64 lines
2.3 KiB
ReStructuredText
64 lines
2.3 KiB
ReStructuredText
|
H264 Decoding
|
||
|
|
=================================
|
||
|
|
|
||
|
|
:link_to_translation:`zh_CN:[中文]`
|
||
|
|
|
||
|
|
1.H264 Decoding
|
||
|
|
--------------------
|
||
|
|
|
||
|
|
The H264 decoding part is implemented on the mobile phone. This article uses the ffmpeg library for decoding
|
||
|
|
|
||
|
|
2.ffmpeg Official Website
|
||
|
|
----------------------------------------
|
||
|
|
|
||
|
|
https://ffmpeg.org/
|
||
|
|
|
||
|
|
3.Decoding Process Diagram
|
||
|
|
----------------------------------------
|
||
|
|
|
||
|
|
.. figure:: ../../../_static/ffmpeg_h264_decode_flow.png
|
||
|
|
:align: center
|
||
|
|
:alt: fill specific color
|
||
|
|
:figclass: align-center
|
||
|
|
|
||
|
|
Figure 1. decode_flow
|
||
|
|
|
||
|
|
4.Process Explanation
|
||
|
|
-------------------------------
|
||
|
|
|
||
|
|
1. Register Decoder:``avcodec_register_all``
|
||
|
|
|
||
|
|
Used to register all compiled codecs.
|
||
|
|
|
||
|
|
2. Find Decoder:``avcodec_find_decoder``
|
||
|
|
|
||
|
|
Searches for the decoder corresponding to the specified decoder ID. The decoder ID for H.264 is: AV_CODEC_ID_H264.
|
||
|
|
|
||
|
|
3. Initialize Parser:``av_parser_init``
|
||
|
|
|
||
|
|
Initializes the parser context. The parser can be used for preprocessing input data before processing the data stream, such as finding frame boundaries, extracting key information, etc.
|
||
|
|
|
||
|
|
4. Allocate Decoder Context:``avcodec_alloc_context3``
|
||
|
|
|
||
|
|
Allocates an AVCodecContext structure and sets default values. AVCodecContext is a data structure in ffmpeg used to store the context information of the codec.
|
||
|
|
|
||
|
|
5. Open Decoder:``avcodec_open2``
|
||
|
|
|
||
|
|
Opens the decoder and initializes its context (AVCodecContext). After initializing the decoder context, this function opens the decoder to start the decoding process.
|
||
|
|
|
||
|
|
6. Receive Data
|
||
|
|
|
||
|
|
Before starting decoding each time, check if the data starts with an sps frame. Only start decoding when the data with sps beginning is found; otherwise, discard the data without decoding.
|
||
|
|
|
||
|
|
7. Send Data to Parser:``av_parser_parse2``
|
||
|
|
|
||
|
|
Parses the input media data stream, usually used to find frame boundaries.
|
||
|
|
|
||
|
|
8. Send Data to Decoder:``avcodec_send_packet``
|
||
|
|
|
||
|
|
Used to send data packets to the decoder for decoding. It sends the input data packet to the decoder context for decoding.
|
||
|
|
|
||
|
|
9. Receive Decoded Data Frames:``avcodec_receive_frame``
|
||
|
|
|
||
|
|
Used to receive the decoded frames from the decoder. After using the ``avcodec_send_packet`` function to send data packets for decoding, ``avcodec_receive_frame`` can be used to obtain the decoded image frames from the decoder.
|