ts2shout is a small software that acts as a filter or cgi programm to convert a single program mpeg transport stream directly to a webradio compatible shoutcast stream. The generated shoutcast stream includes broadcast and title information. I actually use it for playing DVB-C and DVB-S on Squeezebox players. Please see my page http://www.siski.de/~carsten/radio-streaming-squeezebox.html for more information.
It is based on the dvbshout application by Nicholas J. Humfrey (https://github.com/njh/dvbshout) - but the use case here is completly different. ts2shout uses autoconfiguration by reading PAT and PMT tables from a single program mpeg transport stream and outputs a native mpeg (or shoutcast) stream directly.
There are two modes, one is very like a traditional unix filter: a mpeg transport stream is read from stdin and it is filtered to stdout. Error and info messages are directed to sterr.
The other mode is CGI (common gateway interface) mode. This avoids hasseling around with shoutcast like parameters (like icy-sr, icy-br etc.) which are read directly from the mpeg transport stream and output in the corresponding and required header lines. To support this mode libcurl is used for accessing tvheadend. This means libcurl (and the corresponding libcurl-dev package for compiling) is required for ts2shout to work. In CGI mode it is expected that you have a tvheadend or a similar streaming server available that provides the MPEG transport stream by a network protocol libcurl understands. Please see a configuration example for apache2 below.
Additionally I started to implement DSM-CC to fetch stream images. This requires the availability of libz. Please install libz and the corresponding libz-dev package. Currently dsmcc is disabled in the code, because it is a pre-alpha version just writing the stream data files into a cache dir.
ts2shout outputs a shoutcast stream with the "best" found mpeg audio stream in PAT/PMT and the "current programm" from the mpeg EIT ("EPG") translated to "StreamTitle" all 8192 bytes inside the mpeg stream. This is a standard for most shoutcast radio-stations and described in https://stackoverflow.com/questions/44050266/get-info-from-streaming-radio. ts2shout should work with most DVB-C/DVB-S radio stations and is tested by myself with Unitymedia DVB-C, Hotbird 13E and Astra 19.2E DVB-S satellite reception.
Many radio stations on DVB-S (and also DVB-C) supports RDS within the MPEG transport stream. The RDS data (Radio Data System, that thing from the Analog Radio on FM) is just inserted inside the padding bytes MPEG1/2 audio stream. If the stream uses AAC-LATM for audio RDS is also searched in a defined private stream in a separate MPEG pid. There is a method to include RDS also in AAC-LATM audio directly. As at is necessary to fully decode the AAC data for this I made a modified version of ffmpeg that is capable of extracting the RDS data out of AAC and provide it to ts2shout. Please see my patched version of FFmpeg and the Makefile of this project.
If you supply the environment variable RDS or the command line parameter rds it is preferred over MPEG EIT/EPG data. This is useful to get title and artist information in the shoutcast stream.
Just compile the application with make on your linux box and install it manually with "make install", which defaults to /usr/local/bin. It's possible to use it without ffmpeg, for a lot of use cases libcurl and libz is sufficient. ffmpeg is required to fetch inline RDS data from AAC audio.
To achive decoding of inline AAC RDS data, please download my version of ffmpeg and place the source directory directly next to the source directory of ts2shout. Please install all the supportive stuff necessary for compiling of ffmpeg and compile ffmpeg. You should take care to not install this version of ffmpeg to avoid interfering with your system installation. Just compile it to get the ".a" libraries. After successfully compiling ffmpeg you can enable "USE_FFMPEG" in the Makefile of ts2shout. Now you can just compile ts2shout. The resulting executable of ts2shout will be quite big, because it contains most of ffmpeg's libraries to decode AAC RDS.
Your source directories should be placed like this
mamba:~/src>ls -ald ts2shout FFmpeg drwxr-xr-x FFmpeg/ drwxr-xr-x ts2shout/
Some german cultural programmes also have AC3 streams available. If you supply the command line parameter "ac3" or (in CGI mode) the value "1" in the environment variable "AC3" AC-3 stream will be preferred. Some programmes have AC-3 only (NDR-Kultur), in this case you'll always get AC-3 audio. Native AC-3 currently doesn't work with the squeezebox but you can fetch the stream and it can be decoded by ffmpeg or played with mplayer.
ts2shout can be used in conjunction with tvheadend and apache2 as follows:
Activate modules (if not already done) mod_cgi and mod_action (a2enmod cgi ; a2enmod action) on your apache2. Take care that ts2shout is placed in the cgi-bin of the webserver, on debian it is searched in /usr/lib/cgi-bin. Be careful with permissions and reachability - ts2shout is not intended for worldwide accessibility.
On a Debian system change /etc/apache2/sites-available/000-default.conf as follows. Use a virtual server from your apache2 installation as you wish.
It assumes that your tvheadend is reachable via localhost. Change and adjust the radio URLs, the tvheadend URLs and the allow/deny lists as you need them.
If shoutcast Streamtitles should be used inside the stream is autodetected by the Icy-MetaData header given in the http request.
<Location /radio>
Order Deny,Allow
Deny from all
Allow from 172.16.0.0/24
Allow from ::1
Allow from 127.0.0.1
<If "%{HTTP:Icy-MetaData} in {'1'}">
SetEnv "MetaData" "1"
</If>
# If you prefer RDS data
SetEnv RDS 1
SetEnv TVHEADEND "http://localhost:9981/stream/channelname"
# The radio stations are called e.g.
# /radio/SWR1%20BW
# if you want to fetch a radio channel named "SWR1 BW" in tvheadend
# Only A-Z/a-z/0-9, space and "-" are possible due to the regex.
# Change names containing other characters in tvheadend frontend
# You can also use http://localhost:9981/stream/channelnumber
# and use channel numbers
SetEnvIf REQUEST_URI "([-A-Za-z0-9_ ]*)$" PROGRAMMNO=$1
Options +ExecCGI
Action ts2shout /cgi-bin/ts2shout virtual
SetHandler ts2shout
</Location>