Decoded Audio - Flex Integration vs Desktop Audio

221 views
Skip to first unread message

K9YK Radio

unread,
Dec 29, 2025, 3:48:19 PM12/29/25
to digitalvoice
Since I've had a few days to listen to FreeDV decoded (received) audio I've noticed that the decoded audio coming from the Flex integration (both the RPi AppImage, and my own built Ubuntu instance) produced quite a bit a robotic noise when receiving a station and the person pauses between words, or inhales while speaking.   The desktop version of FreeDV doesn't seem to have this issue.   I'm testing running both the Flex integration and desktop version on my Flex at the same time using two different slices. One slice is in FDVU mode and the other slice is in USB feeding the desktop version via DAX.   What causes the difference in behavior/decode quality between the two versions?

I've also noticed that the decoded audio level coming from the Flex integration version is very low compared to analog SSB audio coming from the radio. This makes me keep reaching for the volume control when switching back and forth between FreeDV and SSB.  Can something be done to increase the decoded audio level for the Flex integrated version?

Mark (K9YK)

Stefan N.

unread,
Dec 30, 2025, 4:37:09 AM12/30/25
to digitalvoice
My experience with the waveform (RSPi 4+) so far:

Switching AGC from „mid“ to „slow“ produces a better RX audio.
No complaints from other OM‘s about any low audio level from my side.
Going to check that out the coming days via diff. WebSdr‘s.
Thought that the mic adjustment (level controlled by the meter) has the same function than in SSB?

Would be helpful to put basic infos in the message window (name and QTH) and to have a option to edit the Locator. 
I know, that is already in the pipeline👍
I also don’t see any SNR‘s on the screen🤔

🥳 Happy new (FDV) year🎉
Stefan - DF6VI 

Mooneer Salem

unread,
Dec 30, 2025, 12:49:29 PM12/30/25
to digita...@googlegroups.com
Hi Mark,

Does the volume slider on the right hand side in SmartSDR have any effect on the loudness for you? (The 100 shown below, for example.)

image.png

Thanks,

-Mooneer K6AQ

--
You received this message because you are subscribed to the Google Groups "digitalvoice" group.
To unsubscribe from this group and stop receiving emails from it, send an email to digitalvoice...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/digitalvoice/a6a5ea44-b556-44f4-8866-44182b6e0f86n%40googlegroups.com.

Mooneer Salem

unread,
Dec 30, 2025, 12:50:10 PM12/30/25
to digita...@googlegroups.com
Hi Stefan,

Interesting re: AGC. Do you notice the same thing if you use the FreeDV application instead of the waveform?

Thanks,

-Mooneer K6AQ

--
You received this message because you are subscribed to the Google Groups "digitalvoice" group.
To unsubscribe from this group and stop receiving emails from it, send an email to digitalvoice...@googlegroups.com.

Stefan N.

unread,
Dec 31, 2025, 1:00:03 PM12/31/25
to digita...@googlegroups.com
Hi Mooneer,

no, it's only happening with the waveform. 
BTW: The Audioslider has it's usual funktion, even the level is not very high at maximum gain.
Had a QSO these days where in inhaling from the OP was very present and sounds like Darth Vader sometimes ;-)
I have the feeling it's also only with the waveform, but nor sure. Heard him one day later via the mac software and it was much weaker.
Not sure, if he changed something on his equipment? That's possible too.

73, Stefan

You received this message because you are subscribed to a topic in the Google Groups "digitalvoice" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/digitalvoice/pxLRiDofAwQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to digitalvoice...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/digitalvoice/CAAzDfxNDmHhJWn%3DLL%2BeuP%3D04f4Kn0c3YfB%3DSxg1KE393HKReSg%40mail.gmail.com.

Mooneer Salem

unread,
Dec 31, 2025, 8:01:29 PM12/31/25
to digita...@googlegroups.com
Hi all,

Give the build in this PR a shot and let me know how it goes: https://github.com/drowe67/freedv-gui/pull/1156. (It's still building as of this writing so it might be a bit before you can click on the download link for the Flex AppImage. This can be found at https://github.com/drowe67/freedv-gui/actions/runs/20629471918?pr=1156 once the build completes.)

Thanks,

-Mooneer K6AQ

K9YK Radio

unread,
Jan 1, 2026, 12:58:22 PMJan 1
to digitalvoice
Hi Mooneer!   I just cloned and built from branch 1156 - ms-flex-audio-limiter and the decoded audio sounds better.  I'm now not hearing the robotic sound between words or when people are inhaling like I had been when using the build from the trunk.  This is much better!

