As the next round of P25 PII systems take hold in my area... OP25 is the goto solution... but...
Audio dying is not a good thing... to send out silence on feeds....
Seems this is related to the metatag update when using -M tags.json option with rx.py...
Whats frustrating is
Pi3B+ with a cheapo junky asian dongle WORKS with -M tags.json and the audio goes on... .. I have another issue not related to that, which a commit fixes, just can't update to handle that right now...
A DUAL CORE laptop with a junk asian dongle... the audio dies after awhile.
SAME DUAL CORE laptop with a V3 Blog dongle... audio dies after awhile
SAME DUAL CORE laptop with V3 blod dongle, REMOVE the -M option and the audio seems to be stable out to loop0!
URRRGGRRRHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH!
The only thing I can seem to possibly pinpoint is my standard A1 whipping or flogging boi,,, puslecrapdio! RURRRRRHHGHHH!
The laptop has it the Pi3 doesn't....
Any one else seeing similar ? ? ? ? ?
Once I have a new image setup for my Pi feed boxes.. and if that works reliably great...it seems to on my existing feed, other issue not witstanding.....
Jusy annoying that a laptop has these issues where as a Pi3B+ doesn't......
I am grasping for answers here... and as I said... my standard A1 whipping or flogging boi,,, puslecrapdio! just gets my attention and foucs as the problem area.
So to wrap up my screed and vent BS post.. .:) ;)
Any seeing similar with xUbuntu's 18.04 or even other distros on lap/desktops v. Pi3B+???
Welp... there goes my theory in the time it took to type this OP25 is now not playing audio out loop0 again, and no -M option was used.... CC channel did rotate... (Harris nee M/A-COM system)??? So maybe CC rotation kills audio ? ? ? ?
Any one with ideas? ? ? ? Bueller????
This is not a unique problem especially with RPi3 and aloop/darkice. My recommended solution is to switch over to liquidsoap which avoids alsa and aloop altogether. Instructions can be found in the README-rpi3-liquidsoap file in the top level op25 directory.
Graham
On Fri, Feb 22, 2019, 7:13 PM op25@zellners.com wrote:
As the next round of P25 PII systems take hold in my area... OP25 is the goto solution... but...
Audio dying is not a good thing... to send out silence on feeds....
Seems this is related to the metatag update when using -M tags.json option with rx.py...
Whats frustrating is
Pi3B+ with a cheapo junky asian dongle WORKS with -M tags.json and the audio goes on... .. I have another issue not related to that, which a commit fixes, just can't update to handle that right now...
A DUAL CORE laptop with a junk asian dongle... the audio dies after awhile.
SAME DUAL CORE laptop with a V3 Blog dongle... audio dies after awhile
SAME DUAL CORE laptop with V3 blod dongle, REMOVE the -M option and the audio seems to be stable out to loop0!
URRRGGRRRHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH!
The only thing I can seem to possibly pinpoint is my standard A1 whipping or flogging boi,,, puslecrapdio! RURRRRRHHGHHH!
The laptop has it the Pi3 doesn't....
Any one else seeing similar ? ? ? ? ?
Once I have a new image setup for my Pi feed boxes.. and if that works reliably great...it seems to on my existing feed, other issue not witstanding.....
Jusy annoying that a laptop has these issues where as a Pi3B+ doesn't......
I am grasping for answers here... and as I said... my standard A1 whipping or flogging boi,,, puslecrapdio! just gets my attention and foucs as the problem area.
So to wrap up my screed and vent BS post.. .:) ;)
Any seeing similar with xUbuntu's 18.04 or even other distros on lap/desktops v. Pi3B+???
Welp... there goes my theory in the time it took to type this OP25 is now not playing audio out loop0 again, and no -M option was used.... CC channel did rotate... (Harris nee M/A-COM system)??? So maybe CC rotation kills audio ? ? ? ?
Any one with ideas? ? ? ? Bueller????
