Recherche avancée

Sur d’autres sites (8803)

• Python buffered IO ending early streaming with multiple pipes

  5 octobre 2022, par Malibu

  I'm trying to make a continuous livestream of videos downloaded via yt-dlp. I need to port this (working) bash command into Python.

  


  (
    youtube-dl -v --buffer-size 16k https://youtube.com/watch?v=QiInzFHIDp4 -o - | ffmpeg -i - -f mpegts -c copy - ;
    youtube-dl -v --buffer-size 16k https://youtube.com/watch?v=QiInzFHIDp4 -o - | ffmpeg -i - -f mpegts -c copy - ;
) | ffmpeg -re -i - -c:v libx264 -f flv rtmp://127.0.0.1/live/H1P_x5WPF


  


  My Python attempt is cutting off the last 2 seconds of each video. My suspicion is that although the first pipe, yt-dlp, has an empty stdout, there is still data travelling between the second and third pipe. I haven't been able to figure out a way to properly handle the data between those two pipes at the end of the video.

  


  from subprocess import Popen, PIPE, DEVNULL

COPY_BUFSIZE = 65424

playlist = [
    {
        # 15 second video
        "url": "https://youtube.com/watch?v=QiInzFHIDp4"
    },
    {
        # 15 second video
        "url": "https://youtube.com/watch?v=QiInzFHIDp4"
    },
    {
        # 15 second video
        "url": "https://youtube.com/watch?v=QiInzFHIDp4"
    },
]

if __name__ == "__main__":
    stream_cmd = [
        "ffmpeg", "-loglevel", "error",
        "-hide_banner", "-re", "-i", "-",
        "-c:v", "libx264",
        "-f", "flv",
        "-b:v", "3000k", "-minrate", "3000k",
        "-maxrate", "3000k", "-bufsize", "3000k",
        "-r", "25", "-pix_fmt", "yuv420p",
        "rtmp://127.0.0.1/live/H1P_x5WPF"
    ]
    print(f'Stream command:\n"{" ".join(stream_cmd)}"')

    encoder_cmd = [
        "ffmpeg", "-re", "-i", "-", "-f", "mpegts",
        "-c", "copy", "-"
    ]
    print(f'Encoder command:\n"{" ".join(encoder_cmd)}"')

    stream_p = Popen(stream_cmd, stdin=PIPE, stderr=DEVNULL)

    for video in playlist:
        yt_dlp_cmd = [
            "yt-dlp", "-q",
            video["url"],
            "-o", "-"
        ]

        print("Now playing: " + video["url"])

        with Popen(yt_dlp_cmd, stdout=PIPE) as yt_dlp_p:
            with Popen(encoder_cmd, stdin=PIPE, stdout=PIPE, stderr=DEVNULL) as encoder_p:
                while True:
                    yt_dlp_buf = yt_dlp_p.stdout.read(COPY_BUFSIZE)
                    print("READ: yt_dlp")
                    if not yt_dlp_buf:
                        print("yt-dlp buffer empty")
                        # Handle any data in 2nd/3rd pipes before breaking?
                        break

                    written = encoder_p.stdin.write(yt_dlp_buf)
                    print("WRITE: encoder. Bytes: " + str(written))

                    encoder_buf = encoder_p.stdout.read(COPY_BUFSIZE)
                    # if not encoder_buf:
                    #     print("encoder_buf empty")
                    #     break
                    print("READ: encoder")

                    stream_bytes_written = stream_p.stdin.write(encoder_buf)
                    print("WRITE: stream, Bytes: " + str(stream_bytes_written))


  


  Running Python 3.6.9 on MacOS.

  


• FFMPEG build error when building app on OS X

  26 juillet 2014, par konsti

  I installed FFMPEG on OS X 10.9 with homebrew (and earlier with macport) and the ffmpeg command in console is working. But when trying to include it to my XCode project and building my Mac app I get build errors :

  #include 
  
  #include 

  stdarg.h

  User defined issues

  This header only supports __MWERKS__.

  When trying to add it as preprocessor macro I get more errors stating that mw_stdarg.h file not found. Did I miss some configuration somewhere ? Please help :)

• ffmpeg : convert a video by using another video's settings

  11 janvier 2015, par orcaman

  I have an mp4 file containing an H.264 video. This video has particular settings that I am not sure about, but I want to reproduce them on other videos that I encode (those other videos are already H.264 encoded, but they have different properties - bitrate, size, etc.)

  I know that ffprobe can be used to extract some information from the original video, and I guess that using this information I could try to reproduce the conversion settings required for use with other videos.

  Question is : do I have to do this manually, in the sense that I need to map the output of ffprobe to ffmpeg flags ? Is there a better way to do this, to make sure I don’t miss anything ?