[arm-allstar] software update status

Wed Sep 2 12:21:54 EST 2015

Hi LaRoy,

I hesitate to speculate as to why the variable audio delay is more
prevalent on the Rpi2, vs. other platforms. So, what I'm suggesting below
is speculation.

One factor -may- be that on the BBB, while single core, that core is
significantly faster (for integer performance)  than a single RPi2
core....Other possibilities include the multi-processor environment of the
RPi2 is massively more complex at the OS level than single-processor. Of
course, the USB subsystem on the RPi2 is very different, too, etc., etc.

I've also had some comments from other experienced developers suggesting
that the Linux CPU clock speed scaling code (in the kernel) is the root
cause and that hand-cooking a kernel with scaling completely removed fixed
this.  I will comment that the scaling code was once a contributer to the
issue, I confirmed this myself.  However, in more recent kernel versions
(including the kernel in the RPI2 1.0 ARM-AllStar release), that code
seems to be much better behaved and I've not been able to obviously
duplicate the earlier issues.

Anyhow, the re-sampling method I'm developing now does fix the issue on
all these boards without resorting to kernel changes, etc.

73, David KB4FXC

On Wed, 2 Sep 2015, LaRoy McCann wrote:

> David,
> I am curious as to why I only notice this audio problem on the Pi2 and 
> not the BBB.
> 
> I have a repeater running on each the Pi2 and the BBB and only the Pi2 
> repeater has this issue.
> These two repeaters are linked together thru a wireless link full time.
> I know there is a slight delay on the BBB from input audio to output 
> audio, but I have never been able to un-key fast enough to hear myself 
> on the BBB.
> 
> Just seems odd that a faster processor has the problems.
> I know the processor is a multi-core processor but my guess is that 
> everything still runs on one core and that one core is slower than the BBB.
> 
> 73's
> LaRoy, K5TW
> 
> On 9/1/2015 1:24 AM, David McGough wrote:
> > Hi Bob,
> >
> > The variable audio delay issue is occasionally seen on PC-computer
> > platforms, too, particularly with the simpleusb driver. This is NOT the
> > audio delay that is unavoidable due to buffering as the audio samples pass
> > through Asterisk. Nor is this issue related to the Asterisk jitter buffer,
> > etc.
> >
> > Basically stated, the problem is that the input sample rate -dynamically-
> > varies from the output sample rate. Depending on various factors, you may
> > have too many samples, or too few, at any given moment. With the current
> > code, the main issue seen is that too many output sound frames get queued,
> > causing a frequently increasing in length input to output audio
> > delay--sometimes with delays of maybe as much as one second (really
> > annoying!). If there is an output buffer under-run, you hear a pop/click
> > due to running out of samples.
> >
> > Initially, I suspected some kind of subtle bug in the simpleusb sound
> > code. But, after experimentation, I've been able to easily duplicate this
> > issue using my native ALSA stand-alone test software or even ALSA
> > applications like arecord and aplay, etc.
> >
> > One solution to this problem is to use a dynamic buffer that effectively
> > re-samples the audio, gently up-sampling or down-sampling, depending on
> > the current trends for overrun or under-run.  I've got working software
> > that accomplishes this now. And, it runs full-duplex beautifully for 24+
> > hours with no change in audio delay and no logged clicks or pops.
> >
> > Hopefully, I'll get a few days time when I'm rested to focus on completing
> > the integration of this code into simpleusb.  So far, I've had no time at
> > all to complete this.
> >
> > But, hopefully soon.
> >
> > 73, David KB4FXC
> >
> >
> >
> > On Mon, 31 Aug 2015, kk6ecm wrote:
> >
> >> Couple of questions:
> >>
> >> 1) Is this audio delay the same as experienced with ACID?
> >>
> >> 2) What is the nature of the fix that reduces the delay?
> >>
> >> Thanks,
> >> Bob
> >> kk6ecm
> >>
> >> Sent from iPad
> >>
> >>
> >>> On Aug 27, 2015, at 9:33 AM, David McGough <kb4fxc at inttek.net> wrote:
> >>>
> >>>
> >>> Hi LaRoy,
> >>>
> >>> Progress has been slow in recent weeks, due primarily to family
> >>> commitments that I've got. At this point, the update is waiting on me.
> >>>
> >>> I am hopeful to minimally have a bug-fix interim release soon. The big
> >>> item on the list being resolution for the variable audio delay.
> >>>
> >>> I hope to have more details soon.
> >>>
> >>>
> >>> 73, David KB4FXC
> >>>
> >>>
> >>>
> >>>> On Thu, 27 Aug 2015, LaRoy McCann wrote:
> >>>>
> >>>> Doug,
> >>>> Just curious on how things are progressing with a new software update
> >>>> for the PI2?
> >>>>
> >>>> Thanks,
> >>>>
> >>>> LaRoy McCann, K5TW
> >>>>
> >>>>
> >>>>
> >>>>
> >>>> ---
> >>>> This email has been checked for viruses by Avast antivirus software.
> >>>> https://www.avast.com/antivirus
> >>>>
> >>>> _______________________________________________
> >>>>
> >>>> arm-allstar mailing list
> >>>> arm-allstar at hamvoip.org
> >>>> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
> >>>>
> >>>> Visit the BBB and RPi2 web page - http://hamvoip.org
> >>> _______________________________________________
> >>>
> >>> arm-allstar mailing list
> >>> arm-allstar at hamvoip.org
> >>> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
> >>>
> >>> Visit the BBB and RPi2 web page - http://hamvoip.org
> >> _______________________________________________
> >>
> >> arm-allstar mailing list
> >> arm-allstar at hamvoip.org
> >> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
> >>
> >> Visit the BBB and RPi2 web page - http://hamvoip.org
> >>
> > _______________________________________________
> >
> > arm-allstar mailing list
> > arm-allstar at hamvoip.org
> > http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
> >
> > Visit the BBB and RPi2 web page - http://hamvoip.org
> >
> 
> 
> ---
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
> 
> _______________________________________________
> 
> arm-allstar mailing list
> arm-allstar at hamvoip.org
> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
> 
> Visit the BBB and RPi2 web page - http://hamvoip.org
> 