Recherche avancée

Sur d’autres sites (4162)

• Downloading youtube mp3 - metadata encoding issue (python, youtube-dl, ffmpeg)

  21 mai 2015, par mopsiok

  I’m trying to download audio from youtube with youtube-dl.exe and ffmpeg.exe (Windows 7), but I am having some troubles with encoding. I have to parse metadata manually, because when I try to use

  --metadata-from-title "%(artist) - %(title)" --extract-audio --audio-format mp3 https://www.youtube.com/watch?v=DaU94Ld3fuM

  I get ERROR : Could not interpret title of video as "%(artist) - %(title)"

  Anyway, I wrote some code to save metadata with ffmpeg :

  def download(url, title_first=False):
  
      if (0 == subprocess.call('youtube-dl --extract-audio --audio-format mp3 %s' % url)):
  
          #saves file in current directory in format: VID_TITLE-VID_ID.mp3
  
          video_id = url[url.find('=')+1:] #video id from URL (after ?v=)
  
          for f in os.listdir('.'):
  
              if video_id in f:
  
                  filename = f
  
                  break
  
          os.rename(filename, video_id+'.mp3') #name without non-ascii chars (for tests)
  
          video_title = filename[: filename.find(video_id)-1]
  
  
  
          output = video_title + '.mp3'
  
          title, artist = '', ''
  
          try: #parsing the title
  
              x = video_title.find('-')
  
              artist = video_title[:x].strip()
  
              title = video_title[x+1:].strip()
  
              if (title_first): output = '%s - %s.mp3' % (title, artist)
  
          except:
  
              pass
  
  
  
          x = 'ffmpeg -i "%s" -metadata title="%s" -metadata artist="%s" -acodec copy -id3v2_version 3 -write_id3v1 1 "%s"' \
  
                          % (video_id+'.mp3', title, artist, output)
  
          print x
  
          subprocess.call(x)

  The file is downloaded and then cropped to given start and duration times (the code above is a simplified version). Filename is fine, but when I open the file with AIMP3, it shows rubbish instead of non-ascii characters :

  

  I’ve tried to re-encode the final command with iso-8859-2, utf-8 and mbcs :

  x = x.decode('cp1250').encode('iso-8859-2')

  But non-ascii chars are still not readable. Passing an unicode command returns UnicodeEncodeError...

  Any idea how to solve this problem ?

• Why i'm failing to download mp4 file using youtube-dl in python

  9 décembre 2019, par Joao

  Python noob :(

  I created a script that downloads any youtube video.

  Then the script inserts a subtitle into the video and that works fine.

  However, I have a problem, I can’t make the script always download the video in mp4. Most of the time it is downloaded in mkv and i can’t understand why that happens and how can I change the code in order to always download the video in the format I prefer.

  The function in charge to download the video is this one :

  def video_download():
  
      #global resolution
  
      try:
  
          print(linkvideo)
  
          ydl_opts = {
  
              'format': 'bestvideo[ext=mp4]+bestaudio/best',
  
              'outtmpl': video_id+"."+'%(ext)s'
  
              #'bestvideo': 'mp4',
  
              #'bestaudio': 'm4a',
  
              #'ext': 'mp4'
  
          }
  
          with youtube_dl.YoutubeDL(ydl_opts) as ydl:
  
              ydl.download([linkvideo])
  
              result = ydl.extract_info("{}".format(linkvideo))
  
              title = result.get("id", None)
  
              videoext = result.get("ext", None)
  
              resolution = result.get("resolution", None)
  
      except:
  
          print("falhou")
  
          pass
  
      global ficheiro
  
      ficheiro2 = str(title)+"."+"mkv"
  
      ficheiro =""
  
      for i in ficheiro2:
  
          if (i == '"' or i =="'" or i =="/"):
  
              pass
  
          else:
  
              ficheiro += i

  Can any one help me ? Thanks in advance.

  I got it, thanks to everyone who tried to help me.

  The problem is this : when we try to download any video with the opts i’ve selected, youtube-dl will download the best video/audio combination in the same file, so I’ve added the following in my code :

  ytdl = YoutubeDL()
  
      result = ytdl.extract_info(linkvideo, download=False)
  
      formats = result['formats']
  
      formato = ""
  
      formatoaudio =  ""
  
      extaudio = ""
  
      for format in formats:
  
          if format['format_note'] == "1080p" and format['ext'] == "mp4":
  
              print(format)
  
              formato = (format['format_id'])
  
              break
  
          else:
  
              print("no 1080p mp4?!")
  
  
  
      if formato == "":
  
          for format in formats:
  
              if format['format_note'] == "720p" and format['ext'] == "mp4":
  
                  print(format)
  
                  formato = (format['format_id'])
  
                  break
  
              else:
  
                  print("no 720p mp4")

  Then I did the same to the sound file, checking if there is any m4a or webm sound file. To finish I’m joining both files using FFmpeg.