Quoting Graham Norbury gnorbury@bondcar.com:
This is not a unique problem especially with RPi3 and aloop/darkice. My recommended solution is to switch over to liquidsoap which avoids alsa and aloop altogether. Instructions can be found in the README-rpi3-liquidsoap file in the top level op25 directory.
First, MY PI3 op25 is humming along with darkice and audio quite fine... its only issue is when metatag updates fail causing a crash.. thats fixed in a later version, but I can't update that part right now.. listening to it right now after days and days of running....
Second, LAPTOP setup is the issue.....
Thrid, I planned to do a different test today, but a hardware snafu nixed that, till later..
Fourth... I have had the laptop taking audio from VLC output to loop0 and its been playing for hours...this is continous audio... I just took my mp3 NAS and dumped it into VLC! :)
Fifth, I do NOT LIKE liquidsoap. I have my reasons, mainly I KNOW DARKICE, and use it extensively for years.
So with audio flowing non stop locally on the laptop
And audio flowing from a VLC Playlist out aloop0
op25 seems to be the issue...'
Now the the only difference in this is a CHANGE IN SYSTEMS from a local about 3 air miles to a new system coming online, tower is 12 air miles away... I was using op25 to monitor another part of this system that was about 20 air miless away....
The signal is about ~30-31db for each channel, CC or traffic, and audio is clean and crisp.. when I let it only play on local .... I was using the same laptop for the local system and feeding it out darkice/aloop....I just was not monitoring it close to see if it was dying and was futzing with what it was monitoring etc.. so I can't 100% verify that it was pumping out audio 24/7 on that system...
So one thing could be signal strength, but that should only matter in the recovery of the CC and audio. Constellation looks good, eye pattern looks good... steady stream of CC data, even dealing with the rotations.. Harris system.... they have taken their rolling CC from EDACS and added it to their P25 CAI, more so than just normal.. which when monitoring the part of this system system 20 miles away, I saw CC decode issues, missing audio... to be expected.. the system is setup so as not to cover me from that site(s)... its a dual zone setup similar to their existing system... not really sure why keep this in P25.. but considering the vendor is the same, it doesn't surprise me....
I probably will take and revert this back to the local system overnight to see if audio comes out of this thing 24/7 and then proceed from there...
I will most likely try my Pi3 setup with this system when I can set it up... and then see if it works... with dakrice/aloop...
What I am after is getting way to pinpoint where the issue is.. op25, aloop, what ever... Some may have an affinity for liquidsoap, just as I have affinity for darkice.. I use ALSA plugins to act as sort of split to take a single stereo input and send out 2 feeds with out issues..24/7/365 for over 10 years. And like I Said I am listening to the PI3B with darkice/aloop feed out out audio right now.. just a different system.... which I am hoping is not the culprit... Motorola v. Harris?, signal... (Note: I use commercial antennas and LMR cable, the antenna for these systems is 25 feet in the air above local scrub and I actually will put a yagi onto this in the final setup... its on the omni right now
Ultimately this has to run on Pi3b+'s as my setup for op25... and I need to setup probably about 3-4 for this system once its full blown use.. so if I get things to work on the Pi3b+ as a test then I will just go with that... and chalk the laptop upto an anamoly... I would mostly use it for local audio any way..
The laptop was pushed into this due to the users in one group switching sooner than I was planning on them....
Localizing the problem is going to take some detective work. 1. Determine whether op25 is receiving, decoding and outputting audio. Depending how you run your log level, I'd suggest first looking for "voice update" messages which are produced at level "-v 1" (or greater) every time op25 tunes a voice channel. Once that is confirmed, I'd look for occasional "PCM underrun" messages that are occasionally emitted by sockaudio when there are inevitable lost frames or processing delays. You should be able to get a sense of the approximate frequency looking back at the log when audio is known to be good. Sudden absence of these sporadic underruns would suggest either op25 stopped sending pcm or the alsa subsystem had an issue.
