Recherche avancée

Sur d’autres sites (7496)

• Capturing audio and video from different sources, how to sync ?

  16 février 2017, par aerodavo

  Here is my code :

  Lapaki:~ Lapaki$ /Users/Lapaki/Desktop/ffmpeg -f avfoundation -video_size 960x540 -pixel_format uyvy422 -framerate ntsc -thread_queue_size 8B -i "XI:none" -f avfoundation -thread_queue_size 8B -i "none:XI" -vf 'crop=iw-240:ih:120:0' -af 'asetpts=PTS+.58735/TB' -pix_fmt yuv420p -aspect 4:3 -s 720x480 -q:v 3 -maxrate 5000k -bufsize 2000k -acodec ac3 -ac 2 -ab 256k -ar 48000 -f dvd /Users/Lapaki/Desktop/FF\ Test/`date +%F`\ `date +%H_%M_%S`.mpg
  
  ffmpeg version 3.2.3-tessus Copyright (c) 2000-2017 the FFmpeg developers
  
    built with Apple LLVM version 8.0.0 (clang-800.0.42.1)
  
    configuration: --cc=/usr/bin/clang --prefix=/opt/ffmpeg --extra-version=tessus --enable-avisynth --enable-fontconfig --enable-gpl --enable-libass --enable-libbluray --enable-libfreetype --enable-libgsm --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopus --enable-libschroedinger --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libvidstab --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libx265 --enable-libxavs --enable-libxvid --enable-libzmq --enable-version3 --disable-ffplay --disable-indev=qtkit --disable-indev=x11grab_xcb
  
    libavutil      55. 34.101 / 55. 34.101
  
    libavcodec     57. 64.101 / 57. 64.101
  
    libavformat    57. 56.101 / 57. 56.101
  
    libavdevice    57.  1.100 / 57.  1.100
  
    libavfilter     6. 65.100 /  6. 65.100
  
    libswscale      4.  2.100 /  4.  2.100
  
    libswresample   2.  3.100 /  2.  3.100
  
    libpostproc    54.  1.100 / 54.  1.100
  
  Input #0, avfoundation, from 'XI:none':
  
    Duration: N/A, start: 648413.295900, bitrate: N/A
  
      Stream #0:0: Video: rawvideo (UYVY / 0x59565955), uyvy422, 960x540, 29.97 fps, 29.97 tbr, 1000k tbn, 1000k tbc
  
  Input #1, avfoundation, from 'none:XI':
  
    Duration: N/A, start: 648413.884042, bitrate: 3072 kb/s
  
      Stream #1:0: Audio: pcm_f32le, 48000 Hz, stereo, flt, 3072 kb/s
  
  Output #0, dvd, to '/Users/Lapaki/Desktop/FF Test/2017-02-16 04_16_33.mpg':
  
    Metadata:
  
      encoder         : Lavf57.56.101
  
      Stream #0:0: Video: mpeg2video (Main), yuv420p, 720x480 [SAR 8:9 DAR 4:3], q=2-31, 200 kb/s, 29.97 fps, 90k tbn, 29.97 tbc
  
      Metadata:
  
        encoder         : Lavc57.64.101 mpeg2video
  
      Side data:
  
        cpb: bitrate max/min/avg: 5000000/0/200000 buffer size: 2000000 vbv_delay: -1
  
      Stream #0:1: Audio: ac3, 48000 Hz, stereo, fltp, 256 kb/s
  
      Metadata:
  
        encoder         : Lavc57.64.101 ac3
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (rawvideo (native) -> mpeg2video (native))
  
    Stream #1:0 -> #0:1 (pcm_f32le (native) -> ac3 (native))
  
