1 Topic by nikolaib (edited by nikolaib 2016-03-16 22:03:58)

  • Registered: 2016-03-16
  • Posts: 4

Topic: USB-Audio and OSX, Known issue by RME, how about you?

Hi
I just want to put some focus on a Mac vs USB audio issue we experienced. It might help other users.

We set up a new system in august 2015, consisting of a brand new Mac Pro, and a MADIface XT.
The application requires us to run this with low latency, 15 hours a day, we tested with 32 samples, and it ran fine. So we set it to 64 to be safer.
The mac was cleaned, and all scheduled services and tasks where disabled, all power save features disabled. Turned off wifi, etc, and took the computer offline. All seems good.

Then we started having these noises like once or twice a day. Raising buffer to 128 does not change it.
It would sound like a sawtooth tone. looking at recordings we figured the noise was the last buffer looped for about 2 seconds. Something was clearly blocking audio.
Contacted RME through our local distributor and this is the response from RME:

This is a bug in newer OSX versions that affects any USB audio device.
The system seems unable to run for more than a few hours at a stretch withouth an interruption. As a simple workaround, it will help to unplug and reconnect the audio device

What? Why is this not more talked about? Can apple just break USB-audio and nobody cares?

So we put the madifaceXT in the trash, got a Thubdebolt PCI chassis and put in an old RME MADI PCIe card we had.
Now, the system is working fine. without USB

So just a heads up to anyone buying a USB audio device, planning to use it on a mac..

  • Jeff
  • Jeff
  • Moderator
  • Offline
  • From: USA
  • Registered: 2007-05-21
  • Posts: 4,619

Re: USB-Audio and OSX, Known issue by RME, how about you?

Is this the Fusion Drive (SSD+HDD) on the Mac? That is known to cause issues.

Regards,
Jeff Petersen
Synthax Inc.

Jeff's Website

3 Reply by Randyman... (edited by Randyman... 2016-03-16 23:21:56)

  • From: Houston, TX
  • Registered: 2007-05-11
  • Posts: 1,191

Re: USB-Audio and OSX, Known issue by RME, how about you?

nikolaib wrote:

So we put the madifaceXT in the trash, got a Thubdebolt PCI chassis and put in an old RME MADI PCIe card we had.
Now, the system is working fine. without USB

You do realize the MADIface-XT is also a ePCIe (External PCIe) interface (basically a MADI-FX PCIe card in a 1/2RU enclosure plus a few killer features)?  I've never even run my MADIface-XT over USB - strictly PCIe on PC over here smile

You can use a PCIe-to-ePCIe card in that same TB PCIe chassis, and get the same performance you would if the MADIface-XT was an internal PCIe card.  No need to trash it or swap to an interface with 1/3rd the I/O and no TM-FX (I seriously doubt you threw away a $3K interface, but I get your point)

cool

MADIface-XT+ARC / 3x HDSP MADI / ADI648
2x SSL Alphalink MADI AX
2x Multiface / 2x Digiface /2x ADI8

4 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 24,837

Re: USB-Audio and OSX, Known issue by RME, how about you?

nikolaib wrote:

What? Why is this not more talked about?.

The net has plenty of it. Just look in the right places.

nikolaib wrote:

run this with low latency, 15 hours a day

Like it or not, but in 99.9% such applications are done on a Windows PC. And there it will work even for weeks, months and years. And: While PCIe is less affected, it can happen there too. Also: the lower the buffer size, the sooner the drop-out might occur. With 128 you might have got around the 15 hours without noticing a problem.

Regards
Matthias Carstens
RME

MC's Website

  • Registered: 2009-05-02
  • Posts: 105

Re: USB-Audio and OSX, Known issue by RME, how about you?

MC wrote:
nikolaib wrote:

run this with low latency, 15 hours a day

Like it or not, but in 99.9% such applications are done on a Windows PC.

I could be wrong, but I think 99,9% of the users don't need such applications, that run 15 hours a day at system limit wink Perhaps because of this it's "not more talked about this".

The problem isn't really a problem, if there are alternatives (even on Mac). As Randyman mentioned, you can run Madiface XT over PCIe. Did you try this? Or you could try older Versions of OSX and/or different SSDs or HDs, if the "bug" is caused by the Fusion Drive, as Jeff mentioned.

  • Registered: 2016-03-16
  • Posts: 4

Re: USB-Audio and OSX, Known issue by RME, how about you?

Hi
Thank you for all your replies.
No I did not actually trash the MadiFaceXT, it is still a great product. smile
The original setup was with a MADI-Usb interface, and I then upgraded it to a XT, hoping that would help.
Yes, I could connect it vie pcie, but I figured I'd safe and use the madi pci card, I dont need FX, and I only need 64 channels.
Anyway, my point is more about information,

The net has plenty of it. Just look in the right places.

