Recherche avancée

Sur d’autres sites (4717)

• Overlay circular video with transparency with maskedmerge

  15 juillet 2017, par cgenco

  I have a square video from Snap Spectacles (1088x1088) that I want to overlay on itself zoomed in and blurred.

  Example input frame :

  

  Generated zoomed in and blurred background :

  

  Desired output :

  

  I think I can do this with ffmpeg’s maskedmerge, but I’m having trouble finding examples.

  There’s an example of maskedmerge that merges two videos of the same size and dynamically removes a green screen, and another that merges videos with transparency.

  Here’s the closest I’ve been able to get :

  ffmpeg -i background.jpg -vf "movie=input.jpg[inner];[in][inner] overlay=#{offset}:0 [out]" -c:a copy output.jpg

  

  tl ;dr : given the first two frames, how could I generate the third frame (as video) ?

• create a movie in python [on hold]

  22 mai 2017, par masoud

  I have some files with names (’Den_car_resample’ +’sdf’+ str(n)+’.dat’) where n changes from 0 to 25. I wrote a code to read these files and plots the results.
  now I want to create a movie from these plots. at the end of the program, I used the avconv command to do that.
  but, unfortunately my code creates a movie but it is empty.
  I don’t know the reason exactly but I think, first, I have to define a frame to each plot and then create a movie.
  can anyone please tell me how can I define a frame and also the add bit_rate of the movie.

  import sys
  
  import subprocess   
  
  import sdf
  
  import numpy as np
  
  import matplotlib.pyplot as plt  
  
  import time
  
  import matplotlib.animation as Animation
  
  from matplotlib.font_manager import FontProperties
  
  fp = FontProperties('Symbola')
  
  
  
  ##################### information from EPOCH input.deck
  
  nx,ny= 1200, 1600
  
  xmin=-100e-6
  
  xmax = 110e-6
  
  ymin = -200e-6
  
  ymax = 200e-6
  
  X =np.linspace(xmin,xmax,nx)   
  
  Y =np.linspace(ymin,ymax,ny)
  
  
  
  #################
  
  for n in range(0,25):
  
    nstr = str(n)
  
  
  
  ######################..... reading Density of carbon
  
    filename ="Den_car_resample" +'_sdf_'+ str(n)+'.dat'
  
  
  
    with open(filename, 'rb') as f:                        
  
      data = np.fromfile(f, dtype='float64', count=nx*ny)  
  
  
  
    Den_car  = np.reshape(data, [ny, nx], order='F')
  
    Den_car= np.log10(Den_car)
  
  
  
    ######################  Display Carbon density
  
  
  
    fig = plt.imshow(Den_car,  extent=[X.min()*1e6, X.max()*1e6, Y.min()*1e6,Y.max()*1e6], vmin=24, vmax=29, cmap='brg', aspect='auto') 
  
    plt.suptitle('Den_car')
  
    plt.title('sdf '+ str(n)+'; Time= '+str(n*50)+'ps',color='green', fontsize=15)
  
    plt.xlabel('x($\mu$m)')
  
    plt.ylabel('y($\mu$m)')
  
    plt.text(-80,-40,'Den_Carbon',color='red', fontsize=15)
  
    plt.colorbar()
  
    plt.savefig( 'fig%06d.png' % n, bbox_inches='tight')          
  
  
  
    plt.pause(.1)
  
    plt.clf()  
  
    plt.close()
  
  
  
    ######################  Create movie
  
  
  
    subprocess.call("avconv -framerate 1 -i fig%06d.png -c:v libx264 -profile:v high -crf 20".split())
  
  
  
  sys.exit()

• Remapping multiple Mp4 videos into a single one with ffmpeg

  21 septembre 2016, par Josep Bosch

  I’m interested in remapping multiple (6) MP4 videos into a high resolution final video according to lookup tables I calculated. The idea is convert 6 independent videos in a 360º video according to an equirectangular projection.

  
  Example of equirectangular video here

  Is there a way to do this remapping with ffmpeg or any other linux program ?
  Right now I’m extracting all the frames from the videos, creating the equirectangular individual images and joining them again into a video. There must be a better way for this...

  UPDATE :

  Following Mulyva’s suggestion I first remap each individual video using the remap filter. Those parts of the panoramic video not covered are interpreted like chromakey pixels using :

  ffmpeg -i videos/camera1.MP4 -i camera0_map_x_radius5.pgm -i camera0_map_y_radius5.pgm -lavfi remap -qscale 1 out0.MP4

  

  

  Then I try to overlay them using the chomakey filter :

   ffmpeg -i out0.MP4 -i out1.MP4 -filter_complex "[1:v]chromakey=0x12da11:0.2:0.2[chromakey_res];[0:v][chromakey_res]overlay=eof_action=pass[out]" -map "[out]" out.mp4

  As you can see, the final result has an undesirable green shadow. Any idea of how to remove it ?