  Press [q] to stop, [?] for help
  
  [swscaler @ 0x7f8e0c8ab400] Warning: data is not aligned! This can lead to a speedloss
  
  frame=   33 fps=0.0 q=3.0 size=     266kB time=00:00:01.06 bitrate=2051.9kbits/sframe=   49 fps= 48 q=3.0 size=     444kB time=00:00:01.54 bitrate=2358.8kbits/sframe=   64 fps= 42 q=3.0 size=     652kB time=00:00:02.08 bitrate=2560.5kbits/sframe=   79 fps= 39 q=3.0 size=     838kB time=00:00:02.59 bitrate=2642.4kbits/sframe=   94 fps= 37 q=3.0 size=    1022kB time=00:00:03.07 bitrate=2720.0kbits/sframe=  109 fps= 36 q=3.0 size=    1208kB time=00:00:03.59 bitrate=2756.5kbits/sframe=  124 fps= 35 q=3.0 size=    1406kB time=00:00:04.07 bitrate=2830.0kbits/sframe=  127 fps= 35 q=3.0 Lsize=    1474kB time=00:00:04.19 bitrate=2876.4kbits/s dup=12 drop=0 speed=1.15x    
  
  video:1310kB audio:113kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 3.604597%

  The start of Input #0 is 648413.295900, and the start of Input #1 is 648413.884042.

  I’ve been able to keep the audio and video in very close sync by subtracting the two values (which I assume are wallclock values), and using the asetpts audio filter to delay the audio stream of the recorded mpeg-2 file by that amount.

  I’d like to be able to do this exactly though, and that value changes slightly every time I start a new capture. Not to mention, I’d like to be able to do this reliably on different machines, where I assume the value will most likely be different, thus using a calculation as opposed to a fixed number is obviously the best option, if it’s possible.

  Is there a way to subtract the wallclock start time of input #0 from the wallclock start time of input #1 ? I’d like to do this inside the asetpts filter, instead of manually finding the difference from a previous run, which again is slightly different every time...

  I was thinking something like -af asetpts=PTS-([1:0]RTCSTART-[0:0]RTCSTART)/TB might work, but I have no idea how to format it.

  Thanks in advance !

• arm : vp9itxfm16 : Make the larger core transforms standalone functions

  24 février 2017, par Martin Storsjö

  arm : vp9itxfm16 : Make the larger core transforms standalone functions
  This work is sponsored by, and copyright, Google.
  This reduces the code size of libavcodec/arm/vp9itxfm_16bpp_neon.o from
  
  17500 to 14516 bytes.
  This gives a small slowdown of a couple tens of cycles, up to around
  
  150 cycles for the full case of the largest transform, but makes
  
  it more feasible to add more optimized versions of these transforms.
  Before :                                 Cortex A7       A8       A9      A53
  
  vp9_inv_dct_dct_16x16_sub4_add_10_neon :    4237.4   3561.5   3971.8   2525.3
  
  vp9_inv_dct_dct_16x16_sub16_add_10_neon :   6371.9   5452.0   5779.3   3910.5
  
  vp9_inv_dct_dct_32x32_sub4_add_10_neon :   22068.8  17867.5  19555.2  13871.6
  
  vp9_inv_dct_dct_32x32_sub32_add_10_neon :  37268.9  38684.2  32314.2  23969.0
  After :
  
  vp9_inv_dct_dct_16x16_sub4_add_10_neon :    4375.1   3571.9   4283.8   2567.2
  
  vp9_inv_dct_dct_16x16_sub16_add_10_neon :   6415.6   5578.9   5844.6   3948.3
  
  vp9_inv_dct_dct_32x32_sub4_add_10_neon :   22653.7  18079.7  19603.7  13905.3
  
  vp9_inv_dct_dct_32x32_sub32_add_10_neon :  37593.2  38862.2  32235.8  24070.9
  Signed-off-by : Martin Storsjö <martin@martin.st>
  

  • [DH] libavcodec/arm/vp9itxfm_16bpp_neon.S

• Resizing 360 stereoscopic video with FFMPEG

  13 mars 2018, par Miriam Tschanen

  I’m trying to prepare a 360 stereoscopic video from our VR app for streaming. We’d like to offer the video in different resolutions to accomodate varying internet speeds. The original file is 3840x2160 and 463MB large.

  I tried using the naive FFMPEG command :
  ffmpeg -i video_3840.mp4 -vf scale=2560:1440 video_2560.mp4

  However, this seems to remove the stereoscopic / 360 properties of the video, at any rate the Windows video player no longer lets me pan around the view and the file size is down to 74MB, which seems a bit extreme.

  Does anyone know which other flags I should set ? Note that I have absolutely no clue about filming or video formats, so I don’t even know what the original file is encoded as. Ideally I’d like to copy over as many settings as possible. The only thing I want to change is the resolution.

  Any help would be greatly appreciated.