Probably my research or google skills are lacking but I feel that is besides the point, I should not need to verify that the product works as advertised.
my point is that this issue should be some how stated by RME when I purchase  products of this price range.
Maybe in the Notes section on the product page.
It is advertised as a interface that can run with low latency on OSX, and that, we all agree, it can not.
Even the our dealer did not know about this (national importer of RME products).

Like it or not, but in 99.9% such applications are done on a Windows PC

Probably true, and I'm a PC fan any day of the week, but this is a mac system.. so, again if we where informed about the issue we would have built a PC system from start.

I could be wrong, but I think 99,9% of the users don't need such applications, that run 15 hours a day at system limit wink Perhaps because of this it's "not more talked about this".

We are way bellow system limits, but I would assume that this would also apply to any system, as long as it is running all day? and that is pretty much any professional setup I would guess. ?

Again thank you all for your replies, and lets hope that Apple, the real crook here, fixes their broken OS.

  • Registered: 2013-07-30
  • Posts: 4

Re: USB-Audio and OSX, Known issue by RME, how about you?

which OSX did you talk ? is it maverick ? yosemite ? el capitan ?

  • Registered: 2009-05-02
  • Posts: 105

Re: USB-Audio and OSX, Known issue by RME, how about you?

nikolaib wrote:

The application requires us to run this with low latency, 15 hours a day, we tested with 32 samples…
Then we started having these noises like once or twice a day. Raising buffer to 128 does not change it.

nikolaib wrote:

I could be wrong, but I think 99,9% of the users don't need such applications, that run 15 hours a day at system limit wink Perhaps because of this it's "not more talked about this".

We are way bellow system limits, but I would assume that this would also apply to any system, as long as it is running all day? and that is pretty much any professional setup I would guess. ?

Sorry, but I doubt, that constantly recording 196 audiotracks a whole day or even 15 hours without any break is an application, that is needed from numerous users in the world (except the NSA tongue that would not use Apple or Windows PCs). Perhaps I misunderstood your application or what you want to do…

  • cyrano
  • DAW Mastermind
  • Offline
  • From: Belgium
  • Registered: 2015-04-28
  • Posts: 695

Re: USB-Audio and OSX, Known issue by RME, how about you?

There are several bugs here.

One is in El Capital and maybe Yosemite too. Could be fixed with the latest update. Several users reported having no problem anymore. It's to soon to tell and I still don't have enough samples to judge if it cures the bug for everyone.

The other one is in the Intel USB peripheral chips that go with the Haswell platform. It affects Windows PC's too, but not on every kind of implementation. Obviously, this can't be fixed. And Intel has a contract that absolves them from bug fixing, so it's the end user that's stuck with it.

A third one seems to be in Core audio, when it comes to long recordings. There's a fix out there that is stupid but simple. I've recorded 6 hour sessions with my old FF400 and USB interfaces without any problem.

I'm amazed RME still doesn't know about it. It seems they just prefer to hate Apple. I can understand that, as big companies are not very likeable, but from a commercial standpoint, it's just plain stupid.

If you need to do very long recordings, you need to prevent Core audio from going to sleep. That's a bug in multi core setups that is really stupid and I'm sure Apple could fix it, but the numbers are against it. Apple doesn't care about less than 1 ppm of their customers. The sad truth is, Apple is on a road leading to the disappearance of OSX. The ipad Pro is one of the important steps in this process. Once all development can be done on ARM/iOS, there's no need for OSX or Intel machines anymore.

The fix for very long recordings consists of a simple shell script (or an Apple script, if you prefer, or Python, or Perl), that speaks a space every 10 seconds or so. It's enough activity to keep Core audio awake. It cures the problem with any DAW I could test, including ProTools, Reaper, Audacity and Cubase.

BTW, I've seen the same kind of problem with MySQL on Linux, before this goes into an old fashioned OS war...

Only BSD seems to avoid it completely. MySQL went into maintenance because the task was running on another core from the main task and it couldn't "see" the heavy processing going on on other cores. Everything went down because of a minor error in the distribution of tasks over different cores.

The problem doesn't even surface for most users, as it depends on hardware and software used.

MB Pro - 2 X FireFace 400, FF800 & DigiFace USB
ADAT gear: Korg, Behri, Fostex, Alesis...

cyrano's Website

  • garry
  • S/Mux specialist
  • Offline
  • From: Wales, UK
  • Registered: 2007-05-12
  • Posts: 251

Re: USB-Audio and OSX, Known issue by RME, how about you?

