M17 Digital Voice , mvoice, is a fully functional, graphical repeater. It uses David Rowes Codec 2 and operates as a complete M17 repeater, only there is no RF component. It can Link to M17 reflectors and it can also do routing! It works best with a USB-based headset with microphone. mvoice uses the default pulseaudio/ALSA input and output device, so for most versions of linux, all you need to do is plug your headset in and you should be ready to go.

Raspberry Pi OS is not ready for mvoice out of the box. While Raspbian has ALSA, it also needs pulseaudio:

sudo apt install -y pulseaudio pavucontrol paprefs

It's probably a good idea to reboot after installing pulseaudio. After it's installed, take a look at the output of aplay -L. Near the beginning is should say:

default
    Playback/recording through the PulseAudio sound server

You might also need to go to the ALSA audio configuration. For Debian Buster, this is found in main menu under Preferences-->Audio Device Settings. Select your headset in the drop-down list of devices, and then configure it and set it to be the default device. Set your speaker and microphone gain somewhere near the top. Once you build and configure mvoice, you can use the Echo feature to help you set these gains adjust the speaker volume of you headset for a comfortable level and adjust the mic gain for the loudest playback without clipping. For my setup, the playback (speaker) was near 100% and the mic gain was at about 50%.

There are several library requirements before you start:

sudo apt install -y git build-essential libasound2-dev libcurl4-gnutls-dev gettext

mvoice now uses FLTK, the Fast Light Tool Kit. You may be able to install it with the package manager:

sudo apt install -y libfltk1.3-dev

You can also build it yourself from the FLTK git repository:

git clone https://github.com/fltk/fltk.git
cd fltk
git checkout branch-1.3
make
sudo make install

Using the OpenDHT library, mvoice now joins a special digital voice, ham-radio DHT network to discover reflector and repeater destinations directly. For these target systems using the DHT, connection information is published and updated directly by the target and is available to mvoice in near-realtime. All the mvoice user needs to know is the callsign of the target.

OpenDHT is available here. Building and installing instructions are in the OpenDHT Wiki. Pascal support and proxy-server support (RESTinio) is not required for mvoice and so can be considered optional. With this in mind, this should work on Debian/Ubuntu-based systems:

# Install OpenDHT dependencies
sudo apt install libncurses5-dev libreadline-dev nettle-dev libgnutls28-dev libargon2-0-dev libmsgpack-dev  libssl-dev libfmt-dev libjsoncpp-dev libhttp-parser-dev libasio-dev cmake pkg-config libcppunit-dev

# clone the repo
git clone https://github.com/savoirfairelinux/opendht.git

# build and install
cd opendht
mkdir build && cd build
cmake -DOPENDHT_PYTHON=OFF -DCMAKE_INSTALL_PREFIX=/usr ..
make
sudo make install

Please note that there is no easy way to uninstall OpenDHT once it's been installed. However, it is based on static header files and libraries and doesn't use any resources except for a small amount of disk space, unless mvoice is running. You can delete your downloaded OpenDHT git repo once you've installed the OpenDHT library.

git clone https://github.com/n7tae/mvoice.git
cd mvoice

There are several compile-time options for building mvoice, including where run-time files are stored, where the mvoice executable is installed, support for 44,100 Hz (CD-quality) audio, debugging support and support for the new Digital Voice Information Network. These are all controlled by your very own mvoice.mk file. Start by creating the file:

cp example.mk mvoice.mk

Use your favorite editor to change values in your mvoice.mk file.

You're now read to build mvoice:

make

If it builds okay, then you can install it:

make install

To clean up object files and other intermediate files:

make clean

Don't delete the build directory because that's the only way to remove mvoice from your system:

make uninstall

If you want a desktop icon to launch mvoice then, from your build directory:

cp mvoice.desktop ~/Desktop

If you don't want to answer the question "What do you want to do with it?" every time you double-click the new desktop icon, then in the Raspberry Pi Menu, go to "Accessories->File Manager->Edit->Preference->General and turn on the "Don't ask options..." check box. The mvoice.desktop file will launch a terminal window to run mvoice. If you really don't want to see this terminal window, then change the "Terminal=true" line to "Terminal=false" in your copy of mvoice.desktop in your Desktop folder.

The Raspberry Pi OS has a bug in xdg-open and that's what mvoice's "Open Dashboard" button uses to view a reflector's dashboard. If Chromium isn't running when you click this button, it will freeze the mvoice gui. To unfreeze the gui, simple kill Chromium. To get around this bug, launch Chromium before you click this "Open Dashboard" button and a new tab with the dashboard will open without a problem.

