Recherche avancée

Sur d’autres sites (14968)

• I want to get the data of the MPEG2ts data stream

  26 juin 2020, par 夏目たかし

  I want to get the data of the MPEG2ts data stream using FFMPEG.

  


  Delivery side :

  


  ffmpeg -re -i hoge.ts -map 0 -c copy -f mpegts udp://127.0.0.1:5004


  


  Receiver :

  


  ffmpeg -i udp://127.0.0.1:5004 -map 0:1 -codec copy -f data stream.txt


  


  I started two command prompts and both were running on one PC.

  


  disp Delivery side message :

  


  Input #0, mpegts, from 'E:/Day_Flight.mpg':
  Duration: 00:03:14.88, start: 10.000000, bitrate: 4187 kb/s
  Program 1
    Stream #0:0[0x1e1]: Video: h264 (Main) ([27][0][0][0] / 0x001B), 
yuv420p(progressive), 1280x720, 60 fps, 60 tbr, 90k tbn, 180k tbc
    Stream #0:1[0x1f1]: Data: klv (KLVA / 0x41564C4B)
Output #0, mpegts, to 'udp://127.0.0.1:5004':
  Metadata:
    encoder         : Lavf58.47.100
    Stream #0:0: Video: h264 (Main) ([27][0][0][0] / 0x001B), yuv420p(progressive), 1280x720, q=2-31, 60 fps, 60 tbr, 90k tbn, 90k tbc
    Stream #0:1: Data: klv (KLVA / 0x41564C4B)
Stream mapping:
  Stream #0:0 -> #0:0 (copy)
  Stream #0:1 -> #0:1 (copy)


  


  disp Receiver message :

  


  Input #0, mpegts, from 'udp://127.0.0.1:5004':
  Duration: N/A, start: 1.400000, bitrate: N/A
  Program 1
    Metadata:
      service_name    : Service01
      service_provider: FFmpeg
    Stream #0:0[0x100]: Video: h264 (Main) ([27][0][0][0] / 0x001B), 
yuv420p(progressive), 1280x720, 60 fps, 60 tbr, 90k tbn, 180k tbc
    Stream #0:1[0x101]: Data: klv (KLVA / 0x41564C4B)
Output #0, data, to 'stream.txt':
  Metadata:
    encoder         : Lavf58.47.100
    Stream #0:0: Data: klv (KLVA / 0x41564C4B)
Stream mapping:
  Stream #0:1 -> #0:0 (copy)


  


  However, the contents of steram.txt are empty.
Please give me some advice.

  


• Is there a way to save a matplotlib animation as a video (with ffmpeg) so that the last frame is held for N seconds ?

  2 décembre 2020, par janluke

  I'm using FFMpegWriter to save a matplotlib animation as a video. I'd like to hold the last frame for some seconds at the end of the video.

  


  As a workaround, one could modify the animation itself by repeating the last frame for a number of extra steps. This number of steps can be computed as a function of the desired "hold duration" and the interval/fps of the animation.

  


  Nonetheless, I'd like to know if there's a cleaner way to do the same without artificially modifying the animation itself and using instead some extra arguments for the writer (which uses ffmpeg in my case). Unfortunately I don't know much of ffmpeg so I'd like to have some help.

  


  Thank you.

  


• PIL image save causes FFMPEG to fail

  6 janvier 2023, par Xorgon

  I have been attempting to convert some videos using FFMPEG with image2pipe using PIL. I have found that when the frame is particularly simple (such as all one colour), it causes FFMPEG to fail with the following message :

  


  [image2pipe @ 000001785b599bc0] Could not find codec parameters for stream 0 (Video: none, none): unknown codec&#xA;Consider increasing the value for the &#x27;analyzeduration&#x27; and &#x27;probesize&#x27; options&#xA;Input #0, image2pipe, from &#x27;pipe:&#x27;:&#xA;  Duration: N/A, bitrate: N/A&#xA;    Stream #0:0: Video: none, none, 24 tbr, 24 tbn, 24 tbc&#xA;Output #0, mp4, to &#x27;<your filepath="filepath" here="here">/test.mp4&#x27;:&#xA;Output file #0 does not contain any stream&#xA;</your>

  


  The minimum code I have found to reproduce this is as follows :

  


  import numpy as np&#xA;from subprocess import Popen, PIPE&#xA;from PIL import Image&#xA;&#xA;output_file = "<your filepath="filepath" here="here">/test.mp4"&#xA;&#xA;p = Popen([&#x27;ffmpeg&#x27;,&#xA;           &#x27;-y&#x27;,  # Overwrite files&#xA;           &#x27;-f&#x27;, &#x27;image2pipe&#x27;,  # Input format&#xA;           &#x27;-r&#x27;, &#x27;24&#x27;,  # Framerate&#xA;           &#x27;-i&#x27;, &#x27;-&#x27;,  # stdin&#xA;           &#x27;-c:v&#x27;, &#x27;libx264&#x27;,  # Codec&#xA;           &#x27;-preset&#x27;, &#x27;slow&#x27;,&#xA;           &#x27;-crf&#x27;, f&#x27;18&#x27;,  # H264 Constant Rate Factor (quality, lower is better)&#xA;           output_file], stdin=PIPE)&#xA;&#xA;# This one works&#xA;# vid = np.random.randint(0, 255, (10, 64, 64))  # Create a 64x64 &#x27;video&#x27; with 10 frames of random noise&#xA;&#xA;# This one does not&#xA;vid = np.full((10, 64, 64), 129)  # Create a 64x64 &#x27;video&#x27; with 10 frames of pure grey&#xA;&#xA;for frame in vid:&#xA;    im = Image.fromarray(np.uint8(frame))&#xA;    im.save(p.stdin, &#x27;JPEG&#x27;)&#xA;&#xA;p.stdin.close()&#xA;p.wait()&#xA;</your>

  


  Notably, if I do the same thing with a randomly generated series of frames (commented as "This one works" in the script above), it will output fine.

  


  One workaround I have found so far is to replace 'JPEG' with 'PNG' in the im.save(...) call. However, I would be interested in understanding what causes it to fail with JPEG.