[App_rpt-users] First experience with ACID app_rpt & usbradio audio popping problem.
Stephen Rodgers
sales at qrvc.com
Wed Nov 18 15:14:33 UTC 2009
Steve Gladden wrote:
> Jim Duuuude wrote:
>> if you dont mind, I would like to ask a completely left-field question.
>>
>> What kind of ethernet hardware do you have on your system, and what kind of
>> bus controller (northbridge, etc). Back in the really old days when we first
>> were using Zapata Telephony cards, we found a number of drivers that had, well
>> lets say, some serious issues in Linux.
>>
>> Thanks for humoring me.
>>
>> JIM WB6NIL
>>
>>
>> ----------------------------------------
>>
>>> Date: Tue, 17 Nov 2009 14:27:01 -0500
>>> From: steve at michiganbroadband.com
>>> To: sales at qrvc.com
>>> CC: app_rpt-users at qrvc.com
>>> Subject: Re: [App_rpt-users] First experience with ACID app_rpt & usbradio audio popping problem.
>>>
>>> Stephen Rodgers wrote:
>>>
>>>> Steve Gladden wrote:
>>>>
>>>>
>>>>> Hi David.. it's not a terribly loud popping it's right with the normal audio
>>>>> levels
>>>>> It really sounds like 'holes' in the processing small abrupt dropouts of the
>>>>> stream..
>>>>> Very random in timing maybe 5-15 hits per minute when I key the radio with
>>>>> PL..
>>>>>
>>>>> Seems to more frequently happen on/around the CWID but still is happening
>>>>> very randomly.
>>>>>
>>>>> CPU utilization is ~10% while radio keyed producing PL and CWID..
>>>>> Have radio continuously keyed (no actual radio) and CDID beacon going every
>>>>> minute.
>>>>>
>>>>> 5-15 'dropouts' per minute.
>>>>>
>>>>> I've tried to isolate it by throwing different tasks at the hardware/CPU to
>>>>> see if the audio dropouts
>>>>> Would increase or not during heavy CPU utilization or maxing out IO on
>>>>> network hard drives etc.
>>>>>
>>>>> Checked alsa mixer to make sure there were no open turned way up audio
>>>>> inputs or mic monitor turned on..
>>>>> Turned down all input levels and this had no effect on the dropouts either.
>>>>>
>>>>> One thing I have not tried (but will not matter) is to hook a usb HDD to the
>>>>> usb port and give the usb controller
>>>>> Itself some activity besides the usbradio port and see if this impacts the
>>>>> issue.
>>>>>
>>>>> This might tend to point me back to the usb controller itself having an
>>>>> issue..
>>>>>
>>>>> I should also try a PCI external USB card on the motherboard and give it
>>>>> it's own interrupt.
>>>>>
>>>>> :-)
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: David McGough [mailto:kb4fxc at inttek.net]
>>>>> Sent: Thursday, November 12, 2009 6:57 PM
>>>>> To: Steve Gladden
>>>>> Cc: app_rpt-users at qrvc.com
>>>>> Subject: Re: [App_rpt-users] First experience with ACID app_rpt & usbradio
>>>>> audio popping problem.
>>>>>
>>>>>
>>>>> Hi Steve,
>>>>>
>>>>> What exactly do you mean by "popping"?? Is it a loud POP! like a drum or
>>>>> is it a momentary audio drop-out or choppiness??
>>>>>
>>>>> The reason I ask about choppiness is that I've experienced this issue.
>>>>> But, I've never had loud POPS! ....
>>>>>
>>>>> 73, David kb4fxc
>>>>>
>>>>>
>>>>>
>>>>> On Wed, 11 Nov 2009, Steve Gladden wrote:
>>>>>
>>>>>
>>>>>
>>>>>> Short version:
>>>>>>
>>>>>> *Installed ACID
>>>>>> *Have audio popping problem with USB transmit (output) audio.
>>>>>> *I'm not surprised didn't expect it to work perfectly first attempt.
>>>>>> *Only tried one combinarion of hardware/cpu/motherboard so far.. I need
>>>>>> to try other motherboards/CPUs.
