The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Help with strange Firewire streaming issue... Audio Interfaces
Old 2nd July 2016
  #1
Gear Maniac
 
WalkoftheEarth's Avatar
 

Help with strange Firewire streaming issue...

Hello. So I've been on this adventure of insanity trying to find and fix the cause of this problem. I will do my best to explain, but first, Here is the setup -
PC: ASUS H87 motherboard, intel i7, Seasonic FL 520 platinum power supply, windows 7 x64, 16g Ram, FW Ti chipset, 2 SSDs (1 for operating, second for audio files)

DAW: Nuendo 7

Interface: Prism Orpheus via Firewire

The issue that i am having *seems* to be some sort of limit of bandwidth coming from the computer. If I stream a single stereo (2 channel) track out of the DAW, all seems ok enough. If I add output busses and stream up to 8 channels out, to an external summing mixer, the sound degrades and gets worse with each added stereo output channel. The sound can be described as slightly grainy & harsh, with limited high and low end

Here is what I have tried differently, with the same results - x86 vs x64, different firewire pcie cards, firewire cables, hdd vs ssd, different Orpheus interfaces, BIOS configurations, endless amounts of "tweaks", amongst some other things.

What I have not tried yet, is -
a different motherboard (thinking of bandwidth limitation), different power supply (perhaps I need 1000w or more for transient demand??) ,and probably some other things that I haven't even thought of thought of yet.

If anyone has any clue as to what could be causing this sort of issue, please let me know what you think. It would be greatly appreciated. Thanks! If you need me to clarify anything, please ask.

Last edited by WalkoftheEarth; 2nd July 2016 at 10:49 PM..
Old 2nd July 2016
  #2
Hmmm...this is above me, and it looks like that's a nice expensive interface...are you monitoring through the summing mixer? I ask because when I monitor my 3 slave DAW PC's together, via a mixing board (which is my Tascam FW-1082 interface/mixer/controller on a 4th PC) it can have a tendency to sound like the audio is degraded. The 1st time I experienced this, it was very bad. It turned out that, within my receiving mixer, I wasn't panning correctly hard left/right. Once I corrected the panning, it sounded fine. This, in addition to the fact, that this Tascam FW-1082, that I was using as my monitor mixer, was receiving 'analog' stereo audio from 3 other DAW's with Aardvark Q10 audio cards. As opposed to listening directly from a DAW software program & that PC's attached Q10 interface. In part, having more in my signal chain to to the point of my monitoring stage can make a difference.
Old 3rd July 2016
  #3
Gear Addict
 

Have you tried the Ieee1394 " Legacy " driver ?
The windows firewire driver changed for win 7 , 8 & 10 , & many older devices recommend installing the Legacy driver into the firewire device port.
It's a little tricky to install as Windows prefers to use most recent driver so you have to force it to install "1394 OHCI Compliant Host Controller ( Legacy) "
version .

driver should be here :
https://support.microsoft.com/en-us/kb/2970191

instructions from focusrite website attached
https://support.focusrite.com/hc/en-...-#LegacyDriver
Old 3rd July 2016
  #4
Gear Maniac
 
WalkoftheEarth's Avatar
 

Thanks for the input Subhertz. I have tried the legacy drivers also. Same result.

Steve - thanks also for your input. I am monitoring through the summing mixer. Although what you describe doesn't seem applicable in my case, it has given me some other thoughts to consider.

Still hunting...
Old 3rd July 2016
  #5
Lives for gear
 

I've never heard anything like that..

However, I always tend to separate the "feel" thing from technical issues, which are necessary to isolate. Although this isn't always simple to do, especially if there isn't other equipment or second DAW available.

First let's assume there's nothing wrong with summing and analog connection part of the setup.. as simple it can sound, for example one channel in stereo pair with reversed polarity can do really mess when summed together with other stems. Also there can be some issue with particular inputs at summing mixer, without generator this can be discovered by sequential reconnecting of the same output signal (verified good output signal let's say from channel 1) to remaining inputs at mixer. You can listen to that or record mixer output and then compare it at DAW.

