最近在写一些跟视频解码相关的简单程序,学习 avcodec 技术时有一些强大的分析工具,分享一下。

码流与封装容器相关

ISO Viewer

isoviewer 是 Mac 下一个检查 mp4 文件构成的 GUI 程序,能直观地看到各个 box 的数据。

undefined

bento4

bento4 是一套专用于 mp4 格式文件分析和编辑的 SDK,支持的功能非常多。

工具 用途
mp4info displays high level info about an MP4 file, including all tracks and codec details
mp4dump displays the entire atom/box structure of an MP4 file
mp4edit add/insert/remove/replace atom/box items of an MP4 file
mp4extract extracts an atom/box from an MP4 file
mp4encrypt encrypts an MP4 file (multiple encryption schemes are supported)
mp4decrypt decrypts an MP4 file (multiple encryption schemes are supported)
mp4dcfpackager encrypts a media file into an OMA DCF file
mp4compact converts 'stsz' tables into 'stz2' tables to create more compact MP4 files
mp4fragment creates a fragmented MP4 file from a non-fragmented one.
mp4split splits a fragmented MP4 file into discrete files
mp4tag show/edit MP4 metadata (iTunes-style and others)
mp4mux multiplexes one or more elementary streams (H264/AVC, H265/HEVC, AAC) into an MP4 file
mp42aac extract a raw AAC elementary stream from an MP4 file
mp42avc extract a raw AVC/H.264 elementary stream from an MP4 file
mp42hls converts an MP4 file to an HLS (HTTP Live Streaming) presentation, including the generation of the segments and .m3u8 playlist as well as AES-128 and SAMPLE-AES (for Fairplay) encryption. This can be used as a replacement for Apple's mediafilesegmenter tool.
mp42ts converts an MP4 file to an MPEG2-TS file.
mp4dash creates an MPEG DASH output from one or more MP4 files, including encryption. This is a full-featured MPEG DASH packager.
mp4dashclone creates a local clone of a remote or local MPEG DASH presentation, optionally encrypting the segments as they are cloned.
mp4hls creates a multi-bitrate HLS master playlist from one or more MP4 files, including support for encryption and I-frame-only playlists. This tool uses the 'mp42hls' low level tool internally, so all the options supported by that low level tool are also available. This can be used as a replacement for Apple's variantplaylistcreator tool.

GPAC

GPAC 是一个为科研和学术领域开发的多媒体的框架,包括几个视频分析和码流编辑的 CLI 工具。官网主要 feature 如下:

GPAC features encoders and multiplexers, publishing and content distribution tools for MP4 and 3GPP or 3GPP2 files and many tools for scene descriptions (BIFS/VRML/X3D converters, SWF/BIFS, SVG/BIFS, etc…). MP4Box provides all these tools in a single command-line application. An exhaustive list of packaging features is available here but a good summary of what MP4Box can do for you is the following:

  • MP4/3GP Conversion from MP3, AVI, MEPG-2 TS, MPEG-PS, AAC, H263, H264, AMR, and many others,
  • 3GPP DIMS Packaging from SVG files,
  • File layout: fragmentation or interleaving, and cleaning,
  • File hinting for RTP/RTSP and QTSS/DSS servers (MPEG-4 / ISMA / 3GP / 3GP2 files),
  • File splitting by size or time, extraction from file and file concatenation,
  • XML information dumping for MP4 and RTP hint tracks,
  • Media Track extractions,
  • ISMA E&A encryption and decryption,
  • 3GPP timed text tools (SUB/SRT/TTXT/TeXML), VobSub import/export,
  • BIFS codec and scene conversion between MP4, BT and XMT-A,
  • LASeR codec and scene conversion between MP4, SAF, SVG and XSR (XML LASeR),
  • XML scene statistics for BIFS scene (BT, XMT-A and MP4),
  • Conversion to and from BT, XMT-A, WRL, X3D and X3DV with support for gzip.
  • Preparing MP4,3GP and MPEG-2 TS files for DASH streaming

主要使用的可能是 MP4Box 工具,功能很多,几个例子:

MP4Box -add file.av1 file_av1.mp4 # 将 av1 码流封装写入 mp4 文件

ffprobe

大名鼎鼎的 ffmpeg 项目一员,一身技能和 API 参数

其中一个用法是 dump 出所有的 frame:

ffprobe -show_frames some.mp4 > probe-mp4.txt
[FRAME]
media_type=video
stream_index=0
key_frame=1
pkt_pts=0
pkt_pts_time=0.000000
pkt_dts=0
pkt_dts_time=0.000000
best_effort_timestamp=0
best_effort_timestamp_time=0.000000
pkt_duration=512
pkt_duration_time=0.040000
pkt_pos=48
pkt_size=27979
width=320
height=564
pix_fmt=yuv420p
sample_aspect_ratio=N/A
pict_type=I
coded_picture_number=0
display_picture_number=0
interlaced_frame=0
top_field_first=0
repeat_pict=0
color_range=unknown
color_space=unknown
color_primaries=unknown
color_transfer=unknown
chroma_location=left
[/FRAME]
[FRAME]
media_type=audio
stream_index=1
key_frame=1
pkt_pts=0
pkt_pts_time=0.000000
pkt_dts=0
pkt_dts_time=0.000000
best_effort_timestamp=0
best_effort_timestamp_time=0.000000
pkt_duration=1024
pkt_duration_time=0.021333
pkt_pos=28243
pkt_size=6
sample_fmt=fltp
nb_samples=1024
channels=2
channel_layout=stereo
[/FRAME]

二进制显示/编辑

Hex Fiend

Hex Fiend, Mac 下的 hex 编辑 GUI 工具,支持按文本或hex搜索/编辑/diff,功能强大。

HexFiend

hexyl

hexyl ,一个轻量 CLI 查看二进制数据的工具,用 rust 编写。优点是快速,好看。

hexyl 效果图