☰

从 ffmpeg/ffprobe 检索每帧时间差

tag-icon • tag-icon macos ffmpeg video ffprobe
从 ffmpeg/ffprobe 检索每帧时间差

我正在尝试从 iPhone 视频中重建物体的速度信息，因此我对确定帧之间的时间差很感兴趣。据我所知，iPhone（HEVC 编码，.mov 文件）以可变帧速率拍摄。由于需要高精度，我无法在恒定帧速率视频的假设下工作。到目前为止，我只能恢复有效反映 CFR 视频的 pts_time 值，但根据其他帖子，这似乎是可能的。我正在 macOS Monterey 12.6 上使用 ffmpeg 5.1.1。

这是我尝试过的：

  1. 从命令行运行“已确认”的 VFR 视频mediainfo input.movw 输出：
Video
ID                                       : 1
Format                                   : HEVC
Format/Info                              : High Efficiency Video Coding
Format profile                           : Main@L5@Main
Codec ID                                 : hvc1
Codec ID/Info                            : High Efficiency Video Coding
Duration                                 : 1 s 372 ms
Source duration                          : 1 s 947 ms
Bit rate                                 : 32.2 Mb/s
Width                                    : 3 840 pixels
Height                                   : 2 160 pixels
Display aspect ratio                     : 16:9
Rotation                                 : 180°
Frame rate mode                          : Variable
Frame rate                               : 25.685 FPS
Minimum frame rate                       : 7.500 FPS
Maximum frame rate                       : 75.000 FPS
Color space                              : YUV
Chroma subsampling                       : 4:2:0
Bit depth                                : 8 bits
Bits/(Pixel*Frame)                       : 0.151
Stream size                              : 4.78 MiB (91%)
Source stream size                       : 5.23 MiB (99%)
Title                                    : Core Media Video
Encoded date                             : UTC 2022-08-27 18:44:20
Tagged date                              : UTC 2022-08-27 18:44:20
Color range                              : Limited
Color primaries                          : BT.709
Transfer characteristics                 : BT.709
Matrix coefficients                      : BT.709
Codec configuration box                  : hvcC

Audio
ID                                       : 2
Format                                   : AAC LC
Format/Info                              : Advanced Audio Codec Low Complexity
Codec ID                                 : mp4a-40-2
Duration                                 : 1 s 372 ms
Source duration                          : 1 s 440 ms
Bit rate mode                            : Variable
Bit rate                                 : 181 kb/s
Channel(s)                               : 2 channels
Channel layout                           : L R
Sampling rate                            : 44.1 kHz
Frame rate                               : 43.066 FPS (1024 SPF)
Compression mode                         : Lossy
Stream size                              : 30.3 KiB (1%)
Source stream size                       : 31.9 KiB (1%)
Title                                    : Core Media Audio
Encoded date                             : UTC 2022-08-27 18:44:20
Tagged date                              : UTC 2022-08-27 18:44:20

Other #1
Type                                     : meta
Duration                                 : 1 s 372 ms
Source duration                          : 9 s 245 ms
Stream size                              : 0.00 Byte
Source stream size                       : 10.0 Bytes