Next let's say all D/A converters and analog outputs on the interface are fine. Again.. in ideal case it's possible to measure that using analyzer, however for the sake of improvised test you can do also some analog loopback with software signal generator and FFT plugin (like Span) at your DAW.. then try for example output 1 to input 1 at interface.. If that would be good, you can repeat that loopback procedure for other analog outputs. You should be able to check most apparent problems with noise, level or distortion.

After that you can focus on data transfer related errors with your interface.. Generally those errors are more common at shorter ASIO buffers, where it can be seen like short bursts of repeated samples or glitches.. this is unlike according to your description, but before further testing I would relax your ASIO buffer settings (to lets say 512s) to exclude another possible problem.
Then I would try to do typical audio data integrity test using digital loopback cable.. Orpheus has ADAT lightpipe I/O, so its pretty good option for that. It should be possible to connect ADAT output to input. That way you can stream 8 channels from DAW to your interface and record it back to another tracks.
If there will be any issues with data integrity during transfer (drivers, firewire, interface..), recorded audio data simply won't null. There is one catch tough.. after recording its usually necessary to sample align source and recorded audio clips.. So it makes sense to start test signal with some short audio peak, which can be used for that alignment in DAW. After alignment procedure and cutting of everything to same length, it is also possible to export all source and recorded tracks to wavs and compare using some external utility (eg. there is plugin for foobar2000, which can do audio comparisons foobar2000: Components Repository - Binary Comparator).. this is handy if you're doing some longer tests.
If you possibly verify, everything matches, then it's not problem of interface, ASIO, computer, bandwidth etc. And it can be also assumed with high percent of probability, if digital I/Os are bit-perfect, then also analog I/O won't be affected by transfer errors.. more tracks or similar voodoo.

Described way is general and bit labour intensive approach to check that.. I have used also some own methods with second "lab" DAW, where I've recorded my special audio files with particular patterns, which were analzyed in real-time in my custom software, but it's not generally usable of course.
However some good vendors (RME) has drivers, which reports statistics of all such transfer related errors etc. This could be also implemented as a special proprietary vendor utility, which would fetch such statistics.. maybe Prism support has also something like that.
I would definitely contact their support about that, if you'd be interested with ruling out transfer issues.

Michal
Old 4th July 2016
  #6
Gear Maniac
 
WalkoftheEarth's Avatar
 

Thanks for the input Michal. Print backs do not null, even when aligned . And in the null, it sounds very digitally distorted. I am almost positive that it is within the CPU....
Old 4th July 2016
  #7
Lives for gear
 

Quote:
Originally Posted by WalkoftheEarth View Post
Thanks for the input Michal. Print backs do not null, even when aligned . And in the null, it sounds very digitally distorted. I am almost positive that it is within the CPU....
That's not good news.. at least you've narrowed possible cause of the problem.
Well then there can be some problem with syncing.. assuming you operate it primarily in local mode.
Clocking issues can be sometimes very tricky.. I've experienced during one problem debugging with some other rig, that ext syncing problems caused intermittent glitches just at last pair of ADAT input all other inputs and channels were completely fine.
Of course, then your Orpheus can be simply broken.. Do you have some chance to try it with different computer (some friend's old MacBook)? Or do you have some response from Prismsound support (eg. some diagnostic steps, utility etc.)?

Michal
Old 5th July 2016
  #8
Gear Maniac
 
WalkoftheEarth's Avatar
 

Im actually testing a second Orpheus right now. Same results. Looking deep into windows mmcs settings. It has to be some sort of priority issue or something with I/O of the CPU. Someone in the world knows all about it im sure, but not this guy. haha If I ever figure this thing out I will publish a book.
Old 5th July 2016
  #9
Gear Maniac
 
WalkoftheEarth's Avatar
 

...one tip off that Ive had, although incredibly hard to discern, is that resetting the firewire bus via unplugging the unit and replugging it, proves to make an immediate improvement. But then within minutes (or less) it gets grainy again. Mind you, this is incredibly subtle and if it wasn't for ridiculous amount of prints on test CDs for outside listening, I would be convinced it's just in my head.

