TalKT is early alpha software, please DO NOT expect it to work. Instead report every problem to Hans-Peter. Thanks.
TalKT is a very simple application which allows voice QSO over packet. It uses a very simple but flexible protocol called VoAX to transfer data. The codec is not part of TalKT itself, so it can be changed if you want. VoAX can handle different formats, in fact it could be used to transfer other data then compressed speech, e.g. images. The current version is designed to use MELP which is used by FlexTalk, too. Since I am unsure about the licensing of MELP, please download it at your own risk. I one of the next releases I will include support for LPC10, which I know to be publicly available. LPC10s quality is significantly worse than MELP, though. Both need about 2k4 throughput for realtime, so you should have at least 9k6. TalKT doesn't do realtime at the moment, but this is planned (VoAX can handle it). TalKT depends on Sox for audio recording/output, it should be part of your distribution.
I have started to implement a conference server which will allow voice-nets over packet radio. It will be - based on VoAX - rather generic, so you can even exchange other data than voice. Or use a different codec. The server will only send data to clients which they are able to handle.
talkt-0.0.4.tar.gz (108 kB)
Sox must be installed with binaries
Get MELP from here
(the fixed point version I think), untar it in a new directory (no path in the
tar!) and compile it by typing
make. Put the binary to
/usr/bin. Then compile TalKT. This is similar to the LinKT
installation, you must have the ax25-util headers and library installed. If it
is in a uncommon directory, run configure with
I will make the location of sox and melp configurable in future versions, you may change it in the sourcecode by now if you need to.