>>>>>> * Tried 3 different USB devices and ports.
>>>>>> * played with moving interrupts disabling motehrboard built-on hardware
>>>>>>
>>>>>>
>>>>> etc.
>>>>>
>>>>>
>>>>>> * CPU runs at about 10% while USB radio is active..
>>>>>> * Loading PC heavily by running gzip -9 seti at home Lots of disk access
>>>>>> lots of network activity seem to have zero impact on the popping
>>>>>> frequentness.
>>>>>>
>>>>>> * I'd like to know if anyone else here has had this or similar problems
>>>>>> with consistent audio from the USB device (CM-108) and what they may
>>>>>> have done
>>>>>> to make it work better.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Long Version:
>>>>>>
>>>>>> I gave it a go and went the 'easy' route just to get started..
>>>>>> I ended up trying it out on that 1.1Ghz AMD Duron CPU rack mount server..
>>>>>> I must say I'm extremely excited & please with this software..
>>>>>> But it would be too good to be true if it all just worked first try..
>>>>>> and of course would take all the 'fun'
>>>>>> out of getting it to work..
>>>>>> Ok enough small talk (haha)
>>>>>> Here's what I have going on:
>>>>>>
>>>>>> Everything works except I am getting intermittent audio popping in the
>>>>>> audio output of the USB sound device.
>>>>>> VERY intermittent maybe about 3-4 'pops' every minute on average.. 10 at
>>>>>> most sometimes.. and other times several minutes go by without a 'pop'.
>>>>>> It seems to mostly occur during the CW ID beacon which I have going
>>>>>> every minute..
>>>>>> The pops that are heard only happen when the 'transmitter' is active
>>>>>> sending PL or CW ID or PLAYBACK...
>>>>>> My test arrangement has "radio key" running to produce continuous PL and
>>>>>> occasional audio popping.
>>>>>> There is no radio actually hooked up yet.
>>>>>>
>>>>>> OK will talk about the hardware a bit.. and probably worse case scenario
>>>>>> here due to an 'all in one' motherboard..
>>>>>> It's a MSI 6378 'budget' Athlon board.. everything built on.
>>>>>> Has USB 1.1 *NOT 2*
>>>>>> I don't know/think this matters for a sound device.
>>>>>>
>>>>>> Here's all of the things I have tried all of which have absolutely no
>>>>>> effect on the intermittent popping...
>>>>>>
>>>>>> 1. Disabling everything not needed on motherboard including ethernet for
>>>>>> test..
>>>>>> ethernet + usb share an interrup line and there is no ability to
>>>>>> re-route this in the bios
>>>>>> so I turned off the ethernet on the motherboard for testing.
>>>>>>
>>>>>> 2. Carefully monitor CPU during audio pops running 'top' around 20
>>>>>> updates per second (this alone takes more CPU load than asterisk
>>>>>> /app_rpt + usb radio.
>>>>>>
>>>>>>
>>>>>> 3. On this motherboard the USB controller and ethernet share a PCI
>>>>>> interuupt line and you cannot change it.
>>>>>> I've tried disabling the built on ethernet and still get the popping..
>>>>>> I've also flood and large pinged the machine which genreates huge
>>>>>> numbers of interrups from the ethernet chip and this has no effect on
>>>>>> the popping..
>>>>>>
>>>>>> 4. Tried other USB fobs with same chipset.
>>>>>> Created various types of heavy system activity to see if it has any
>>>>>> impact on the popping heavy disk use hevy network use and heavy vga use
>>>>>> (text mode)
>>>>>> moving interrupts (changing them) on the motherboard.. getting usb
>>>>>> controller onto a different interuupt (ethernet ends up on same).
>>>>>>
>>>>>> 5. Measured and scoped 5V output on USB to be sure 'popping' sound is
>>>>>> not coming from 5V power supply.
>>>>>>
>>>>>> 6. Am really looking to hear what some others experience has been with
>>>>>> this..
>>>>>> what works and what does not?
