+ * Add a graph described by a string to a graph.
+ *
+ * @param graph the filter graph where to link the parsed graph context
+ * @param filters string to be parsed
+ * @param inputs linked list to the inputs of the graph
+ * @param outputs linked list to the outputs of the graph
+ * @return zero on success, a negative AVERROR code on error
+ */
+int avfilter_graph_parse(AVFilterGraph *graph, const char *filters,
+ AVFilterInOut *inputs, AVFilterInOut *outputs,
+ void *log_ctx);
+
+/**
+ * Add a graph described by a string to a graph.
+ *
+ * @param[in] graph the filter graph where to link the parsed graph context
+ * @param[in] filters string to be parsed
+ * @param[out] inputs a linked list of all free (unlinked) inputs of the
+ * parsed graph will be returned here. It is to be freed
+ * by the caller using avfilter_inout_free().
+ * @param[out] outputs a linked list of all free (unlinked) outputs of the
+ * parsed graph will be returned here. It is to be freed by the
+ * caller using avfilter_inout_free().
+ * @return zero on success, a negative AVERROR code on error
+ *
+ * @note the difference between avfilter_graph_parse2() and
+ * avfilter_graph_parse() is that in avfilter_graph_parse(), the caller provides
+ * the lists of inputs and outputs, which therefore must be known before calling
+ * the function. On the other hand, avfilter_graph_parse2() \em returns the
+ * inputs and outputs that are left unlinked after parsing the graph and the
+ * caller then deals with them. Another difference is that in
+ * avfilter_graph_parse(), the inputs parameter describes inputs of the
+ * <em>already existing</em> part of the graph; i.e. from the point of view of
+ * the newly created part, they are outputs. Similarly the outputs parameter
+ * describes outputs of the already existing filters, which are provided as
+ * inputs to the parsed filters.
+ * avfilter_graph_parse2() takes the opposite approach -- it makes no reference
+ * whatsoever to already existing parts of the graph and the inputs parameter
+ * will on return contain inputs of the newly parsed part of the graph.
+ * Analogously the outputs parameter will contain outputs of the newly created
+ * filters.