Other #2
Type                                     : meta
Duration                                 : 1 s 372 ms
Source duration                          : 9 s 245 ms
Stream size                              : 0.00 Byte
Source stream size                       : 8.00 Bytes
  1. 运行ffmpeg -i input.mov -vf vfrdet -an -f null -（显示它不是 VFR？）w 输出：
MacBook-Pro-6:timing_validation $ ffmpeg -i 30_4k-3ft.mov -vf vfrdet -an -f null -
ffmpeg version 5.1.1 Copyright (c) 2000-2022 the FFmpeg developers
  built with Apple clang version 13.1.6 (clang-1316.0.21.2.5)
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '30_4k-3ft.mov':
  Metadata:
    major_brand     : qt  
    minor_version   : 0
    compatible_brands: qt  
    creation_time   : 2022-08-27T18:44:20.000000Z
    com.apple.quicktime.make: Apple
    com.apple.quicktime.model: iPhone 13 Pro Max
    com.apple.quicktime.software: 15.0
    com.apple.quicktime.creationdate: 2022-08-27T11:21:43-0700
  Duration: 00:00:01.37, start: 0.000000, bitrate: 32207 kb/s
  Stream #0:0[0x1](und): Video: hevc (Main) (hvc1 / 0x31637668), yuv420p(tv, bt709), 3840x2160, 22539 kb/s, 25.68 fps, 30 tbr, 600 tbn (default)
    Metadata:
      creation_time   : 2022-08-27T18:44:20.000000Z
      handler_name    : Core Media Video
      vendor_id       : [0][0][0][0]
      encoder         : HEVC
    Side data:
      displaymatrix: rotation of -180.00 degrees
  Stream #0:1[0x2](und): Audio: aac (LC) (mp4a / 0x6134706D), 44100 Hz, stereo, fltp, 181 kb/s (default)
    Metadata:
      creation_time   : 2022-08-27T18:44:20.000000Z
      handler_name    : Core Media Audio
      vendor_id       : [0][0][0][0]
  Stream #0:2[0x3](und): Data: none (mebx / 0x7862656D), 0 kb/s (default)
    Metadata:
      creation_time   : 2022-08-27T18:44:20.000000Z
      handler_name    : Core Media Metadata
  Stream #0:3[0x4](und): Data: none (mebx / 0x7862656D), 0 kb/s (default)
    Metadata:
      creation_time   : 2022-08-27T18:44:20.000000Z
      handler_name    : Core Media Metadata
Stream mapping:
  Stream #0:0 -> #0:0 (hevc (native) -> wrapped_avframe (native))
Press [q] to stop, [?] for help
Output #0, null, to 'pipe:':
  Metadata:
    major_brand     : qt  
    minor_version   : 0
    compatible_brands: qt  
    com.apple.quicktime.creationdate: 2022-08-27T11:21:43-0700
    com.apple.quicktime.make: Apple
    com.apple.quicktime.model: iPhone 13 Pro Max
    com.apple.quicktime.software: 15.0
    encoder         : Lavf59.27.100
  Stream #0:0(und): Video: wrapped_avframe, yuv420p(tv, bt709, progressive), 3840x2160, q=2-31, 200 kb/s, 30 fps, 30 tbn (default)
    Metadata:
      creation_time   : 2022-08-27T18:44:20.000000Z
      handler_name    : Core Media Video
      vendor_id       : [0][0][0][0]
      encoder         : Lavc59.37.100 wrapped_avframe
    Side data:
      displaymatrix: rotation of -0.00 degrees
frame=    1 fps=0.0 q=-0.0 size=N/A time=00:00:00.03 bitrate=N/A speed=0.0622x  frame=   16 fps= 15 q=-0.0 size=N/A time=00:00:00.53 bitrate=N/A speed=0.498x   frame=   41 fps= 28 q=-0.0 Lsize=N/A time=00:00:01.36 bitrate=N/A speed=0.921x    
video:19kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
[Parsed_vfrdet_0 @ 0x7ff2c8836fc0] VFR:0.000000 (0/40)
  1. 运行ffprobe -select_streams v:0 -show_entries packet=pts_time,duration_time,stream_index input.mov显示与 CFR 有几个偏差（尽管 duration_time 没有反映这一点……），但总 pts_time 值与之前输出中显示的视频持续时间不匹配：
[PACKET]
stream_index=0
pts_time=-0.433333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.300000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.366667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.166667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.233333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.033333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=-0.100000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.100000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.033333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.000000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.066667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.233333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.166667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.133333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.200000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.366667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.300000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.266667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.333333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.500000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.433333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.400000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.466667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.633333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.566667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.533333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.600000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.766667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.700000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.666667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.733333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.900000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.833333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.800000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.866667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.033333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.966667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=0.933333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.000000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.166667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.100000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.066667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.133333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.300000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.233333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.200000
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.266667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.433333
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.366667
duration_time=0.033333
[/PACKET]
[PACKET]
stream_index=0
pts_time=1.333333
duration_time=0.033333
[/PACKET]

一些问题：

  1. 在此过程中是否存在丢帧的情况或者存在其他假设？

  2. 帧间时间差的精度有要求吗？如果精确到毫秒就好了。

  3. 如果这不现实，还有其他库/模块可以做到这一点吗？对任何可以从命令行或 python/C++ 运行的开源软件感兴趣。

相关内容