OP25 Distorted Audio

[j_mar]

Hi All!

I'm having the same issue as this thread:

OP25 Audio Slow and Choppy

I'm using a brand new RPI 4b (1gig) and new NooElec NESDR smart v4 running raspbian buster. I am trying to listen to the North Fulton Regional Radio System (NFRRSA). My first setup had audio that sounded terrible, so I tried the solution from the original thread by getting a brand new class 10 SD card and wrote a new raspbian buster image (didn't copy the old one in case of user errors). Got everything set back up using John's Tech Blog instructions and am having the same problem.

My audio sounds like this while running op25 through VNC. I've got the speaker hooked up directly via cable to the pi and not bluetooth. The audio plays perfectly with the same setup over VNC while in gqrx or youtube. My gut tells me I'm doing something wrong on my end or that there's some driver issue going on but I don't have the knowledge to troubleshoot it further.

My setup does lock onto the control channel and pull in audio. Here's my err log and trunk file (I've tried both CQPSK and C4FM).

The command to launch op25 looks like this: ./rx.py --args 'rtl' -N 'LNA:47' -S 2500000 -x 2 -f 771.98125e6 -o 17e3 -T trunk.tsv -V -2 -U 2> stderr-stream0.2

The only other thing I noticed is that htop shows a lot of instances of rx.py running but I'm not sure if that's normal or something funky going on.

Thanks in advance for any help as I'm stumped!

 

[boatbod]

If you are using a regular RTL device reduce your sample rate to -S 960000 as there is no need to run wide open and doing so can cause undesirable behavior.
Next, please post screen shots of your mixer plot (#5), datascope plot (#4) and constellation plot (#2)
Lastly, try adding -O "pulse" and let me know if this helps or makes things worse.

 

[j_mar]

If you are using a regular RTL device reduce your sample rate to -S 960000 as there is no need to run wide open and doing so can cause undesirable behavior.
Next, please post screen shots of your mixer plot (#5), datascope plot (#4) and constellation plot (#2)
Lastly, try adding -O "pulse" and let me know if this helps or makes things worse.

1. No noticible change when reducing the sample rate down.
2. Adding pulse doesn't seem to make much of a difference. Maybe, if anything, it's a tad better but it could be me wanting an improvement.
3. Mixer Plot, Datascope, and Constellation

Thank you boatbod for helping me out on this.

 

[wgbecks]

I suspect that if you have a speaker plugged into the Raspberry Pi's headphone jack that the distorted audio is due to the fact the the Pi uses a four (4) pole jack that has analog video on one of the poles. Take a look at the plug at the end of the cable from you speakers. I am willing to bet that it has three (3) pole (T,R,S) type plug.

You'll need a 3-pole to 4-pole adapter to prevent the analog video signal from bridging to the audio channel that's responsible for the noise or distortion that you're hearing. There are several variations of such an adapter. The Sennheiser PCV 05 Combo Audio Adapter is one such device that is high quality and works very well for this application.

You might also be interested in setting up your own private streaming audio server. See the README-rpi3-liquidsoap in the ~/op25 folder.

Bill

 

[j_mar]

I suspect that if you have a speaker plugged into the Raspberry Pi's headphone jack that the distorted audio is due to the fact the the Pi uses a four (4) pole jack that has analog video on one of the poles. Take a look at the plug at the end of the cable from you speakers. I am willing to bet that it has three (3) pole (T,R,S) type plug.

You'll need a 3-pole to 4-pole adapter to prevent the analog video signal from bridging to the audio channel that's responsible for the noise or distortion that you're hearing. There are several variations of such an adapter. The Sennheiser PCV 05 Combo Audio Adapter is one such device that is high quality and works very well for this application.

You might also be interested in setting up your own private streaming audio server. See the README-rpi3-liquidsoap in the ~/op25 folder.

Bill

Bill-
Thanks for the reply.... you were spot on! I was using a 3 pole cable. I did a sort of bootleg solution since I had a USB soundcard soundcard laying around. I tried running the audio out from the pi into the headphone jack in on the USB card and another cable from the out on the soundcard to the speaker and now everything is crystal clear. I guess the soundcard can decode the signal properly. I'm definitely going to go the liquidsoap route now that I know it's all working.

Thanks so much Bill and boatbod for the help.

 

[boatbod]

Your plots are interesting. Good signal strength in the mixer plot, but look at all the smearing of the constellation plot (more of an X rather than 4 distinct clusters in the corners of a square). Probably simulcast/multipath distortion in action. The datascope also seems to be having a hard time settling down. Was this taken while voice was active or idle on the control channel?

 

[wgbecks]

It might be helpful to see the FFT plot. I just looked at your other plots and agree it looks like simulcast/multipath distortion. The FFT plot will reveal if by chance the SDR is saturated with strong signals that result inn IM products that beat up the desired signal and it might just be a simple matter of reducing LNA gain. In the event its not overload then relocating the antenna might be helpful in finding a spot that receives fewer reflections thus cleaning up the recovered data.

Bill

 

[j_mar]

boatbod & Bill-

Even though the audio issue is better, I am not curious about the signal. I redid the plots HERE. I intentionally let it run during silence and with voice traffic on each of the plots.