I also tried using -m and -l command line options to set the message and locator and those are also working fine.

Thankk you so much!
Mark (K9YK) 

K9YK Radio

unread,
Jan 1, 2026, 1:02:43 PMJan 1
to digitalvoice
Yes, the audio slider does increase/decrease the volume of the freedv decoded audio.  The problem is when switching from FreeDV back to SSB, I have to turn down the volume as SSB audio is louder.  When switching from SSB to FreeDV, I have to turn the volume up, and the FreeDV audio is quieter.  Having to keep turning the volume up and down when switching modes is somewhat of a pain.   

Could this be related to the RX EQ in the Flex. Does the FreeDV decoded audio pass through the Flex RX EQ?  If so, this could be part of the difference.  I'll have to do some experimenting with the effects of the EQ. 

Mark (K9YK)

Stefan N.

unread,
Jan 1, 2026, 1:27:56 PMJan 1
to digita...@googlegroups.com
Mooneer,

thanks for the links, but I'm on the line.
Nothing to download so far. Just a „copy“ option, which I can’t handle.

73, Stefan

You received this message because you are subscribed to a topic in the Google Groups "digitalvoice" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/digitalvoice/pxLRiDofAwQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to digitalvoice...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/digitalvoice/f72f0826-9f8f-452c-862b-994889e851d5n%40googlegroups.com.

Alan Blind

unread,
Jan 1, 2026, 3:02:29 PMJan 1
to digitalvoice
I have identified the issue and a fix.   I leave it to Mooneer to decide if it is a Flex firmware or FreeDV issue, or no issue at all.

First, everything pointed to two instances of Flex AppImage was running, the Flex "residual" Meter that would not clear and the two entries on FreeDV recorder (although the second entry had the comment, that was only included in the update RPI version.

Second, repeated attempts to remove the "residual" SDR meter from the Flex radio list, failed, including factory resets and firmware changes.  As reported, any new boot created a new, second, SNR meter that did not work.

Then, Dave, WO2X said that any open socket may be causing the Flex Radio to not clear the residual meter created from an API.  So......instead of just shutting down the Flex API instances, I shutdown the power to both RPIs, containing, but NOT running Flex AppImage files.

Success, after a factory reset, the "residual" SNR Flex meter was gone!  This shows the RPI itself was holding the socket open, as seen by the Flex Radio, even though the Flex AppImage was shutdown on each of the two RPIs.

Also, only one FreeDV entry is now seen, as expected.

Alan. WA9WUD

Mooneer Salem

unread,
Jan 2, 2026, 4:33:03 AMJan 2
to digita...@googlegroups.com
Hi Mark,

Glad that the audio sounds better now. :)

Anyway, I'm not actually sure re: the RX equalizer. Definitely worth experimenting and letting us know!

Thanks,

-Mooneer K6AQ

Mooneer Salem

unread,
Jan 2, 2026, 4:36:10 AMJan 2
to digita...@googlegroups.com
Hi Alan,

Interesting that the socket would stay open even after the process itself was killed. So far I haven't heard any other reports like this, so maybe keep an eye out and see if it happens again?

Thanks,

-Mooneer K6AQ

K9YK Radio

unread,
Jan 2, 2026, 9:41:26 AMJan 2
to digitalvoice
I checked yesterday afternoon, and can confirm that the FreeDV decoded audio does NOT pass through the Flex receive EQ.   When receiving FreeDv, turning the receive EQ on/off makes no change in the received audio.  If I switch to analog USB, then a definite difference can be heard.  This may go a long way to explaining why analog reception on my Flex sounds louder than FreeDV.  

Would it be possible to add a config/command line option to the Flex integrated version to boost or cut the received FreeDV audio level like there is in the GUI version? 

Mark (K9YK)

Mooneer Salem

unread,
Jan 2, 2026, 12:46:14 PMJan 2
to digita...@googlegroups.com
Hi Mark,

See https://github.com/drowe67/freedv-gui/pull/1159. I haven't had a chance to test this change yet but let me know how it goes.

