Changes between Version 2 and Version 3 of Branch_clientDeviceConfig

Author:

jmsizun (IP: 127.0.0.1)

Timestamp:

25.05.2010 23:53:20 (3 years ago)

Comment:

added info on extensions of process-type transcoding

Branch_clientDeviceConfig

@@ -146 +146 @@
-
-Notes: file extension value are to be provided with a dot. UPnP classes are to be provided with the fully qualified name.
+
+== Updates to process-style transcoders ==
+
+Support for transcoders with type "process"  has been extended as follows.
+data output by the 'process' is bufferized in a FIFO, up to a given size. Above this size, the process is paused. To be frank, the concepts here are heavily 'borrowed' from [[http://mediatomb.cc mediatomb]].
+Here are to the available parameters, besides existing {{{type}}}, {{{name}}} and {{{target}}}:
+
+{{{ pipeline }}} : the command to launch the transcoder process. This command takes the form of a python formatting string, with the following variables:
+ - either a single {{{%s}}} or {{{%uri}}} variable, corresponding to the source URI or file path. The command has to output transcoded data to its standard output.
+ - either both {{{%uri}}} and {{{%output}}} variables, corresponding to the source URI or file path, and file path to a pipe file that coherence will stream to the client
+
+{{{ source }}} : the mimetype (as attribute {{{mimetype}}}) for which the transcoder will be applied. For other data types, no transcoded resources will be proposed.
+
+{{{priority}}} : defines how the transcoded resource will be published, besides the original one. Available values are:
+ - -1: the transcoded resource will be published before the original one(s)
+ - 0: the transcoded resource will be published in place of the original one(s)
+ - 1: the transcoded resource will be published after the original one(s)
+ - 2: the transcoded resource will be published at the end of the list
+
+{{{buffer}}} : (optional)tuning parameters for the buffering taking place at the server side, when streaming the transcoded resources. Attributes are:
+ - {{{size}}} : total size of the buffer
+ - {{{chunk-size}}} : size of data chunks read from the FIFO and served to the client
+Values may be tuned according to your setting, to smoothen playback or optimize server resource consumption.
+
+Below a sample transcoder configuration (see svn for the source code for the given {{{bash_memcoder_srt2mpeg}}} command: it takes a srt file, accompanying a video file, and transcode it to mpeg embedding the subtitles).
+
+{{{
+<transcoder>
+  <pipeline>misc/transcoders/bash_memcoder_srt2mpeg %uri %output</pipeline>
+  <type>process</type>
+  <name>srt2video</name>
+  <target>video/mpeg</target>
+  <!--<buffer size="5242880" chunk-size="8192"/>-->
+  <source mimetype="application/x-srt"/>
+  <priority>0</priority>
+</transcoder>
+}}}