Ben Chuanlong Du's Blog

And let it direct your passion with reason.

Manipulate Videos Using MoviePy in Python

Installation

In [2]:
!pip3 install -q moviepy

Use moviepy to Manipulate Videos

Download a short video.

In [3]:
!wget --no-check-certificate https://www.sample-videos.com/video123/mp4/720/big_buck_bunny_720p_5mb.mp4 -O bunny.mp4

--2022-05-21 17:32:50--  https://www.sample-videos.com/video123/mp4/720/big_buck_bunny_720p_5mb.mp4
Resolving www.sample-videos.com (www.sample-videos.com)... 103.145.51.95
Connecting to www.sample-videos.com (www.sample-videos.com)|103.145.51.95|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 5253880 (5.0M) [video/mp4]
Saving to: ‘bunny.mp4’

bunny.mp4           100%[===================>]   5.01M  2.15MB/s    in 2.3s    

2022-05-21 17:32:53 (2.15 MB/s) - ‘bunny.mp4’ saved [5253880/5253880]
In [6]:
import numpy as np
from PIL import Image
import cv2
from moviepy.editor import VideoFileClip

Extract and check the first frame.

In [7]:
vidcap = cv2.VideoCapture("bunny.mp4")
In [8]:
success, arr = vidcap.read()
In [9]:
arr
Out[9]:
array([[[ 46, 112, 104],
        [ 31,  97,  89],
        [ 21,  92,  83],
        ...,
        [251, 223, 205],
        [251, 223, 205],
        [251, 223, 205]],

       [[ 44, 110, 102],
        [ 30,  96,  88],
        [ 22,  93,  84],
        ...,
        [251, 223, 205],
        [251, 223, 205],
        [251, 223, 205]],

       [[ 38, 104,  96],
        [ 26,  92,  84],
        [ 26,  97,  88],
        ...,
        [251, 223, 205],
        [251, 223, 205],
        [251, 223, 205]],

       ...,

       [[ 43, 131, 138],
        [ 57, 145, 152],
        [ 70, 153, 161],
        ...,
        [ 24, 117, 109],
        [ 27, 135, 116],
        [ 46, 154, 135]],

       [[ 54, 145, 151],
        [ 67, 158, 164],
        [ 72, 160, 167],
        ...,
        [ 27, 115, 108],
        [ 25, 128, 110],
        [ 46, 149, 131]],

       [[ 62, 153, 159],
        [ 68, 159, 165],
        [ 70, 158, 165],
        ...,
        [ 33, 121, 114],
        [ 28, 131, 113],
        [ 42, 145, 127]]], dtype=uint8)
In [11]:
arr.shape
Out[11]:
(720, 1280, 3)

The numpy array is the BGR (instead of RGB) representation of the image, which means that you have to flip the last dimension of the numpy array so that it can loaded using the Pillow library correctly.

In [10]:
Image.fromarray(np.flip(arr, 2))
Out[10]:
In [9]:
img.size
Out[9]:
(1280, 720)

Extract a sub video with part of the original frame.

In [12]:
video = VideoFileClip("bunny.mp4").subclip(0, 10).crop(x2=640)
video.write_videofile("sub.mp4")

Moviepy - Building video sub.mp4.
MoviePy - Writing audio in subTEMP_MPY_wvf_snd.mp3

                                                                    

MoviePy - Done.
Moviepy - Writing video sub.mp4


                                                               
Moviepy - Done !
Moviepy - video ready sub.mp4



















In [13]:



    


vidcap = cv2.VideoCapture("sub.mp4")



    













In [14]:



    


sucess, arr = vidcap.read()



    













In [15]:



    


img = Image.fromarray(np.flip(arr, 2))
img



    















    
Out[15]:


























In [16]:



    


img.size



    















    
Out[16]:








(640, 720)





















References













In [ ]:



    


 



    








 





    
  


  

    
Comments