[QLab] RME Fireface 800s making digital noise

Brian Conner

unread,

Apr 27, 2010, 10:44:08 AM4/27/10

to ql...@lists.figure53.com

I was just wondering if anyone's experienced any noise "eventually" showing up from Firefaces and QLab.

Mac mini 2.0GHz, OS 10.6.3, 4GB RAM, firewire 800 to two daisy-chained RME Fireface 800s

Aggregate device created.

"Tonight during our invited dress, about 30 minutes into it, we started to get the digital noise issue again. I'm not positive it was exactly the same sound. The designer described it as starting the same, but then it quickly got worse and became like a digitally distorted loop that was very loud. The computer had been on since 9am, but at about 7.45 we had quit Qlab and restarted it (just the program, not the full computer). Interestingly just quitting the program and restarting it cleaned it up, and now, in intermission, it has behaved for the past 35 minutes or so"

I am surprised that the RME gear is problematic at all. A key ingredient is the restarting QLab seems to cure it for a while.

The next thing is to start reinstalling and/or rolling back RME drivers I suppose.

System Log shows a many strange messages:

MIDIServer[114] PlugIn Fireface MIDI.plugin -- /Library/Audio/MIDI Drivers/ does not contain a supported architecture.

QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 1

4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 1

4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 0

4/27/10 8:53:24 AM QLab[110] audioDeviceNominalSampleRateDidChange: Fireface 800 (D80)

4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 2

4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 3

4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 4

that keeps repeating through channels 1-28 for a while, then there's a TON of these this morning:

4/27/10 9:01:39 AM QLab[110] ERROR: -10877 [É’ˇˇ] row=28, column=28, FFSoundDevice.m: 1017

4/27/10 9:01:39 AM QLab[110] ERROR: -10877 [É’ˇˇ] row=76, column=28, FFSoundDevice.m: 1017

4/27/10 9:01:39 AM QLab[110] ERROR: -10877 [É’ˇˇ] row=124, column=28, FFSoundDevice.m: 1017

4/27/10 9:01:39 AM QLab[110] ERROR: -10877 [É’ˇˇ] row=172, column=28, FFSoundDevice.m: 1017

4/27/10 9:01:39 AM QLab[110] ERROR: -10877 [É’ˇˇ] row=220, column=28, FFSoundDevice.m: 1017

Unfortunately, replacing Fireface units and cables, and confirming word clock setup has not fixed this.

The next thing is to start reinstalling and/or rolling back RME drivers I suppose.

Am I missing something? Sample rate setting in the Audio/MIDI setup or something?

This is an immediate problem so I apologize for the slightly hands-in-the-air request for help.

The fact that things are OK for a while is weird.

Christopher Ashworth

unread,

Apr 27, 2010, 11:05:51 AM4/27/10

to Discussion and support for QLab users.

On Apr 27, 2010, at 10:44 AM, Brian Conner wrote:

> I was just wondering if anyone's experienced any noise "eventually"
> showing up from Firefaces and QLab.
>
> Mac mini 2.0GHz, OS 10.6.3, 4GB RAM, firewire 800 to two daisy-
> chained RME Fireface 800s
> Aggregate device created.

This smells like an aggregate device problem to me. How easy is it to
recreate, and if it's relatively easy, does it happen with just one
device?

These are not good. I'm assuming you're not changing the sample rate
on the device, so the fact that it is changing of its own accord is
bad. Could conceivably be related to the aggregate device, if
CoreAudio is struggling to keep the devices synchronized.

Although of course I've never seen the aggregation code directly, I
know there's a lot of work involved in keeping devices meshed into an
aggregate. Since audio devices do not ever actually run at exactly the
sample rate they claim, sending audio to separate physical devices
requires a fair amount of cleverness and computation. It seems to me
this could also relate to the fact that the problem appears only after
a period of time, if the devices are slowly drifting from each other
and CoreAudio can't figure out how to keep them together.... that's a
guess, but I think a not unreasonable one.

-C
________________________________________________________
WHEN REPLYING, PLEASE QUOTE ONLY WHAT YOU NEED. Thanks!
Change your preferences or unsubscribe here:
http://lists.figure53.com/listinfo.cgi/qlab-figure53.com

--
Change your preferences or unsubscribe here:
http://groups.google.com/group/qlab

Brian Conner

unread,

Apr 27, 2010, 12:08:36 PM4/27/10

to Discussion and support for QLab users.