Is there some sort of software to check if the I/O is "bit perfect" as they say?
Old 5th July 2016
  #10
Lives for gear
 
foamboy's Avatar
Have you verified that the summing mixer isn't the culprit?

I had a similar deal and discovered that my mixer was finally crapping out.

Just a suggestion.

fb
Old 5th July 2016
  #11
Lives for gear
 

Quote:
Originally Posted by WalkoftheEarth View Post
Im actually testing a second Orpheus right now. Same results. Looking deep into windows mmcs settings. It has to be some sort of priority issue or something with I/O of the CPU. Someone in the world knows all about it im sure, but not this guy. haha If I ever figure this thing out I will publish a book.
What about other Orpheus' users, I've never heard from anyone, incl. mastering guys, that any mentioned system tweak is really necessary for bit-perfect transfer.. Or adjustment of system priories would change data.
Other thing is, that streaming audio doesn't really put too much stress on current computers, that some special extra special setup is needed.. (of course if you want to run at really low latencies, it's another story, but as I've mentioned previously.. it's advisable to run such integrity tests with long ASIO buffers to exclude this kind of problem).

MMCSS is special scheduler and isn't really meant for tweaking by the user.. Default settings of its classes is good.
Programmer of the particular application (eg. DAW) either use it or not.. If yes, then it means, priority of particular thread (like ASIO I/O thread) within application is automatically managed by operating system to prioritize it over other threads and processes.

Quote:
Originally Posted by WalkoftheEarth View Post
...one tip off that Ive had, although incredibly hard to discern, is that resetting the firewire bus via unplugging the unit and replugging it, proves to make an immediate improvement. But then within minutes (or less) it gets grainy again. Mind you, this is incredibly subtle and if it wasn't for ridiculous amount of prints on test CDs for outside listening, I would be convinced it's just in my head.
I highly discourage you from hot-plugging of firewire cable.. just don't do that .
Although in theory and standard, it's hot pluggable.. One of disadvantages of this bus is susceptibility of this physical connection to damage by ESD (electrostatic discharge) and voltage spikes after such plugging.
I've seen so many different damaged firewire devices from camcorders, interfaces, hardrives to adapters in computers by this, that it's even funny.
That's why also vendors later recommend to plug it before computer startup.

Michal
Old 5th July 2016
  #12
Lives for gear
 

Quote:
Is there some sort of software to check if the I/O is "bit perfect" as they say?
If you followed that digital "nulling" procedure, I've mentioned.. you've just did bit-perfect test.
You said, it won't null, so if I assume, it was properly done, it means there is really problem with audio transfers to and from the interface. When everything works as intended, files played through digital loopback completely nulls up to last bit.

From my point, this is the culprit and this is what you should concentrate with next procedures and mainly during communication with Prismsound support, they should definitely walk you through some troubleshooting steps..
It's very hard to convince or talk to someone, some output channels are very slightly grainy, or it's necessary to do crazy system tweaks, others don't do..
but digital I/O should be bit-perfect, which isn't at your case.

My previous hint with another computer was meant to complete exclude of influences and variables of your system. Maybe also your Prism dealer could help with that.. loopback test with ADAT cable is easy to do even with trial version of Reaper.

Aside from that, if you'd be affected by problems with short latencies or your system won't keep up with load.. DAW application should "know" that and some applications report that as underrun or stops playback. Possibly you'll hear that as intermittent glitches. This kind of issue definitely shouldn't exhibit as continuous data corruption.

Michal
Top Mentioned Products
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump