Recherche avancée

Sur d’autres sites (6183)

• Matplotlib pipe canvas.draw() to ffmpeg - unexpected result [duplicate]

  31 juillet 2022, par Narusan

  I'm using this code from here to try and pipe multiple matplotlib plots into ffmpeg to write a video file :

  


  import numpy as np
import matplotlib.pyplot as plt
import subprocess

xlist = np.random.randint(100,size=100)
ylist = np.random.randint(100, size=100)
color = np.random.randint(2, size=100)

f = plt.figure(figsize=(5,5), dpi = 300)
canvas_width, canvas_height = f.canvas.get_width_height()
ax = f.add_axes([0,0,1,1])
ax.axis('off')


# Open an ffmpeg process
outf = 'ffmpeg.mp4'
cmdstring = ('ffmpeg',
    '-y', '-r', '30', # overwrite, 30fps
    '-s', '%dx%d' % (canvas_width, canvas_height), # size of image string
    '-pix_fmt', 'argb', # format
    '-f', 'rawvideo',  '-i', '-', # tell ffmpeg to expect raw video from the pipe
    '-vcodec', 'mpeg4', outf) # output encoding
p = subprocess.Popen(cmdstring, stdin=subprocess.PIPE)

# Draw 1000 frames and write to the pipe
for frame in range(10):
    print("Working on frame")
    # draw the frame
    f = plt.figure(figsize=(5,5), dpi=300)
    ax = f.add_axes([0,0,1,1])
    ax.scatter(xlist, ylist,
               c=color, cmap = 'viridis')
    f.canvas.draw()
    plt.show()

    # extract the image as an ARGB string
    string = f.canvas.tostring_argb()
    # write to pipe
    p.stdin.write(string)

# Finish up
p.communicate()


  


  While plt.show() does show the correct plot (see image below), the video that ffmpeg creates is a bit different than what plt.show() shows. I am presuming the issue is with f.canvas.draw(), but I'm not sure how to get a look at what canvas.draw() actually plots.

  


  plot.show() :


  


  ffmpeg video (imgur link)

  


• How to convert a Video to a Slideshow with synced audio ? [closed]

  25 mai 2013, par Henry Mazza

  I want a simple way to convert a Ted Talk Presentation to a SlideShow + (synced)Audio format so I can listen to it in my car. I don't want to lose most of the visuals as it does in the audio format and also want to reduce the storage/cell data needs of the video format.

  So far I already extracted the key scenes with timing and the audio, now I must glue this together in a synced fashion.

  Possible ways I found but couldn't make work :

  • MP4Box to make a .m4b (audiobook/enhanced podcast) with mp4chap to set each image as a chapter image (but I could find proper documentation on how to do this)
  • FFMPEG to make a flat movie with the images (but I couldn't make each image stay for a random period of time)

  Ultimately I will automate this process in my VPS and make my private podcast server, so no fancy tools that don't work in Linux, please.

  EDIT : actually the Podcast is a lousy idea, the artwork must be in a square aspect ratio and has great compatibility problems with varius players. Still studing the Book format.

• avformat/hlsenc : optimize help message default value.

  5 juillet 2017, par Steven Liu

  avformat/hlsenc : optimize help message default value.
  show the hls_segment_type default always 0, show the flag name better
  Signed-off-by : Steven Liu <lq@chinaffmpeg.org>
  

  • [DH] libavformat/hlsenc.c