You have a few different options.
We run an Xorg server using the dummy driver on our streaming server. This allows you to start an X environment without a monitor being connected. We then connect to that from a remote PC using x2go and run Stereo Tool GUI on the virtual X. This gives us the full flexibility to log in over an SSH tunnel and tweak Stereo Tool. We use this with jack and a liquidsoap script to generate our streams at various bitrates. It's very stable (two years running) apart from the odd xrun from jackd (although I have recently learnt a lot about running Stereo Tool on Linux having built a project for TechCon in the UK). You could always skip the x2go part and run it in a dummy Xorg server and then use the web server to connect in and edit settings.
Another way to do it without an X environment is to run jackd along with jack-stdout and pipe it into Stereo Tool CMD version (via sox as you have to get the format and 'endianness' right IIRC). You can run that under a screen instance to come back and monitor the process' stderr. We used this in project with netjack2 - we essentially took audio in from a physical linux computer in our studio, sent it over the network as a PCM stream using jack and then plugged it into jack, jack-stdout, ST, jack-stdin, darkice all hosted on a virtual server. It needed quite a lot of scripting and process monitoring but this worked also. We have to set the settings up on a local copy of ST and then load the .sts file into the command line.
That said, we are moving our production streaming client over to Windows 7 now since you can achieve far lower latencies without frame drops etc using ASIO. I think there needs to be a way to add liquidsoap scripting to ST's output module since you can't stream to an icecast server using VLC (seemingly).
|