Recherche avancée

Sur d’autres sites (7741)

• Rescaling and slowing down a movie at the same time with ffmpeg

  21 décembre 2016, par ejl62

  I would like with ffmpeg to slow down a movie I am creating using the flag :

  -filter:v "setpts=2.0*PTS"

  However the height of my still images is not divisible by 2, so to avoid the error : height not divisible by 2 (1238x833), I am using the flag :

  -vf scale="trunc(iw/2)*2:trunc(ih/2)*2"

  (I also tried -vf scale=1238:-2).

  When I do this the film is generated but it isn’t slowed down, like if the -filter:v "setpts=2.0*PTS" wasn’t there.

  Is there something particular to do in order to have both option working at the same time ?

  Here is the complete command I am using :

  ffmpeg -an -i ./movie/cphmd1.%05d.ppm -vcodec libx264 -pix_fmt yuv420p -b:v 5000k -r 24 -crf 18 -filter:v "setpts=2.0*PTS" -vf scale="trunc(iw/2)*2:trunc(ih/2)*2" -preset slow -f mp4 cphmd1_slower.mp4

  Many thanks in advance !

• How to convert FFmpeg command to FFmpeg-Python code ?

  11 mars 2024, par Валентин Арно

  I have this command-line code :

  


  ffmpeg -hide_banner -hwaccel_device auto -hwaccel auto -thread_queue_size 2048 -f dshow -rtbufsize 2048M -pixel_format bgr24 -i video="screen-capture-recorder"  -thread_queue_size 2048 -f dshow -rtbufsize 2048M -channel_layout stereo -i audio="CABLE Output (VB-Audio Virtual Cable)" -filter_complex "[0:v]crop=1920:1040:0:0,scale=in_range=full:out_range=full:eval=init:interl=false:flags=bitexact+accurate_rnd+full_chroma_int,fps=fps=30.000,pp=fa[v1];[1:a]volume=volume=1.2[a1]" -map "[v1]" -map "[a1]" -c:a copy -c:v libx264 -preset ultrafast -tune film  -crf 17 -r 30 -force_key_frames expr:gte(t,n_forced*1) -sc_threshold 0 -pix_fmt yuvj422p -max_muxing_queue_size 2048 -copyts -start_at_zero -y -t 5 "Y:\test\ output_1".mkv


  


  Error : Stream specifier ':a' in filtergraph description ...matches no streams.

  


  And I want to convert it to ffmpeg-python code in Python.

  


  But how can I do it ?

  


  This is what I have done so far :

  


  video_ = ffmpeg.input('video=screen-capture-recorder',
                     thread_queue_size=2048,
                     rtbufsize='2048M',
                     pixel_format='bgr24',
                     framerate=30,
                     f='dshow'
                     )

audio_ = ffmpeg.input('audio=virtual-audio-capturer',
                     thread_queue_size=2048,
                     rtbufsize='2048M',
                     channel_layout='stereo',
                     f='dshow'
                     )
print(' '.join(  
      ffmpeg
      .filter(video_,'fps', fps=30.000)
      .filter(video_,'crop', 1920,1040,0,0)
      .filter(video_,'pp','fa')
      .filter(video_,'scale', in_range='full', out_range ='full', eval='init', interl='false', flags='bitexact+accurate_rnd+full_chroma_int')
      .filter(audio_,'volume',volume=1.2)

      .output('Y:\\test\\output_1.mkv', acodec='copy', vcodec="libx264", preset='ultrafast',
              tune='film', crf=17, r=30, force_key_frames='expr:gte(t,n_forced*1)',
              sc_threshold=0, pix_fmt='yuv420p', max_muxing_queue_size=2048,
              start_at_zero=None, t=15)
      
      .global_args('-hide_banner')
      .global_args('-hwaccel_device', 'auto')
      .global_args('-hwaccel', 'auto')
      .global_args('-report')
      .overwrite_output()
      .compile()
))


  


  Thanks.

  


• avcodec/h264dec : fix possible out-of-bounds array access

  25 août 2021, par Niklas Haas

  avcodec/h264dec : fix possible out-of-bounds array access
  If slice_type is > 9, the access to ff_h264_golomb_to_pict_type is
  
  out-of-bounds. Fix this by simply setting the slice_type to 0 in this
  
  case.
  This is completely inconsequential because the value is only being used
  
  to being used as an offset in the calculation of the film grain seed
  
  value, a corruption of which is practically invisible.
  Fixes coverity ticket #1490802
  Signed-off-by : James Almer <jamrial@gmail.com>
  

  • [DH] libavcodec/h264dec.c