Recherche avancée

Sur d’autres sites (6489)

• FFMPEG add circular mask to videos, convert to black and white and concatenate

  3 mai 2018, par Yassine

  Hello everyone i’m a beginner and i would appreciate your help.

  I’m making a mobile application that generates custom video resumes based on the user’s videos taken from his phone, the user has to upload 5 different videos to the server from the mobile application, in the server side i want to :

  • Add a .png circular mask to each video.
  • Make each video black and white.

  • Concatenate the videos with other already existing title videos
    (e.g [userVideo1] [title1] [userVideo2] [title2]...) Visual Example

    [Edit : I would like more features]

  • Add background music
  • Add watermark logo in the middle
  • Remove silent footage from the beginning and from the end
  • Some input videos might be rotated, i want to rotate videos back to normal if they are rotated.

  So far i managed to add the circular mask, make the videos black and white and concatenate 3 videos including a premade title video, but the second user video has no sound in the output.

  This is the script i ended up with :

  ffmpeg -i uservid1.mov -i uservid2.mp4 -i mask.png -i title1.mp4  -preset 
  
  ultrafast -filter_complex "
  
  
  
  [2:v][0:v]scale2ref[s1][s2];     
  
  [s2][s1]overlay[vid1];  
  
  [2:v][1:v]scale2ref[s3][s4]; 
  
  [s4][s3]overlay[vid2];  
  
  [vid1]hue=s=0[v0]; 
  
  [vid2]hue=s=0[v1];  
  
  [v0]scale=720x400[in0]; 
  
  [v1]scale=720x400[in1]; 
  
  [3:v]scale=720x400[in3]; 
  
  [in0]setsar=sar=0[final0]; 
  
  [in1]setsar=sar=0[final1];      
  
  [in3]setsar=sar=0[final3]; 
  
  [final0][final3][final1]concat=n=3;"
  
  
  
  -codec:a copy finalCV.mp4  

• TypeError : expected str, bytes or os.PathLike object, not module when trying to sream openCv frames to rtmp server

  30 novembre 2022, par seriously

  I am using openCv and face-recognition api to detect a face using a webcam then compare it with a previously taken image to check and see if the people on both images are the same and the openCv and face-recognition part of the code works properly now what I am trying to achieve is to stream the openCv processed video frames to an rtmp server so for this I am trying to use ffmpeg and running the command using subprocess but when I run the code I get error TypeError: expected str, bytes or os.PathLike object, not module. But I am writing the frames as bytes to stdin hence p.stdin.write(frame.tobytes()). How can I fix it and properly stream my openCv frames to an rtmp server using ffmpeg. Thanks in advance.

  


  Traceback (most recent call last):&#xA;  File "C:\Users\blah\blah\test.py", line 52, in <module>&#xA;    p = subprocess.Popen(command, stdin=subprocess.PIPE, shell=False)&#xA;  File "C:\Python310\lib\subprocess.py", line 969, in __init__&#xA;    self._execute_child(args, executable, preexec_fn, close_fds,&#xA;  File "C:\Python310\lib\subprocess.py", line 1378, in _execute_child&#xA;    args = list2cmdline(args)&#xA;  File "C:\Python310\lib\subprocess.py", line 561, in list2cmdline&#xA;    for arg in map(os.fsdecode, seq):&#xA;  File "C:\Python310\lib\os.py", line 822, in fsdecode&#xA;    filename = fspath(filename)  # Does type-checking of `filename`.&#xA;TypeError: expected str, bytes or os.PathLike object, not module&#xA;</module>

  


  import cv2
import numpy as np
import face_recognition
import os
import subprocess
import ffmpeg

path = '../attendance_imgs'
imgs = []
classNames = []
myList = os.listdir(path)

for cls in myList:
    curruntImg = cv2.imread(f'{path}/{cls}')
    imgs.append(curruntImg)
    classNames.append(os.path.splitext(cls)[0])

def findEncodings(imgs):
    encodeList = []
    for img in imgs:
        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
        encode = face_recognition.face_encodings(img)[0]
        encodeList.append(encode)
    return encodeList

encodeListKnown = findEncodings(imgs)
print('Encoding Complete')

cap = cv2.VideoCapture(0)

rtmp_url = "rtmp://127.0.0.1:1935/stream/webcam"

fps = int(cap.get(cv2.CAP_PROP_FPS))
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))