Thanks,

-Mooneer K6AQ

Stefan N.

unread,
Jan 2, 2026, 1:34:59 PMJan 2
to digita...@googlegroups.com

Hi Mooneer,

that's what i tried so far:
Pasted that in a 2.nd Terminal window:

stefan@raspberrypi:~ $ ./FreeDV-FlexRadio-2.2.0-dev-97fc-aarch64.AppImage -m "this is a test" -l JN58tc
bash: ./FreeDV-FlexRadio-2.2.0-dev-97fc-aarch64.AppImage: No such file or directory

Tried the same changing to the directory where the AppImage is stored:
stefan@raspberrypi:~ $ cd /home/stefan/Desktop
stefan@raspberrypi:~/Desktop $ ./FreeDV-FlexRadio-2.2.0-dev-97fc-aarch64.AppImage -m "this is a test" -l JN58tc
bash: ./FreeDV-FlexRadio-2.2.0-dev-97fc-aarch64.AppImage: No such file or directory

Then pasted the same link it in the active terminal window

Same respond and nothing happend

I guess it's just a missunderstanding in how to send the command to the running app, right?

Thanks. Stefan

P.S. If that works, are there any additional meters (SNR) to find?

You received this message because you are subscribed to a topic in the Google Groups "digitalvoice" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/digitalvoice/pxLRiDofAwQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to digitalvoice...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/digitalvoice/CAAzDfxPXfo44PXUWF1cCUAY4oCGD%3Dk%3Dp19zbPE%2B1iYoB%2Bu3vSA%40mail.gmail.com.

Mooneer Salem

unread,
Jan 2, 2026, 6:24:51 PMJan 2
to digita...@googlegroups.com
Hi Stefan,

Can you confirm that you downloaded the file from GitHub? If so, you'll need to unzip it first before you'll have access to the actual AppImage file.

Thanks,

-Mooneer K6AQ

K9YK Radio

unread,
Jan 2, 2026, 9:40:58 PMJan 2
to digitalvoice
Hi Mooner,
From the changes I see in the branch, it looks like -r as a command line option should work to adjust the received audio level, but the application just exited almost as soon as it started.  Then I tried --rx-volume as the option and that seemed to work will need to play with it a little more tomorrow to check out the audio levels a little more closely.

Mark (K9YK)

stefan netzer

unread,
Jan 3, 2026, 4:22:55 AMJan 3
to digita...@googlegroups.com
Good morning Mooneer,

i downloaded the file via the link you sent me (it wasn‘t zipped). The AppImage was running already and i was testing, (had two QSO‘s) when i startet to try the „plugins“.
I saw the zipped file on GitHub. Should i download this for another try? The version seems to be the same, so…?

BTW: What’s the recommended way to send the additional stuff to the radio? Should a second terminal window be opened, or should it just be pasted in the one that is already open with the running waveform?

The Micgain via the M has the same function as usual, but it’s not so easy to find the right gain (the level adjustment is different to SSB) as in the software version where we have a monitor screen. I checked my audio via WebSdr and the gain was too hight, even it looks good in the meter.

Thanks,

Stefan 

Am 03.01.2026 um 03:41 schrieb K9YK Radio <k9yk...@gmail.com>:

Hi Mooner,
You received this message because you are subscribed to a topic in the Google Groups "digitalvoice" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/digitalvoice/pxLRiDofAwQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to digitalvoice...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/digitalvoice/0ea91486-90eb-478d-8ec8-d625e32f6a54n%40googlegroups.com.

Mooneer Salem

unread,
Jan 3, 2026, 12:00:28 PMJan 3
to digita...@googlegroups.com
Hi all,

I updated the PR to fix the issue causing -r not to work.

Stefan: I'd try the version that is currently being built now (https://github.com/drowe67/freedv-gui/actions/runs/20680195895?pr=1159), which will need to be unzipped once downloaded. Any other versions of the AppImage should also be killed before running this one (which you should be able to do by pressing Ctrl-C in the same window as the one already running).

Thanks,

-Mooneer K6AQ

Stefan N.

unread,
Jan 4, 2026, 10:16:11 AMJan 4
to digita...@googlegroups.com
Putting text in the message box via the -m command.
Just found out that:  "XYZ_testing_flex_waveform" shows the hole text in "message" . Without using underlines, only  "XYZ" is shown in the reporter. 

73, Stefan

Mooneer Salem

unread,
Jan 5, 2026, 12:02:12 AMJan 5
to digita...@googlegroups.com
Hi Stefan,

Does only "XYZ" show up when you surround the text with quotes, i.e.

-m "XYZ testing flex waveform"

What about if you use single instead of double quotes?

