Recherche avancée

Sur d’autres sites (9220)

• How to use FFMPEG on Python/Windows10 with Pipe for Screen recording ?

  20 septembre 2020, par Trmotta

  I'd like to record the screen with ffmpeg as it seems to be the only player out there who can record a region of the screen along with the mouse cursor.

  


  The following code was adapted from i want to display mouse pointer in my recording but it doesn't work on a Windows 10 (x64) setup (using Python 3.6).

  


  #!/usr/bin/env python3

# ffmpeg -y -pix_fmt bgr0 -f avfoundation -r 20 -t 10 -i 1 -vf scale=w=3840:h=2160 -f rawvideo /dev/null

import sys
import cv2
import time
import subprocess
import numpy as np

w,h = 100, 100

def ffmpegGrab():
    """Generator to read frames from ffmpeg subprocess"""

    #ffmpeg -f gdigrab -framerate 30 -offset_x 10 -offset_y 20 -video_size 640x480 -show_region 1 -i desktop output.mkv #CODE THAT ACTUALLY WORKS WITH FFMPEG CLI

    cmd = 'D:/Downloads/ffmpeg-20200831-4a11a6f-win64-static/ffmpeg-20200831-4a11a6f-win64-static/bin/ffmpeg.exe -f gdigrab -framerate 30 -offset_x 10 -offset_y 20 -video_size 100x100 -show_region 1 -i desktop -f image2pipe, -pix_fmt bgr24 -vcodec rawvideo -an -sn' 

    proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, shell=True)
    #out, err = proc.communicate()
    while True:
        frame = proc.stdout.read(w*h*3)
        yield np.frombuffer(frame, dtype=np.uint8).reshape((h,w,3))

# Get frame generator
gen = ffmpegGrab()

# Get start time
start = time.time()

# Read video frames from ffmpeg in loop
nFrames = 0
while True:
    # Read next frame from ffmpeg
    frame = next(gen)
    nFrames += 1

    cv2.imshow('screenshot', frame)

    if cv2.waitKey(1) == ord("q"):
        break

    fps = nFrames/(time.time()-start)
    print(f'FPS: {fps}')


cv2.destroyAllWindows()
out.release()


  


  By using 'cmd' as stated above, I'll get the following error :

  


  b"ffmpeg version git-2020-08-31-4a11a6f Copyright (c) 2000-2020 the FFmpeg developers\r\n  built with gcc 10.2.1 (GCC) 20200805\r\n  configuration: --enable-gpl --enable-version3 --enable-sdl2 --enable-fontconfig --enable-gnutls --enable-iconv --enable-libass --enable-libdav1d --enable-libbluray --enable-libfreetype --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libsrt --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-libvmaf --enable-libvorbis --enable-libvo-amrwbenc --enable-libmysofa --enable-libspeex --enable-libxvid --enable-libaom --enable-libgsm --enable-librav1e --enable-libsvtav1 --disable-w32threads --enable-libmfx --enable-ffnvcodec --enable-cuda-llvm --enable-cuvid --enable-d3d11va --enable-nvenc --enable-nvdec --enable-dxva2 --enable-avisynth --enable-libopenmpt --enable-amf\r\n  libavutil      56. 58.100 / 56. 58.100\r\n  libavcodec     58.101.101 / 58.101.101\r\n  libavformat    58. 51.101 / 58. 51.101\r\n  libavdevice    58. 11.101 / 58. 11.101\r\n  libavfilter     7. 87.100 /  7. 87.100\r\n  libswscale      5.  8.100 /  5.  8.100\r\n  libswresample   3.  8.100 /  3.  8.100\r\n  libpostproc    55.  8.100 / 55.  8.100\r\nTrailing option(s) found in the command: may be ignored.\r\n[gdigrab @ 0000017ab857f100] Capturing whole desktop as 100x100x32 at (10,20)\r\nInput #0, gdigrab, from 'desktop':\r\n  Duration: N/A, start: 1599021857.538752, bitrate: 9612 kb/s\r\n    Stream #0:0: Video: bmp, bgra, 100x100, 9612 kb/s, 30 fps, 30 tbr, 1000k tbn, 1000k tbc\r\n**At least one output file must be specified**\r\n"


  


  Which is the contents of proc (and also of proc.communicate). The program crashes right after when trying to resize this message to an image of size 100x100.

  


  I do not want to have an output file. I need to use Python subprocess along with Pipe in order to directly deliver those screen frames to my Python code, no IO required at all.

  


  If I try the following :

  


  cmd = 'D :/Downloads/ffmpeg-20200831-4a11a6f-win64-static/ffmpeg-20200831-4a11a6f-win64-static/bin/ffmpeg.exe -f gdigrab -framerate 30 -offset_x 10 -offset_y 20 -video_size 100x100 -i desktop -pix_fmt bgr24 -vcodec rawvideo -an -sn -f image2pipe'

  


  proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True)


  


  Then 'frame', inside 'while True', is filled with b''.

  


  Tried using the following libraries with no success, as I couldnt either find how to capture the mouse cursor or capture the screen at all : https://github.com/abhiTronix/vidgear, https://github.com/kkroening/ffmpeg-python

  


  What am I missing ?
Thank you.

  


• Is it possible to pipe an ffmpeg output with multiple files (HLS or DASH)

  29 août 2020, par New Dev

  I'm using FFmpeg to generate fragmented MP4s in a dash and HLS format.

  


  ffmpeg -i input.mov -f dash -seg_duration 6 -hls_playlist true output.mbd


  


  The above (simplified) command outputs multiple files, in addition to output.mdb (e.g. actual segments, master.m3u8, etc...)

  


  Is there a way to get each produced file into their individual and separate output streams ?

  


  ---

  


  Broader context :

  


  I'm trying to build a transcoder in Node.js running in Google Cloud, with the idea being that it writes directly to a Google Storage through a writable stream. I can only create a stream per file, but since the number of files is dynamic, I'm not sure how to obtain a stream from each file.

  


• FFMPEG bottleneck in relaying data from a dshow camera to stdout PIPE without any processing or conversion

  19 août 2020, par koonyook

  I have a USB camera (FSCAM_CU135) that can encode the video to MJPEG internally and it supports DirectShow. My goal is to retrieve the binary stream of the encoded video as it is (without decoding or preview) and send it to my program for further processing.

  


  I choose to use FFMPEG to read the MJPEG stream and pipe to stdout so that I can read it using Python's subprocess.Popen .

  


  ffmpeg -y -f dshow -vsync 2 -rtbufsize 1000M -video_size 1920x1440 -vcodec mjpeg -i video="FSCAM_CU135" -vcodec copy -f mjpeg pipe:1


  


  At this resolution, the camera is able to capture and transmit at 60 fps.
In this case, I expect FFMPEG to pass the data as fast as possible with no calculation.
With the output of FFMPEG I can tell how fast it moves the data from rtbuffer to the output pipe.

  


  With just one camera, FFMPEG works with no problem and move the data at 60 fps.
However, when I run 2 cameras simultaneously, the cameras still generate data at 60 fps but FFMPEG can only move the data around 55 fps. This means that I am unable to consume the video in realtime and the buffer consumption will be larger over time.

  


  I guess that FFMPEG didn't just simply move the data but did some processing such as searching for the beginning, the end, and the timestamp of each video frame so that it can count frames and report.
Is there a way to force FFMPEG to not doing those things and focus on passing the data only to make it faster ?

  


  If I purely use directshow API without FFMPEG, can it be faster ?