Recherche avancée

Sur d’autres sites (3254)

• Add unique color watermark per frame with FFMpeg

  26 octobre 2018, par Dustin Kerstein

  I’ve been able to add a random color watermark with this code :

  ffmpeg -y -r 100 -i "N%3d.tif" -c:v libx264 -vf "drawbox=y=0:color=random@1:width=8:height=ih:t=fill,scale=1920:1080" -crf 30 -g 10 -profile:v high -level 4.1 -pix_fmt yuv420p test.mp4

  And I know that it’s doable with a script and processing each input frame individually, but I would really like to find a way with FFMpeg to add the watermark during the actual video encoding. It needs to be a unique color per frame. Any ideas on how to accomplish this ?

  Thanks !

• ValueError : Image size of 100010x50004 pixels is too large. It must be less than 2^16 in each direction

  3 avril 2021, par Slow

  I am using matplotlib.pyplot to plot an animation video from data files.
Below is my code, I am getting this error each time I run the code.

  


  


  ValueError : Image size of 100010x50004 pixels is too large. It must be
less than 2^16 in each direction.

  


  


  According to some suggestions, I have restarted my kernel and update it, but still getting this issue.

  


  import matplotlib
matplotlib.use('Qt5Agg')
from matplotlib.animation import FFMpegWriter
import matplotlib.pyplot as plt
import numpy as np
import readArray as ra
# import polygon as poly

def fieldMovie( fid, sx, sy, dt):
    """
    save the field and vesicle deformation movie
    
    input: 
        fid: name
        sx, sy: coordinates
        dt: time step

    Returns
    -------
    None.

    """
    # sx = ra.readArray('sx.afdat')
    # sy = ra.readArray('sy.afdat') 
    xbon = ra.readArray('xbound.afdat')
    ybon = ra.readArray('ybound.afdat')
    
    Tend = int(len(sx)/1)
    m = 1350
    n = 135
    x = np.arange(0, m)
    y = np.arange(0, n)
    X, Y = np.meshgrid(x, y)
    
    metadata = dict(title='field', artist='Matplotlib', 
                    comment='draw field')
    writer = FFMpegWriter(fps=7, metadata=metadata)
    
    fig = plt.figure(figsize=(10, 5))
    with writer.saving(fig, fid, Tend):
        for i in range(0, 9999, 2):
            time = dt*i
            fig.clf()
            sx_this = sx[i]
            sx_this = np.r_[sx_this, [sx_this[0,:]]]
            sy_this = sy[i]
            sy_this = np.r_[sy_this, [sy_this[0,:]]]
            sx_min = np.min(sx_this, axis=0)
            sx_this = sx_this - np.floor(sx_min/m)*m
        
            plt.fill(sx_this, sy_this, 'r', linewidth=1.5)
            sx_this = sx_this - m
            plt.fill(sx_this, sy_this, 'r', linewidth=1.5)
            plt.plot(xbon, ybon, '.' ,color = 'k' , markersize=1.0)
            plt.axis('equal')
            plt.axis('on')
            plt.xlim(0, m-1)
            title1 = 'test '
            title2 = 'at T = %.4f s, Ht = 18%%' % time
            plt.title(title1+title2, fontsize=16, fontweight='bold')
            ax = plt.gca()
            ax.set_position([0.05, 0.05, 0.8, 0.85])
            plt.pause(0.1)
            writer.grab_frame()
            plt.axes().set_aspect('equal')
            
sx = ra.readArray('sx.afdat')
sy = ra.readArray('sy.afdat')
dt = 0.002 * 1.5e-7 / 0.0003 * 700001 / 10000
fieldMovie('Animation.mp4', sx, sy, dt)


  


• How to add black borders to video

  7 décembre 2023, par Arnaud Rochez

  So I'm using ffmpeg to convert a video to 19201080 px, I found two ways to do so, the first one would be to stretch the video to 19201080, but then it looks kinda stretched. I used this command for this :

  


  ./ffmpeg_darwin -i SRC -vf scale=1920:1080,setdar=16:9 DEST


  


  The other option is the same without setdar but this just adapts the resolution to the one it started from (1728*1080).

  


  I would like to fill the 192 pixels of the width with a black border. Is there some kind of option to do so ? Or is there maybe another command line that could achieve this ?