addition to crackles, there was also dropouts. And the Fireface USB Settings window showed errors, starting at 1/5, then 2/10, 3/15, 4/20 in the matter of minutes. Once I rolled back the driver to version 30.0.15.1199 (date: 3/6/2022), the crackling stopped and it's 0/0 errors again.

I don't understand why a graphics card driver can have such big impact on audio interface performance? And, should I continue to let the graphics card remain on the current version of the driver?

Thank you very much!

David

Re: Thoughts on RME drivers with NVIDIA

You are lucky that you have an nvidia driver that is not causing trouble, stay on the older version.

The graphics card driver is the cause of the problems, and it would be the same trouble with other audio interfaces, not just RME.

Regards,
Jeff Petersen
Synthax Inc.

3 (edited by davidlaijiajun 2023-01-27 20:21:09)

Re: Thoughts on RME drivers with NVIDIA

Jeff wrote:

You are lucky that you have an nvidia driver that is not causing trouble, stay on the older version.

The graphics card driver is the cause of the problems, and it would be the same trouble with other audio interfaces, not just RME.

Thank you! It amazes me that these things have to work very well with each other, even graphics card drivers matter!!!

4 (edited by mkok 2023-01-27 21:18:46)

Re: Thoughts on RME drivers with NVIDIA

Definitely a known issue. It’s with any audio interface it causes a problem as well. When I used one I only used the driver windows installed and never installed the nvidia latest and their enhancements. I just use the intel inbuilt video on my pc these days. My pc has one job and that’s music production and I keep it that way. I use an iPad or a second pc for anything else.

Babyface Pro Fs, Behringer ADA8200, win 10/11 PCs, Cubase/Wavelab, Adam A7X monitors.

5

Re: Thoughts on RME drivers with NVIDIA

You can configure the nVidia to being inactive, only using the Intel one. The nVidia is superfluous unless you play games.

Also run LatencyMon from Resplendence to see if your notebook can run at lower latencies. As this is a Dell the expected answer is 'not at all'.

Regards
Matthias Carstens
RME

Re: Thoughts on RME drivers with NVIDIA

MC wrote:

You can configure the nVidia to being inactive, only using the Intel one. The nVidia is superfluous unless you play games.

Also run LatencyMon from Resplendence to see if your notebook can run at lower latency at all. As this is a Dell the expected answer is 'not at all'.

Well, I can actually run this at 64 samples without any problems, so long as I don't update my graphics driver. So, all good. I was just curious -- how does this work that even the graphics card driver would impact the performance of the audio interface.

7 (edited by ramses 2023-01-28 10:43:36)

Re: Thoughts on RME drivers with NVIDIA

But the reverse case is also possible, that an nVidia CPU can relieve the system and prevent audio dropouts.

About 8 years ago, I tried connecting my old UFX to a Lenovo business laptop (T-series), which had an nVidia chip installed in addition to the CPU-internal Intel graphics.

The problem here was that I could only play music without interruption when browsing the Internet with Firefox (MusicBee with ASIO support) if I set the ASIO buffer size to at least 256 samples (at single speed).

I found this strange and didn't expect it because the laptop had excellent DPC latencies down to around 60 microseconds (measure in idle mode with LatencyMon) and the CPU load was very low. In addition, USB2 from the Intel chipset is usually rather uncritical, as one reads.

At that time, the issues could be completely solved by setting the graphics driver to use the nVidia GPU for the Firefox application.

After that, there were no more audio dropouts, even with the lowest setting of the ASIO buffer size of only 48 samples (@single speed).

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14

Re: Thoughts on RME drivers with NVIDIA

ramses wrote:

But the reverse case is also possible, that an nVidia CPU can relieve the system and prevent audio dropouts.

About 8 years ago, I tried connecting my old UFX to a Lenovo business laptop (T-series), which had an nVidia chip installed in addition to the CPU-internal Intel graphics.

The problem here was that I could only play music without interruption when browsing the Internet with Firefox (MusicBee with ASIO support) if I set the ASIO buffer size to at least 256 samples (at single speed).

I found this strange and didn't expect it because the laptop had excellent DPC latencies down to around 60 microseconds (measure in idle mode with LatencyMon) and the CPU load was very low. In addition, USB2 from the Intel chipset is usually rather uncritical, as one reads.

At that time, the issues could be completely solved by setting the graphics driver to use the nVidia GPU for the Firefox application.

After that, there were no more audio dropouts, even with the lowest setting of the ASIO buffer size of only 48 samples (@single speed).

Very interesting. I have to say, if I only use the integrated graphics, I do get more stable performance on my Babyface Pro, on the cost of louder fan noise from my laptop. However, the NVIDIA will give me a quiet laptop, but there will be USB transmission errors. My current solution is the default setting of "auto select", which gives me 0 USB transmit errors and a quiet laptop.