As an experiement I did some plots for another system also during silence and with voice traffic. HERE they are. FWIW it seems that this one locks onto talkgroups much faster and seems more active within op25. I am new to these systems so it could just me way more active than the original one [which is likely and due to the amount of departments on the same system] and nothing related to anthing signal related.

 

[boatbod]

boatbod & Bill-

Even though the audio issue is better, I am not curious about the signal. I redid the plots HERE. I intentionally let it run during silence and with voice traffic on each of the plots.

As an experiement I did some plots for another system also during silence and with voice traffic. HERE they are. FWIW it seems that this one locks onto talkgroups much faster and seems more active within op25. I am new to these systems so it could just me way more active than the original one [which is likely and due to the amount of departments on the same system] and nothing related to anthing signal related.

The plots from the second system are coming in nice and clean. The first... not so much!
By any chance is the first system simulcast and the second stand-alone?
On my local P25 simulcast system I see the same smeared constellation plot on the control channel, but the voice channels lock it really cleanly. Interesting as that may be, I unfortunately have no explanation or means of improvement. Try moving the antenna around a little. Sometimes just a few inches or feet can make a significant impact.

 

[j_mar]

The plots from the second system are coming in nice and clean. The first... not so much!
By any chance is the first system simulcast and the second stand-alone?
On my local P25 simulcast system I see the same smeared constellation plot on the control channel, but the voice channels lock it really cleanly. Interesting as that may be, I unfortunately have no explanation or means of improvement. Try moving the antenna around a little. Sometimes just a few inches or feet can make a significant impact.

They are both listed as simulcast systems on RR, but I don't know first hand for sure. I moved the antenna (using a smaller UHF antenna) to a different part of the house and reran the plots on the original system. To my amateur eye it maybe looks a little better.

New Plots

 

[wgbecks]

Still some multipath but looking much better with fewer errors. Do you have an outdoor scanner antenna that you can try? What does the terrain look like surrounding your location an those of the nearest sites? IE: Hilly, flat? You may have to resort to a cheap Yagi antenna aimed
directly at the site to get this under control.

A good friend of mine had a similar issue where he is located along an inland lake and where the terrain in his area is somewhat rugged that was causing frequent periods of multipath to the point that the errors were severe enough to disrupt reception. He ended up installing an outdoor
Yagi (beam) antenna and disconnection the op25 SDR from the outdoor scanner antenna at the same elevation above ground to get the multipath under control.

Bill

 

[j_mar]

Still some multipath but looking much better with fewer errors. Do you have an outdoor scanner antenna that you can try? What does the terrain look like surrounding your location an those of the nearest sites? IE: Hilly, flat? You may have to resort to a cheap Yagi antenna aimed
directly at the site to get this under control.

A good friend of mine had a similar issue where he is located along an inland lake and where the terrain in his area is somewhat rugged that was causing frequent periods of multipath to the point that the errors were severe enough to disrupt reception. He ended up installing an outdoor
Yagi (beam) antenna and disconnection the op25 SDR from the outdoor scanner antenna at the same elevation above ground to get the multipath under control.

Bill

So I'm under 5 miles to the closest site and around 10-15miles to most of the other sites in the simulcast. It's mostly flat to the close site, just some gentle hills. There's some woody areas but nothing really dense and a bunch of roads in between me and the site. Do you think my proximity to multiple simulcast locations can be a contributing factor?

I don't have an external antenna but can probably get my hands on one and see if it helps. Otherwise a yagi may not be a terrible idea! I have a Ham Radio Outlet nearby and can see what they may have in stock or can go the Amazon route.

 

[boatbod]

So I'm under 5 miles to the closest site and around 10-15miles to most of the other sites in the simulcast. It's mostly flat to the close site, just some gentle hills. There's some woody areas but nothing really dense and a bunch of roads in between me and the site. Do you think my proximity to multiple simulcast locations can be a contributing factor?

I don't have an external antenna but can probably get my hands on one and see if it helps. Otherwise a yagi may not be a terrible idea! I have a Ham Radio Outlet nearby and can see what they may have in stock or can go the Amazon route.

Have you tried turning your LNA gain down?

 

[j_mar]

Have you tried turning your LNA gain down?

I know you suggested that earlier in the thread but I haven’t tried it yet. Is there a range that might be better or should I try to start dropping the number and look for better looking plots?

 

[boatbod]

I know you suggested that earlier in the thread but I haven’t tried it yet. Is there a range that might be better or should I try to start dropping the number and look for better looking plots?

I'd suggest starting in the mid-30's and work from there downward to see if you can find a sweet spot. Ultimately a directional antenna might be the only solution, but lower gain comes at no-cost!

 

[j_mar]

I'd suggest starting in the mid-30's and work from there downward to see if you can find a sweet spot. Ultimately a directional antenna might be the only solution, but lower gain comes at no-cost!

Awesome, I'm going to try that and report back. I'd much prefer a free solution!

 

[j_mar]

I've been messing around with the LNA but am not really noticing a difference as of yet. I've gotten to the high 20's so far. I did finally get the liquidsoap/icecast2 setup going (for some reason it took me a few tries) and now there's a weird thing going on where op25 doesn't seem to pick up on a signal as quickly or as well when I'm running liquidsoap. It'll work for a bit and then just silence but will seem to work well again if I close everything out and restart both scripts. It's probably a config issue so I'll work on getting that squared away and keep experimenting with the LNA.