I'm so amazed people still use Mac for audio purposes.. 99% of threads about issues with hardware/software/compatibility is usually about Macs. At least you got the apple logo though. Apple just doesn't care about compatibity with anything else that doesn't also have an apple logo on it.

_

RME HDSP9652 | | RME Digiface USB
  • Registered: 2009-05-02
  • Posts: 105

Re: USB-Audio and OSX, Known issue by RME, how about you?

garry wrote:

99% of threads about issues with hardware/software/compatibility is usually about Macs.

If you analyze all threads in this forum, you will never get 99% of real problems caused by Macs. Sorry, that's nonsense (for me it seems to be the opposite). Most problems I find here are caused by users, independent from Win or Mac. I am in both worlds and never had any Audio problems with Macs or WinPCs. On the other side I never have recorded 15+ hours nonstop roll

  • gigi
  • ADAT user
  • Offline
  • Registered: 2010-02-27
  • Posts: 29

Re: USB-Audio and OSX, Known issue by RME, how about you?

nikolaib wrote:

The original setup was with a MADI-Usb interface, and I then upgraded it to a XT, hoping that would help.
Yes, I could connect it vie pcie, but I figured I'd safe and use the madi pci card, I dont need FX, and I only need 64 channels.
Anyway, my point is more about information,

The net has plenty of it. Just look in the right places.

Probably my research or google skills are lacking but I feel that is besides the point, I should not need to verify that the product  works as advertised.
my point is that this issue should be some how stated by RME when I purchase  products of this price range.
Maybe in the Notes section on the product page.
It is advertised as a interface that can run with low latency on OSX, and that, we all agree, it can not.
Even the our dealer did not know about this (national importer of RME products).

Now, this is a bit confusing for us OSX users who are currently seeking to invest in new RME USB interfaces. Right on the top of RME's main page it is stated that RME's products work without known issues under El Capitan, but in this thread Mathias seems to admit the existence of widespread stability problems on OSX. I don't see how both pieces of information can be compatible.

If, from what has been written here, USB-audio is broken on OSX, and given the number of professional MAC-based studios out there, I agree with nikolaib that it would be fair for RME to inform users that RME's USB interfaces are currently not stable when used with recent Apple products. Even if it is Apple's fault. If RME could elaborate on whether this is the case, it would be much appreciated.

In my case, and probably for many others, 100% stability and reliability is the main reason for having been unconditional users of the awesome palette of RME products over many, many years. One or two glitches every 15 hours might seem unimportant for some, but for a professional studio one or two ruined recordings a day is unacceptable.

many thanks
gigi

13 Reply by garry (edited by garry 2016-04-01 20:27:13)

  • garry
  • S/Mux specialist
  • Offline
  • From: Wales, UK
  • Registered: 2007-05-12
  • Posts: 251

Re: USB-Audio and OSX, Known issue by RME, how about you?

loonypac wrote:
garry wrote:

99% of threads about issues with hardware/software/compatibility is usually about Macs.

If you analyze all threads in this forum, you will never get 99% of real problems caused by Macs. Sorry, that's nonsense

I didn't mean only this forum (RME) its across most if not all audio forums, if its a "problem" you will normally see OSX involved

_

RME HDSP9652 | | RME Digiface USB
  • Registered: 2016-03-16
  • Posts: 4

Re: USB-Audio and OSX, Known issue by RME, how about you?

Thank you cyrano for some VERY interesting information.

loonypac: That was never our intention, as staded, 64ch in/out low latency, no RME-FX, stable operation, was our goal

gigi: Happy to hear someone else concerned.

loonypac: As I said, I'm usually a PC guy, and not a apple fan, but even if your statistics was right,  you can't deny that OSX has a major share in the music industry, so we have to deal with it.

  • Registered: 2016-02-02
  • Posts: 1

Re: USB-Audio and OSX, Known issue by RME, how about you?

gigi wrote:

One or two glitches every 15 hours might seem unimportant for some, but for a professional studio one or two ruined recordings a day is unacceptable.

I agree! I had a couple of long sessions recently and experienced a few glitches, fortunately between takes.

All the best,
Manuel

  • Registered: 2016-03-16
  • Posts: 4

Re: USB-Audio and OSX, Known issue by RME, how about you?

Does anyone (like RME) know if Apple has resolved this issue in Sierra?

17 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 24,837

Re: USB-Audio and OSX, Known issue by RME, how about you?

https://www.forum.rme-audio.de/viewtopic.php?id=24193

Regards
Matthias Carstens
RME

MC's Website