Plug in your headset and start mvoice: Open a shell and type mvoice if ~/bin is in your search path, otherwise, move to your home directory and type bin/mvoice, or where ever you installed it. Most log messages will be displayed within the log window of mvoice, but a few messages, especially if there are problems, may also be printed in the shell you launch mvoice from.

mvoice can operate in linking or routing mode. If you want to be able to receive incoming callsign routes, you'll need to port-forward UDP 17000 on your network firewall. If you can't, you can still connect to M17 reflectors. You can originate a direct, callsign route, but if there is a long pause in the QSO, you're firewall will shutdown the port until you key up again.

The first thing mvoice will do is to download the list of registered reflectors from the M17project.org website and save it in your configuration directory. It will do this every time you start mvoice.

Once it launches, click the Settings button and make sure to set your callsign and the codec setting on the M17 page. You can usually leave the audio settings on "default". Also enable IPv6 if your internet provider supports it. Click the Okay button and your settings will be saved in your configuration directory.

On the main window, set your destination callsign and IP address. You can select a reflector from the dropdown list and that will fill the Destination callsign and IP address, or you can type the callsign into the Callsign Entry box. If the IP address can be found from your local data, the IP will be automatically set, otherwise, you will need to know the IP address of your destination. Once you have valid values for both, you can save these for future use, if it's not already in your database. Please remember that when you are setting a new destination, the callsign and IP entry will "approve" their input strictly based on syntax and not whether the callsign or an IP address is actually a place where an M17 destination exists.

Please note that you don't have to save a destination to use it, but it will not be available in the drop-down selection until you do save it. If you haven't saved a destination and you select another destination from the dropdown list, your unsaved destination will no longer be available. If you are going to use an M17 reflector, a valid callsign is exactly seven characters long, M17-xxx where xxx is made up of letters or numbers. Or, now you can also link to the new URF reflectors. A URF reflector callsign is exactly six characters long, beginning with URF.

Once you set the reflector callsign, you can select the reflector module with the radio buttons from A to Z. If you are callsign routing to an individual, you can provide a module in the ninth position, but it isn't necessary. If mvoice has found a URL for the target, you can open the reflector's dashboard with the Open Dashboard button.

For Linking, you can select a reflector. The link button will only be activated after you have entered a valid target in the destination callsign and IP address. This validation does not check to see if reflect exists at that IP and the module you are requesting is actually configured and operational, or even if it actually exists.

There are three "transmitting" buttons on mvoice that are explained below, but first, an explanation:

Both the Echo Test and the PTT buttons are toggle buttons. You press and release these buttons to start their function and press and release again to stop their function. Both of these buttons also have built-in timers and will show the approximate "key on" time.

1. Echo Test is a toggle switch. Click it to start recording an echo test and your voice will be encoded using the Codec 2 mode you selected in the Settings dialog. Note that the button will turn yellow when you are recording your test. Click the button again and mvoice will decode the saved AMBE data and play it back to you. This is a great way to check the operation of your setup as well as the volume levels and make sure your headset is working properly. Whenever you complete an Echo Test, you'll get a statistical summary of your recorded audio, the volume (in decibels relative to an arbitrary reference that is properly adjusted audio) and the percentage of when your audio was within 6 dB of clipping (50% of the maximum amplitude). For the echo test, you'll generally want 0% clip and a volume of around 0 dB, although most anything above -10 dB is still copyable. So first adjust your system microphone gain to achieve the desired dB, and then adjust you system volume level for a comfortable listening level. Note that mvoice has no volume or gain controls for either TX or RX. You'll use your Linux gui audio settings for this.

2. PTT the large PTT button is also toggle switch. Click it and it will turn yellow also and mvoice will encode your audio and send it to your destination. Click the button again to return mvoice to receiving. When you do, you'll see your volume statistics. mvoice will also report volume statistics on incoming audio streams. There is a locking mechanism in PTT: When you start a transmission by toggling the PTT to the "key on" state, mvoice will attempt to lock the gateway. If the lock is achieved, all is good. If the gateway is busy, the PTT lock will fail and the PTT button will be returned to the "key off" state. While response of this lock is quick, it does take a finite time to execute! So here is a best practice: Key up, and wait one or two seconds before you start talking. By waiting, you're insuring that the PTT lock has been successfully achieved.

3. Quick Key is a single press button that will send a 200 millisecond transmission. This is useful when trying to get the attention of a Net Control Operator when your are participating in a Net. It could also be useful if you are trying to do a direct callsign route when you are behind a firewall you can't configure.

Thanks to IU5HKX, mvoice is now available in Italian.

If you would like to add another language to mvoice GUI to use a different language, I would welcome a pull request. To get started, see https://www.gnu.org/software/gettext/manual/html_node/index.html.

de n7tae (at) tearly (dot) net