From 54f7e79d4706a8343dad1d8da51b7d3d3b2cd3b2 Mon Sep 17 00:00:00 2001 From: Anton Khirnov Date: Tue, 4 Feb 2014 17:35:18 +0100 Subject: lavf: add some basic documentation of the muxing process --- libavformat/avformat.h | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+) (limited to 'libavformat/avformat.h') diff --git a/libavformat/avformat.h b/libavformat/avformat.h index 8ff4422f04..4084b39f86 100644 --- a/libavformat/avformat.h +++ b/libavformat/avformat.h @@ -173,6 +173,58 @@ * * @defgroup lavf_encoding Muxing * @{ + * Muxers take encoded data in the form of @ref AVPacket "AVPackets" and write + * it into files or other output bytestreams in the specified container format. + * + * The main API functions for muxing are avformat_write_header() for writing the + * file header, av_write_frame() / av_interleaved_write_frame() for writing the + * packets and av_write_trailer() for finalizing the file. + * + * At the beginning of the muxing process, the caller must first call + * avformat_alloc_context() to create a muxing context. The caller then sets up + * the muxer by filling the various fields in this context: + * + * - The @ref AVFormatContext.oformat "oformat" field must be set to select the + * muxer that will be used. + * - Unless the format is of the AVFMT_NOFILE type, the @ref AVFormatContext.pb + * "pb" field must be set to an opened IO context, either returned from + * avio_open2() or a custom one. + * - Unless the format is of the AVFMT_NOSTREAMS type, at least one stream must + * be created with the avformat_new_stream() function. The caller should fill + * the @ref AVStream.codec "stream codec context" information, such as the + * codec @ref AVCodecContext.codec_type "type", @ref AVCodecContext.codec_id + * "id" and other parameters (e.g. width / height, the pixel or sample format, + * etc.) as known. The @ref AVCodecContext.time_base "codec timebase" should + * be set to the timebase that the caller desires to use for this stream (note + * that the timebase actually used by the muxer can be different, as will be + * described later). + * - The caller may fill in additional information, such as @ref + * AVFormatContext.metadata "global" or @ref AVStream.metadata "per-stream" + * metadata, @ref AVFormatContext.chapters "chapters", @ref + * AVFormatContext.programs "programs", etc. as described in the + * AVFormatContext documentation. Whether such information will actually be + * stored in the output depends on what the container format and the muxer + * support. + * + * When the muxing context is fully set up, the caller must call + * avformat_write_header() to initialize the muxer internals and write the file + * header. Whether anything actually is written to the IO context at this step + * depends on the muxer, but this function must always be called. Any muxer + * private options must be passed in the options parameter to this function. + * + * The data is then sent to the muxer by repeatedly calling av_write_frame() or + * av_interleaved_write_frame() (consult those functions' documentation for + * discussion on the difference between them; only one of them may be used with + * a single muxing context, they should not be mixed). Do note that the timing + * information on the packets sent to the muxer must be in the corresponding + * AVStream's timebase. That timebase is set by the muxer (in the + * avformat_write_header() step) and may be different from the timebase the + * caller set on the codec context. + * + * Once all the data has been written, the caller must call av_write_trailer() + * to flush any buffered packets and finalize the output file, then close the IO + * context (if any) and finally free the muxing context with + * avformat_free_context(). * @} * * @defgroup lavf_io I/O Read/Write -- cgit v1.2.3