Thanks again for your super quick and knowledgeable reply.
After browsing RME Support Forums, I am frightened.
There seems to be much vagueness as far as causes and solutions across a plethora of problems.
There are many "known issues" with unknown causes and varying degrees of re-creatability of said issues.
For now, the quick fix attempt is swapping for different interfaces (MOTUs) and troubleshooting further once the RMEs are back in my hands.

P.S. I have never felt a hotter piece of gear than the aptly named Firefaces.

On Apr 27, 2010, at 11:05 AM, Christopher Ashworth wrote:

> On Apr 27, 2010, at 10:44 AM, Brian Conner wrote:
>
>> I was just wondering if anyone's experienced any noise "eventually" showing up from Firefaces and QLab.
>>

>> Mac mini 2.0GHz, OS 10.6.3, 4GB RAM, firewire 800 to two daisy-chained RME Fireface 800s

>> Aggregate device created.
>
> This smells like an aggregate device problem to me. How easy is it to recreate, and if it's relatively easy, does it happen with just one device?
>
>> "Tonight during our invited dress, about 30 minutes into it, we started to get the digital noise issue again. I'm not positive it was exactly the same sound. The designer described it as starting the same, but then it quickly got worse and became like a digitally distorted loop that was very loud. The computer had been on since 9am, but at about 7.45 we had quit Qlab and restarted it (just the program, not the full computer). Interestingly just quitting the program and restarting it cleaned it up, and now, in intermission, it has behaved for the past 35 minutes or so"
>>
>> I am surprised that the RME gear is problematic at all. A key ingredient is the restarting QLab seems to cure it for a while.
>>
>> The next thing is to start reinstalling and/or rolling back RME drivers I suppose.
>>
>> System Log shows a many strange messages:
>>
>> MIDIServer[114] PlugIn Fireface MIDI.plugin -- /Library/Audio/MIDI Drivers/ does not contain a supported architecture.
>>
>> QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 1
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 1
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 0
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 0
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceNominalSampleRateDidChange: Fireface 800 (D80)
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceNominalSampleRateDidChange: Fireface 800 (D80)
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceNominalSampleRateDidChange: Fireface 800 (D80)
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceNominalSampleRateDidChange: Fireface 800 (D80)
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 2
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 2
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 3
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 3
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 4
>> 4/27/10 8:53:24 AM QLab[110] audioDeviceStreamDescriptionDidChange: Fireface 800 (D80) forChannel: 4
>>
>> that keeps repeating through channels 1-28 for a while, then there's a TON of these this morning:
>
> These are not good. I'm assuming you're not changing the sample rate on the device, so the fact that it is changing of its own accord is bad. Could conceivably be related to the aggregate device, if CoreAudio is struggling to keep the devices synchronized.
>
> Although of course I've never seen the aggregation code directly, I know there's a lot of work involved in keeping devices meshed into an aggregate. Since audio devices do not ever actually run at exactly the sample rate they claim, sending audio to separate physical devices requires a fair amount of cleverness and computation. It seems to me this could also relate to the fact that the problem appears only after a period of time, if the devices are slowly drifting from each other and CoreAudio can't figure out how to keep them together.... that's a guess, but I think a not unreasonable one.

Sean Dougall

unread,

Apr 27, 2010, 12:59:33 PM4/27/10

to Discussion and support for QLab users.

One way to fix this (or at least to rule it out as a cause of the noise issues) is to make sure both Firefaces aren't working off their own internal clocks. I.e., provide a word clock connection from one Fireface to the other, or have them both reference some external clock source. That way you'll guarantee that they won't drift, and CoreAudio won't have to do any adjusting.

And if that doesn't fix it, then there's probably something else going on...

Sean

On Apr 27, 2010, at 8:05 AM, Christopher Ashworth wrote:

> These are not good. I'm assuming you're not changing the sample rate on the device, so the fact that it is changing of its own accord is bad. Could conceivably be related to the aggregate device, if CoreAudio is struggling to keep the devices synchronized.

Brian Conner

unread,

Apr 27, 2010, 1:07:45 PM4/27/10

to Discussion and support for QLab users.

Thanks. Word clock is setup properly. Plus, we're talking about running for over 6 hours just fine. I'm leaning toward thermal.

I first suggested trying a single interface, not creating an Aggregate Device to see how that worked, but the client really needed 14 analog outs and is seriously time constrained.

Basically - I need to troubleshoot it further on my own time after they have a working show. Thanks for your input!