2. Compare the behavior when outputting to 'aloop' vs 'default' (speaker) device. In the past when I've seen the sudden absence of audio, it has only affected output to aloop.
3. Consider temporarily testing with the internal audio player disabled (remove the -U and make sure you have -V -w) then run an instance of nc | aplay Generally 'nc' is quite chatty when it's running so you should be able to see if raw udp samples stop arriving at some point.
Let me know what happens. Graham
On 2/24/19 7:34 PM, op25@zellners.com wrote:
Quoting Graham Norbury gnorbury@bondcar.com:
This is not a unique problem especially with RPi3 and aloop/darkice. My recommended solution is to switch over to liquidsoap which avoids alsa and aloop altogether. Instructions can be found in the README-rpi3-liquidsoap file in the top level op25 directory.
First, MY PI3 op25 is humming along with darkice and audio quite fine... its only issue is when metatag updates fail causing a crash.. thats fixed in a later version, but I can't update that part right now.. listening to it right now after days and days of running....
Second, LAPTOP setup is the issue.....
Thrid, I planned to do a different test today, but a hardware snafu nixed that, till later..
Fourth... I have had the laptop taking audio from VLC output to loop0 and its been playing for hours...this is continous audio... I just took my mp3 NAS and dumped it into VLC! :)
Fifth, I do NOT LIKE liquidsoap. I have my reasons, mainly I KNOW DARKICE, and use it extensively for years.
So with audio flowing non stop locally on the laptop
And audio flowing from a VLC Playlist out aloop0
op25 seems to be the issue...'
Now the the only difference in this is a CHANGE IN SYSTEMS from a local about 3 air miles to a new system coming online, tower is 12 air miles away... I was using op25 to monitor another part of this system that was about 20 air miless away....
The signal is about ~30-31db for each channel, CC or traffic, and audio is clean and crisp.. when I let it only play on local .... I was using the same laptop for the local system and feeding it out darkice/aloop....I just was not monitoring it close to see if it was dying and was futzing with what it was monitoring etc.. so I can't 100% verify that it was pumping out audio 24/7 on that system...
So one thing could be signal strength, but that should only matter in the recovery of the CC and audio. Constellation looks good, eye pattern looks good... steady stream of CC data, even dealing with the rotations.. Harris system.... they have taken their rolling CC from EDACS and added it to their P25 CAI, more so than just normal.. which when monitoring the part of this system system 20 miles away, I saw CC decode issues, missing audio... to be expected.. the system is setup so as not to cover me from that site(s)... its a dual zone setup similar to their existing system... not really sure why keep this in P25.. but considering the vendor is the same, it doesn't surprise me....
I probably will take and revert this back to the local system overnight to see if audio comes out of this thing 24/7 and then proceed from there...
I will most likely try my Pi3 setup with this system when I can set it up... and then see if it works... with dakrice/aloop...
What I am after is getting way to pinpoint where the issue is.. op25, aloop, what ever... Some may have an affinity for liquidsoap, just as I have affinity for darkice.. I use ALSA plugins to act as sort of split to take a single stereo input and send out 2 feeds with out issues..24/7/365 for over 10 years. And like I Said I am listening to the PI3B with darkice/aloop feed out out audio right now.. just a different system.... which I am hoping is not the culprit... Motorola v. Harris?, signal... (Note: I use commercial antennas and LMR cable, the antenna for these systems is 25 feet in the air above local scrub and I actually will put a yagi onto this in the final setup... its on the omni right now
Ultimately this has to run on Pi3b+'s as my setup for op25... and I need to setup probably about 3-4 for this system once its full blown use.. so if I get things to work on the Pi3b+ as a test then I will just go with that... and chalk the laptop upto an anamoly... I would mostly use it for local audio any way..
The laptop was pushed into this due to the users in one group switching sooner than I was planning on them....
-
Graham Norbury -
op25＠zellners.com