DSDPlus FMP24 and NXDN96 - choppy audio

[NPC411]

Hi, everyone! A few months ago, you were all so kind to help with my dual-dongle decoding of NXDN96 via DSDPlus Fastlane and FMP24. Since then, I have moved my setup to a new location with an outdoor-mounted antenna, which has resulted in some choppy audio. While you can still understand the majority of what is being said, you have to listen extremely close and fight the choppiness.

I have attached screenshots of my FMP24 CC and FMP24 VC so you can take a look at my current signal. I have read some about adjusting the gain, but I have not found any clear and precise instructions on where to adjust it, how to adjust it, and to what extent. I look forward to any advice you may have to offer!

FMP24 CC:


FMP24 VC:

 

[wx5uif]

Your gain is way high. Click on the FMP title bar to select that window then press g to lower gain or shift g to raise.

 

[R0am3r]

In addition to manually changing the gain with 'g' or 'G', you can also set the gain via the command line or in a batch file. Example:
FMP24.exe -o20001 -P0.0 -f150.8675 -b4 -g14.4

 

[NPC411]

Your gain is way high. Click on the FMP title bar to select that window then press g to lower gain or shift g to raise.

Thanks, @wx5uif and @R0am3r ! Once I figure out the right amount of gain, I will add the gain control to the batch files. Do I need to adjust the gain in both my CC and VC, and how much should I lower it?

 

[dave3825]

You want to lower both gains. Click the title bar of the cc tuner to focus, then hit g to lowerer it all the way. Then hit G (shift +g) repeatedly while watching the signal. You basically want to raise the signal to where the noise floor just starts to rise. Then click the voice tuner title bar, and repeat the process.

 

[NPC411]

You want to lower both gains. Click the title bar of the cc tuner to focus, then hit g to lowerer it all the way. Then hit G (shift +g) repeatedly while watching the signal. You basically want to raise the signal to where the noise floor just starts to rise. Then click the voice tuner title bar, and repeat the process.

Thanks for your help, @dave3825 ! I adjusted the gain way down for both CC and VC, and things seemed to improve to a degree. However, I am still getting choppy audio at times. I slightly adjusted the PPM offset for both dongles against the local NOAA station (I realize that RTL-SDR V3 dongles need very little PPM offset anyway), but that didn't make a lot of difference. Another post here mentioned to adjust the sampling rate and FFT size for FMP24.cfg; do you know what either of those do and what their setting should be for NXDN96?

 

[NPC411]

Bump for visibility. For FMP24.cfg, what should the sampling rate and FFT size be set to when NXDN96 is being decoded? I am still getting choppy audio. Thanks for everyone's help thus far!

 

[R0am3r]

I looked at my FMP24.cfg file. I have no clue if these are correct, but they seem to work fine.

2.4 ; sampling rate (1.0, 2.0 or 2.4)
1024 ; spectrum window width
32 ; FFT size (8k, 16k, 32k)
10 ; spectrum update rate in Hz

 

[dave3825]

Just looked at mine When starting, I believe it does an fft optimization check and sets it.

2.4                                ; sampling rate (1.0, 2.0 or 2.4)

1024                               ; spectrum window width

64                                 ; FFT size (16k, 32k, 64k)

15                                 ; spectrum update rate in Hz

.

 

[NPC411]

Just looked at mine When starting, I believe it does an fft optimization check and sets it.

2.4                                ; sampling rate (1.0, 2.0 or 2.4)

1024                               ; spectrum window width

64                                 ; FFT size (16k, 32k, 64k)

15                                 ; spectrum update rate in Hz

.

Thanks, @R0am3r and @dave3825 ! I compared my settings, and since I'm using FMP24, my sampling rate seems hard-coded to 2.4. Out of the box, my FFT size was set to 32, and deleting fmp24.dat to force a new FFT optimization doesn't seem to change the FFT size. Just as a test, I tried adjusting it to 16 and 64, but I can't say I noticed any difference (though my untrained ear may be to blame). Does anyone know what FFT size controls?

I have attached a picture of how my gain looks now, though my audio is still choppy and hard to understand at times (I would attach an MP3 sample if the forum would allow it). Since this is NXDN96, would DSD+ settings for scaling factor, damping level, or viewport size make a difference?

 

[dave3825]

I just got in but looking at the spectrum, your using a -2.0 ppm and it looks like its also auto correcting, but it does not look centered to me. Is that pic the cc or voice dongle? The voice is the one that really needs the fine tuning. If the pic is voice, I would focus on that instance of fmp24 and hit a or A (which ever toggles auto tune) and manually play around with the ppm value to try to center it better.

 

[NPC411]

I just got in but looking at the spectrum, your using a -2.0 ppm and it looks like its also auto correcting, but it does not look centered to me. Is that pic the cc or voice dongle? The voice is the one that really needs the fine tuning. If the pic is voice, I would focus on that instance of fmp24 and hit a or A (which ever toggles auto tune) and manually play around with the ppm value to try to center it better.

That picture is the CC instance; the VC instance is a bit trickier because it is only active when there's traffic, which is sporadic and only lasts a few seconds at a time. Is there an easy way to check the VC instance, even if it's idle? Also, the FMP24 window is a little confusing for me to understand; is the green wave in the middle supposed to be the zoomed-in version of where I'm tuned?

 

[dave3825]

Is there an easy way to check the VC instance, even if it's idle?

I do not believe so as a signal is needed. You could fire up the voice instance of fmp24 and manually tune it to the cc and write down the ppm offset. It should get you in the ballpark as far as ppm. Then fire both up and adjust the voice instance to what looked good when it was locked on the cc.

Do you have any other systems that could aid in fine tuning the voice dongle?

 

[NPC411]

I do not believe so as a signal is needed. You could fire up the voice instance of fmp24 and manually tune it to the cc and write down the ppm offset. It should get you in the ballpark as far as ppm. Then fire both up and adjust the voice instance to what looked good when it was locked on the cc.

Do you have any other systems that could aid in fine tuning the voice dongle?

I originally used SDR# to tune each dongle to the local NOAA radio, which is where I got the original PPM offsets of -2.0. However, SDR# only does PPM offsets in whole numbers, so I can't say it's the best for super-fine tuning. I am also using official RTL-SDR V3 dongles, which supposedly require little-to-no offset adjustment. I'm open to using another software to check the fine tuning, but I didn't know what methods would be best and easiest in a Windows-based environment.

 

[R0am3r]

@NPC411 Have you tried starting off your scripts to run with 0.0 PPM? I wonder if that will help clear up the choppy signal? Also, what did you settle on for the gain? The FMP24 waveform looks right, but I can't discern what gain you are using.

 

[NPC411]

@NPC411 Have you tried starting off your scripts to run with 0.0 PPM? I wonder if that will help clear up the choppy signal? Also, what did you settle on for the gain? The FMP24 waveform looks right, but I can't discern what gain you are using.

I originally started with 0.0 PPM, which seemed to be fine at first. However, as the weeks went by, the voices became choppy, and it led me to try adjusting the PPM, which hasn't seemed to help. I ended up using 7.7 for the gain, though I'm open to adjusting it further.