# command and params for ffmpeg
command = [ffmpeg,
           '-y',
           '-f', 'rawvideo',
           '-vcodec', 'rawvideo',
           '-pix_fmt', 'bgr24',
           '-s', "{}x{}".format(width, height),
           '-r', str(fps),
           '-i', '-',
           '-c:v', 'libx264',
           '-pix_fmt', 'yuv420p',
           '-preset', 'ultrafast',
           '-f', 'flv',
           'rtmp://127.0.0.1:1935/stream/webcam']

p = subprocess.Popen(command, stdin=subprocess.PIPE, shell=False)


while True:
    ret, frame, success, img = cap.read()
    if not ret:
        print("frame read failed")
        break
    imgSmall = cv2.resize(img, (0,0), None, 0.25, 0.25)
    imgSmall = cv2.cvtColor(imgSmall, cv2.COLOR_BGR2RGB)

    currentFrameFaces = face_recognition.face_locations(imgSmall)
    currentFrameEncodings = face_recognition.face_encodings(imgSmall, currentFrameFaces)

    for encodeFace, faceLocation in zip(currentFrameEncodings, currentFrameFaces):
        matches = face_recognition.compare_faces(encodeListKnown, encodeFace)
        faceDistance = face_recognition.face_distance(encodeListKnown, encodeFace)
        matchIndex = np.argmin(faceDistance)

        if matches[matchIndex]:
            name = classNames[matchIndex].upper()
            y1, x2, y2, x1 = faceLocation
            y1, x2, y2, x1 = y1 * 4, x2 * 4, y2 * 4, x1 * 4 
            cv2.rectangle(img, (x1, y1), (x2, y2), (0, 255, 0), 2)
            cv2.rectangle(img, (x1, y2 - 35), (x2, y2), (0, 255, 0), cv2.FILLED)
            cv2.putText(img, name, (x1 + 6, y2 - 6), cv2.FONT_HERSHEY_DUPLEX, 1, (255, 255, 255), 2) 

    # write to pipe
    p.stdin.write(frame.tobytes())


  


• FFMPEG MP3 file size much larger than expected on Windows 10

  8 avril 2018, par The Gora

  I’ve been using FFMPEG on Windows to :

  1. Convert iTunes M4A files to MP3s (with a bit rate of 128k) ; and
  2. Create 30 sec sample MP3s of the above MP3s (same bit rate).

  When I run FFMEG on a Windows 7 64 bit machine, the size of the MP3s (both for 1. & 2.) is in line with the rough calculation of :

  (Audio length in seconds) X (Bit rate)

  For example, a 4 minute audio yields an approx. 3.7MB MP3 file ; a 30 second sample MP3 is approx. 470KB.

  However when I run the same FFMPEG binary (copied from the Windows 7 machine) on a Windows 10 64 bit machine, all of the MP3s (both for 1. and for 2.) are inflated by approx 5MB. I’m using the same batch file on both machines to execute FFMEG with the required parameters, so pretty confident the difference is not down to user error.

  My questions are :

  1. Why is there this apparent 5MB overhead on Windows 10 ? and more importantly ;
  2. What can I do to remove this ?

  The large file size is a problem as the sample MP3s are to be put on a website for people to listen to a snippet of the song, and the webpage with multiple tags takes a long time to load completely (several minutes).

  Here is the version and lib info :

  ffmpeg version 3.4.1 Copyright (c) 2000-2017 the FFmpeg developers
 built with gcc 7.2.0 (GCC)
 configuration: --enable-gpl --enable-version3 --enable-sdl2 --enable-bzlib --enable-fontconfig --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libfreetype --enable-libmp3lame --enable-libopenjpeg --enable-libopus --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libtheora --enable-libtwolame --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxml2 --enable-libzimg --enable-lzma --enable-zlib --enable-gmp --enable-libvidstab --enable-libvorbis --enable-cuda --enable-cuvid --enable-d3d11va --enable-nvenc --enable-dxva2 --enable-avisynth --enable-libmfx
 libavutil      55. 78.100 / 55. 78.100
 libavcodec     57.107.100 / 57.107.100
 libavformat    57. 83.100 / 57. 83.100
 libavdevice    57. 10.100 / 57. 10.100
 libavfilter     6.107.100 /  6.107.100
 libswscale      4.  8.100 /  4.  8.100
 libswresample   2.  9.100 /  2.  9.100
 libpostproc    54.  7.100 / 54.  7.100

  And here are the command lines I’m using :

  1. ffmpeg -i input.m4a -id3v2_version 3 -b:a 128k -output.mp3
  2. ffmpeg -i input.m4a -ss 30 -t 30 -af "afade=in:st=30:d=5,afade=out:st=55:d=5" -id3v2_version 3 -b:a 128k -output.mp3