>>>>>> The popping very much sounds like an interrupt issue or some kind of
>>>>>> issue where the task of sending and audio stream is completely dropped
>>>>>> for something else
>>>>>> and then resumed leaving lots of holes heard as pops in the output stream.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>> This is reported occasionally, but I've never experienced it.
>>>>
>>>> Given what you've tried already, I'd try a different motherboard.
>>>>
>>>> Also, remove any usb hobs if you are using one.
>>>>
>>>>
>>>> Steve
>>>> WA6ZFT
>>>>
>>>>
>>>>
>>>>
>>> Hello David... I *may* also be on to something here as I've been able to
>>> also determine that the audio pops are occuring ALWAYS
>>> While app_rpt is doing it's statpost routine! updating stats via the web.
>>>
>>> I set te statpost address to something local on the LAN for now just for
>>> testing but I get the same audio pops whether just on the
>>> LAN or if it's statposting out to the internet.
>>>
>>> The popping definitely happening at the time that app_rpt is sending
>>> HTTP out..
>>> The popping never happens during other access to-from the ethernet or
>>> during IAX interchanges between asterisk and the world..
>>> Seems to be only when app_rpt is stat posting...
>>> Dave I think you are certainl correct about the issue being within app_rpt.
>>> I'm sure the stat posting is not the only issue but I can duplicate the
>>> problem 100% by seeing the popping every time app_rpt
>>> sends out HTTP. :-)
>>>
>>> You may have already found that app_rpt does the 'droppouts' with a ton
>>> of other activities... not just stat posting like I am currently
>>> pointing out...
>>>
>>> Might be a start to something though... since it's duplicatable here. :-)
>>>
>>> Note: I've not figured out how to manually make it statpost.. I just
>>> turend up the debug to 7 and waited for it to say it was statposting
>>> also looking at the ethernet and see the HTTP packets go out at same
>>> time dropous are ocurring.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> --
>>> This message has been scanned for viruses and
>>> dangerous content by MailScanner, and is
>>> believed to be clean.
>>>
>>> _______________________________________________
>>> App_rpt-users mailing list
>>> App_rpt-users at qrvc.com
>>> http://qrvc.com/mailman/listinfo/app_rpt-users
>>>
>>
>>
> I'm on test machine #4 now this one is a P3 @450Mhz on an Asus P2-99
> Motherboard.. no built on video audio or ethernet..
> A bit of a slower CPU on this one the audio dropouts are about the same
> (just as bad no worse) Turning off the stat post
> by commenting out the two lines makes a HUGE difference for the better
> but still not perfect as there is still some dropouts..
> This is totally stock off the ACID install after all updates & setup.
>
> I still (badly) need to try David's suggested source and re-compile
> fixes and see what I get.
> I've certainly found that the stat post part of app_rpt is really making
> the problem show up when it is active.
> Besides needing to apply & try David's fixes I am still at the same time
> trying this on a variety of different hardware..
> So far ACID out of the box has had the same issue of 4 different systems.
>
> On this last attempt with a P2-99 I've gotten away from the VIA chipsets
> and pretty much all of the on motherboard problem/frustration causers.
> This Motherboard uses Intel NB and SB.. A copy of the manual for this
> one is at: http://stuff.michiganbroadband.com/p2-99.pdf
>
> Cheers!
> Steve
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
All the machines you have tried except one are sub 1GHz. And the 1.1GHz machine you were using is a Duron processor
which is the low end of the low end! In our documentation we specify a machine with a clock speed of 1GHz or better. The
DSP routines in chan_usbradio/XPMR may be getting bogged down for an instant during the fork or ethernet packet
transmission of the statpost info. We have not had much luck running ACID on machines under 1GHz and we have not tested
it on Celeron or Duron processors.
There are lots of users out there using ACID, but I suspect if you ask them, they are running processors north of 1GHz
and that most of them are running later vintage processors from Intel and AMD.
If you want to continue using old motherboards with slower clock speeds and old low end processors, then you have to
expect these kinds of issues.
Steve
WA6ZFT
More information about the App_rpt-users
mailing list