Brian Conner

Computer Department

PRG Audio
A Division of Production Resource Group

250 E. Sandford Blvd.

Mount Vernon, NY 10550
914-662-3540 phone

914-662-3731 fax

bco...@prg.com

www.prg.com

John Leonard

unread,

Apr 27, 2010, 1:36:03 PM4/27/10

to Discussion and support for QLab users.

Yes, they do get very hot. I had one die a while back due to heat build-up in dual system. They need space above and below each unit to be safe, I think.

Please let us know when you get to the root of the problem.

Regards,

John

On 27 Apr 2010, at 17:08, Brian Conner wrote:

> P.S. I have never felt a hotter piece of gear than the aptly named Firefaces.

Rich Walsh

unread,

Apr 27, 2010, 1:46:32 PM4/27/10

to Discussion and support for QLab users.

On 27 Apr 2010, at 18:07, Brian Conner wrote:

> Thanks. Word clock is setup properly. Plus, we're talking about
> running for over 6 hours just fine. I'm leaning toward thermal.
> I first suggested trying a single interface, not creating an
> Aggregate Device to see how that worked, but the client really
> needed 14 analog outs and is seriously time constrained.
> Basically - I need to troubleshoot it further on my own time after
> they have a working show. Thanks for your input!

You didn't need to make an aggregate device for this: use one Fireface
connected to the computer with 8 analogue outputs and 6 ADAT Optical
outputs; route the ADAT Optical outputs to the other Fireface in
standalone mode, converting the ADAT Optical to analogue.

We've used the Fireface 800s fairly solidly for about 6 months without
any issues (except in Windows) - but I wouldn't recommend mounting two
next to each other unless you're planning on doing some cooking. I've
found the RME forums to be practically useless for troubleshooting as
they seem unusually full of nonsense: they'd have you believing that
Firefaces are fundamentally incompatible with the current FireWire
chipset in Apple's small-form-factor machines, even though RME fixed
that issue in 2008...

Rich

________________________________________________________
WHEN REPLYING, PLEASE QUOTE ONLY WHAT YOU NEED. Thanks!
Change your preferences or unsubscribe here:
http://lists.figure53.com/listinfo.cgi/qlab-figure53.com

--

Change your preferences or unsubscribe here:

http://groups.google.com/group/qlab

Brian Conner

unread,

Apr 27, 2010, 1:56:54 PM4/27/10

to Discussion and support for QLab users.

That's how this guy used to have some MOTUs hooked up, had some issue, then he switched to the Aggregate Device method with Success. This time with the Firefaces, he's starting with what we found to work just fine, but perhaps we should look into reverting to what you said: Using the 2nd unit just to do the D/A.
Thanks much Rich.

On Apr 27, 2010, at 1:46 PM, Rich Walsh wrote:

> On 27 Apr 2010, at 18:07, Brian Conner wrote:
>
>> Thanks. Word clock is setup properly. Plus, we're talking about running for over 6 hours just fine. I'm leaning toward thermal.
>> I first suggested trying a single interface, not creating an Aggregate Device to see how that worked, but the client really needed 14 analog outs and is seriously time constrained.
>> Basically - I need to troubleshoot it further on my own time after they have a working show. Thanks for your input!
>
> You didn't need to make an aggregate device for this: use one Fireface connected to the computer with 8 analogue outputs and 6 ADAT Optical outputs; route the ADAT Optical outputs to the other Fireface in standalone mode, converting the ADAT Optical to analogue.
>

> We've used the Fireface 800s fairly solidly for about 6 months without any issues (except in Windows) - but I wouldn't recommend mounting two next to each other unless you're planning on doing some cooking. I've found the RME forums to be practically useless for troubleshooting as they seem unusually full of nonsense: they'd have you believing that Firefaces are fundamentally incompatible with the current FireWire chipset in Apple's small-form-factor machines, even though RME fixed that issue in 2008…

mackerr

unread,

Apr 27, 2010, 2:21:14 PM4/27/10

to ql...@lists.figure53.com

On Apr 27, 1:56 pm, Brian Conner <bcon...@prg.com> wrote:
> That's how this guy used to have some MOTUs hooked up, had some issue, then he switched to the Aggregate Device method with Success. This time with the Firefaces, he's starting with what we found to work just fine, but perhaps we should look into reverting to what you said: Using the 2nd unit just to do the D/A.
> Thanks much Rich.
>
>

Or for about 1/4 the price of using a FF800 you could use an Aphex
141 to do the DA conversion.

http://www.sweetwater.com/store/detail/141/

You take an expensive item off the rental and save some heat load.

Mac

vengb

unread,

Apr 28, 2010, 8:01:25 AM4/28/10

to Discussion and support for QLab users.

On Apr 27, 2010, at 12:08 PM, Brian Conner wrote:

>
> P.S. I have never felt a hotter piece of gear than the aptly named
> Firefaces.

hmmnnn.. I haven;t noticed, but then it's in a portable rack with an
Octopre that, even with it's internal cross-rack fan, gets hotter
than a minimally ventilated 50w tube amp... the FF doesn't compare.

Jeremy Lee

unread,

Apr 29, 2010, 5:35:39 AM4/29/10

to Discussion and support for QLab users.

Only a guess, but it might be that there's some sort of memory leak/ problem with the Aggregate device setup. I haven't tried it in a show critical situation, but a LONG time ago, there were problems with it. I try to avoid them if possible.

How many outputs do you need? Can you not just use 1 or 2 of the ADAT outs from the first FF to feed converters? Or are you using all 48+ outputs? You can get 24 outs from a single FF800 and not waste FW bandwidth on another unit.

If you're only using the Analog outs, you should definitely clamp down on the FF bandwidth in the FF Settings panel, under "Limit Bandwidth".

Best,

Jeremy

On Apr 27, 2010, at 4:44 PM, Brian Conner wrote:

> Unfortunately, replacing Fireface units and cables, and confirming word clock setup has not fixed this.
>
> The next thing is to start reinstalling and/or rolling back RME drivers I suppose.
>
> Am I missing something? Sample rate setting in the Audio/MIDI setup or something?
>
> This is an immediate problem so I apologize for the slightly hands-in-the-air request for help.
>
> The fact that things are OK for a while is weird.

--
Jeremy Lee
Sound Designer, NYC - USA 829
http://www.jjlee.com

Jeremy Lee

unread,

Apr 29, 2010, 5:35:42 AM4/29/10

to Discussion and support for QLab users.

I did some A/B tests, and the Aphex doesn't even sound as good as the MOTU 828, much less the FF800. With this stuff, you get what you pay for. Fine for a gramophone or other bandwidth constricted system, but certainly not for anything you want to sound good...

On Apr 27, 2010, at 8:21 PM, mackerr wrote:

On Apr 27, 1:56 pm, Brian Conner <bcon...@prg.com> wrote:
That's how this guy used to have some MOTUs hooked up, had some issue, then he switched to the Aggregate Device method with Success. This time with the Firefaces, he's starting with what we found to work just fine, but perhaps we should look into reverting to what you said: Using the 2nd unit just to do the D/A.
Thanks much Rich.

Or for about 1/4 the price of using a FF800 you could use an Aphex
141 to do the DA conversion.

http://www.sweetwater.com/store/detail/141/

You take an expensive item off the rental and save some heat load.

Mac

Jeremy Lee

unread,

Apr 29, 2010, 5:35:40 AM4/29/10

to Discussion and support for QLab users.

If you don't have an outboard converter, run an ADAT cable from the 1st FF into the 2nd one, and use it as a DA converter. Disconnect the FW cable from the 2nd one.

On Apr 27, 2010, at 7:07 PM, Brian Conner wrote:

> Thanks. Word clock is setup properly. Plus, we're talking about running for over 6 hours just fine. I'm leaning toward thermal.
> I first suggested trying a single interface, not creating an Aggregate Device to see how that worked, but the client really needed 14 analog outs and is seriously time constrained.
> Basically - I need to troubleshoot it further on my own time after they have a working show. Thanks for your input!

--
Jeremy Lee
Sound Designer, NYC - USA 829
http://www.jjlee.com

________________________________________________________
WHEN REPLYING, PLEASE QUOTE ONLY WHAT YOU NEED. Thanks!
Change your preferences or unsubscribe here:
http://lists.figure53.com/listinfo.cgi/qlab-figure53.com

--

Change your preferences or unsubscribe here:

http://groups.google.com/group/qlab

Brian Conner

unread,

Apr 29, 2010, 9:41:51 AM4/29/10

to Discussion and support for QLab users.

Thank you all for your input. I will certainly try going back to no-aggregate-device and see how we fare.
Thanks again -
@Jeremy - great info on the limit bandwidth suggestion - thank you.

Reply all

Reply to author

Forward