• From: uk
  • Registered: 2023-07-23
  • Posts: 34

Re: Audio dropouts with UFX iii on MacOS 14.7.1 (23H222)

FZ wrote:

You mean your UFX+ works well with the kernel driver on your MacBook? That's a crazy issue. I’ve seen in the forum that it also affects the AVB and Dante interfaces, always in connection with ARM MacBooks. So mostly Interfaces with high channel counts…

This isn't well understood in this forum and in this sense I too experience a bit of frustration coming across, whenever the issue of USB2 (specifically DriverKit on macOS) vs Thunderbolt (PCIe) is raised.

USB is fine. Its bandwidth is enough. Its latency doesn't seem to be a problem. The issue is with timing, more specifically with the driver code servicing all this data, interruptions in the flow of which are undesired and very noticeable, that driver code seems to not have execution scheduling as tight as we'd like. Even more specifically, macOS forces audio driver code away from the kernel extension layer where it could have higher priority (as kext), and pushes it into the user space where, as DriverKit client, especially for a USB audio device, it simply cannot achieve competitive timing against a kernel component servicing e.g. a PCIe device (NVMe drive in TB hub is a good example, and so is a PCI interface in a TB cage).

So some of us using USB audio, depending on our hardware setups, experience glitches with the DriverKit drivers, and rock solid performance using the Kernel Extension driver, with the same audio interface, same bus, same cable. RME's own video acknowledges the tradeoffs to security, or to audio performance. Having to choose between the two isn't great.

One could argue about bandwidth vs latency all day but the issue subtly involves CPU scheduling, call stack depth, security boundaries, interrupt servicing, bus transfer timing, all sorts of things none of us who disable SIP in frustration care about more than we do about our audio sounding pristine.

I, for one, do not accept the TB complexity argument. Reading this forum for quite a while now, I'm not convinced I'd like to come anywhere near USB the next time I have to buy an audio interface, to use with my Mac.

But I'm very interested in arriving at a detailed, nuanced, grown-up technical explanation of the exact nature of the issue, as well as in getting some clear-cut and honest guidance on what to expect and what to avoid, ideally from RME. Fingers crossed.

Fireface UCX, Focusrite OctoPre, Sound Devices MixPre-D, Mac Studio M2 max, macOS Sonoma

52 Reply by FZ (edited by FZ 2025-01-12 23:38:00)

  • FZ
  • AD/DA expert
  • Offline
  • Registered: 2019-12-08
  • Posts: 40

Re: Audio dropouts with UFX iii on MacOS 14.7.1 (23H222)

gmarinov wrote:

But I'm very interested in arriving at a detailed, nuanced, grown-up technical explanation of the exact nature of the issue, as well as in getting some clear-cut and honest guidance on what to expect and what to avoid, ideally from RME. Fingers crossed.

I'm also waiting for that. It would be helpful to receive a brief statement on the situation. Has there been any communication with Apple regarding the issue, or any other updates? It would be useful to know whether we should be preparing for a solution as Mac users or if we should consider returning the interface and looking for an old UFX+ device.