Thanks,

-Mooneer K6AQ

stefan netzer

unread,
Jan 5, 2026, 3:46:27 AMJan 5
to digita...@googlegroups.com
Hi Mooneer,

Yes, only the first entry appears. I'm going to try using single quotes today.

By the way, I found by chance that the lower the AGC-T level (it's best at 0), the better the audio comes in.

Next, I'll try reducing the audio output gain (is that the '-r'?). That seems to be a bit over-modulated, even with the reduced gain level.

In some cases, the low-cut level can no longer be adjusted on the console, but I'll have to look into that.

Thanks and 73,


Stefan 

Am 05.01.2026 um 06:02 schrieb Mooneer Salem <moo...@gmail.com>:



Michael, DJ2VA

unread,
Jan 5, 2026, 6:49:24 AMJan 5
to digitalvoice
Hi Mooneer, hi Stefan,
it makes no difference. I did the same test yesterday. It makes no difference if I am using single or double quotes. using the command line parameter  -m "DJ2VA test 1234", only DJ2VA shows up. If I use _ instead of spaces all the text in quotes are shown.

73 de Michael, DJ2VA

Mooneer Salem

unread,
Jan 5, 2026, 1:46:47 PMJan 5
to digita...@googlegroups.com
Hi Michael,

Were you doing this from the Pi itself using whatever terminal program comes with its desktop environment? Or did you SSH into the Pi from another machine and run the AppImage that way? If the former, is there any difference in behavior if you SSH instead?

Thanks,

-Mooneer K6AQ

dj2v...@gmail.com

unread,
Jan 5, 2026, 4:44:38 PMJan 5
to digita...@googlegroups.com

Hi Mooner,

yesterday I accessed my Raspi via VNC, opened a terminal and started the FlexAppimage with the command:

 

./FreeDV-FlexRadio-2.2.0-dev-4c69-aarch64.AppImage -m 'Michael_testing_Flex_AppImage'

./FreeDV-FlexRadio-2.2.0-dev-4c69-aarch64.AppImage -m 'Michael testing Flex AppImage'

 

Right now I logged into the Raspi with SSH and used the same commands and the behaviour is exactly the same.

 

Using spaces within the text always shows only the word “Michael”.

Using _ shows the complete text.

 

There’s no difference using single or double quotes.

 

73 de Michael, DJ2VA

Hi Mark,

 

 

Thanks,

 

-Mooneer K6AQ

Mooneer Salem

unread,
Jan 5, 2026, 7:51:48 PMJan 5
to digita...@googlegroups.com
Hi Michael,

Can you enter the following and let me know the output? I wonder if there's a difference depending on the shell being used:

echo $0

FWIW, this shows either "-bash" or "/bin/bash" for my Pi CM4 with Trixie over SSH depending on whether I'm inside a screen session or not.

Also, with commit ID 7986e36 from the master branch and the following command line:

(rade-venv) mooneer@flexradio-test:~/freedv-gui/build_linux $ PYTHONPATH=$(pwd)/rade_integ_src src/integrations/flex/freedv-flex -m "San Diego, CA - FlexRadio waveform testing"

the entire message appears to display for me:

image.png

Thanks,

-Mooneer K6AQ

Mooneer Salem

unread,
Jan 5, 2026, 10:04:24 PMJan 5
to digita...@googlegroups.com
Hi all,

Actually, I just had a realization that the AppImage may be behaving differently vs. running the waveform directly. Can you try the version of the AppImage at https://github.com/drowe67/freedv-gui/actions/runs/20736490756?pr=1168 (once it finishes building) and let me know how it goes?

Thanks,

-Mooneer K6AQ

Michael, DJ2VA

unread,
Jan 6, 2026, 3:07:00 AMJan 6
to digitalvoice
Hi Mooner, here are the results:
SSH via Puttty:
echo $0
=> - bash

VNC Viewer on Raspberry, opening a terminal
echo $0
=> bash    (with no "-" in the beginning)



I also tested the new version (https://github.com/drowe67/freedv-gui/actions/runs/20736490756?pr=1168) mentioned in your last mail. It works correct in both terminals. Using spaces shows the complete string, not just the first word - perfect!

73 de Michael

Stefan N.

unread,
Jan 6, 2026, 5:29:46 AM (14 days ago) Jan 6
to digita...@googlegroups.com
Hi Mooneer,

can confirm,, that with 4e5b the issues are solved.
Great job.

Thanks, Stefan

Reply all
Reply to author
Forward
0 new messages