  Maybe this is not the best solution, but I’m still learning. :)

  Thank you all

• How to apply dynamic watermarking for users watching video in real-time ? [closed]

  3 janvier, par Barun Bhattacharjee

  I am working on a video streaming project where I need to apply a dynamic watermarking (e.g., username and email) in real-time for security purposes. The video is being streamed in DASH format, and the segment files are in .m4s format generated via FFmpeg.

  


  Challenges :
Is it possible to directly apply dynamic watermarking to .m4s segment files ?

  


  Video segments are generated using FFmpeg with the following command :

  


  ffmpeg
    .input(video_path)
    .output(mpd_path,
            format='dash',
            map='0',
            video_bitrate='2400k',
            video_size='1920x1080',
            vcodec='libx264',
            seg_duration='4',  # Sets segment duration to 4 seconds
            acodec='copy')
    .run()



  


  What I tried :
I attempted to use FFmpeg to apply a watermark dynamically to the .m4s files using the drawtext filter, but .m4s files are not always recognized as valid input for FFmpeg operations.

  


  # FFmpeg command to add watermark to m4s file
try:
    # FFmpeg processing
    out, err = (
        ffmpeg
        .input(m4s_file_path)  # Input the segment file
        .filter(
            "drawtext",
            text=user_info,
            fontfile="font/dejavu-sans/DejaVuSans-Bold.ttf",
            fontsize=24,
            fontcolor="white",
            x=10,
            y=10
        )
        .output(
            "pipe:",  # Stream output as a byte stream
            format="mp4",  # Output format as MP4 (compatible with MPEG-DASH)
            vcodec="libx264",
            acodec="copy",
            movflags="frag_keyframe+empty_moov"
        )
        .run(capture_stdout=True, capture_stderr=True)
    )

    logger.info(f"FFmpeg process completed. stdout length: {len(out)}, stderr: {err.decode('utf-8')}")
    logger.error(f"FFmpeg stderr: {err.decode('utf-8')}")
    return out  # Return the processed video stream data


except ffmpeg.Error as e:
    stderr_output = e.stderr.decode('utf-8') if e.stderr else "No stderr available"
    logger.error(f"FFmpeg error: {stderr_output}")

    raise RuntimeError(f"Error processing video: {stderr_output}")



  


  Error I faced :

  


  video-streaming-backend  | [mov,mp4,m4a,3gp,3g2,mj2 @ 0x7f1bf99cc640] Could not find codec parameters for stream 0 (Video: h264 (avc1 / 0x31637661), none(tv, bt709), 1920x1012): unspecified pixel format
video-streaming-backend  | Consider increasing the value for the 'analyzeduration' (10000000) and 'probesize' (5000000) options
video-streaming-backend  | Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'http://web:8000/media/stream_video/chunks/ec1db006-b488-47ad-8220-79a05bcaae39/segments/init-stream0.m4s':
video-streaming-backend  |   Metadata:
video-streaming-backend  |     major_brand     : iso5
video-streaming-backend  |     minor_version   : 512
video-streaming-backend  |     compatible_brands: iso5iso6mp41
video-streaming-backend  |     encoder         : Lavf60.16.100
video-streaming-backend  |   Duration: N/A, bitrate: N/A
video-streaming-backend  |   Stream #0:0[0x1](und): Video: h264 (avc1 / 0x31637661), none(tv, bt709), 1920x1012, SAR 1:1 DAR 480:253, 12288 tbr, 12288 tbn (default)
video-streaming-backend  |     Metadata:
video-streaming-backend  |       handler_name    : VideoHandler
video-streaming-backend  |       vendor_id       : [0][0][0][0]
video-streaming-backend  | Stream mapping:
video-streaming-backend  |   Stream #0:0 (h264) -> drawtext:default
video-streaming-backend  |   drawtext:default -> Stream #0:0 (libx264)
video-streaming-backend  | Press [q] to stop, [?] for help
video-streaming-backend  | Cannot determine format of input stream 0:0 after EOF
video-streaming-backend  | Error marking filters as finished
video-streaming-backend  | Error while filtering: Invalid data found when processing input
video-streaming-backend  | [out#0/mp4 @ 0x7f1bf8e73100] Nothing was written into output file, because at least one of its streams received no packets.
video-streaming-backend  | frame=    0 fps=0.0 q=0.0 Lsize=       0kB time=N/A bitrate=N/A speed=N/A    
video-streaming-backend  | Conversion failed!



  


  These errors have left me wondering if .m4s is a viable format for dynamic watermarking. If it's not, what would be the correct approach ?