Recherche avancée

Sur d’autres sites (7914)

• Piping ffmpeg output into ffplay stdin with boost

  21 décembre 2020, par botiapa

  I'm trying to pipe the output of an ffmpeg process into an ffplay process (Sort of like a playback). My problem is the following : If I copy the output character by character (by character I mean char) it works correctly, other than it consuming a whole lot of cpu power. However when I try to pipe chunks into it (by using a buffer), ffplay for some reason doesn't even recognize the input.

  


  bp::ipstream iso;
bp::ipstream ise;
bp::opstream in;
    
bp::child ffmpeg(bp::search_path("ffmpeg"), bp::args({"-loglevel", "quiet", "-f", "pulse", "-i", "default", "-f", "wav", "-bitexact", "-nostdin", "-"}), bp::std_out > iso, bp::std_err > ise);
bp::child ffplay(bp::search_path("ffplay"), bp::args({"-loglevel", "verbose", "-nodisp", "-f", "wav", "-i", "-"}), bp::std_in < in, bp::std_out > bp::null);


  


  Here are the 2 code snippets for comparison :

  


  Here it is copying char by char

  


  while(ffmpeg.running()) {
    char c;
    c = iso.get();
    in << c;
}


  


  And here it is copying with the help of a buffer

  


  char buffer[1024];
while(ffmpeg.running()) {
    iso.get(buffer, 1024);
    in << buffer;
}


  


  I can provide ffplay output if necessary, however I didn't see any errors or things like that.

  


• Generating thumbnails from multiple videos on Desktop, using ffmpeg, or something similar

  6 février 2013, par Birk

  Hi guys this is a long shot but here goes...

  I basically have what I mentioned in the title running on my server. When I upload a video ffmpeg decomplies it and gives me screenshots, then I pick a screenshot that I want to use for that video. Currently, my server can process 3 videos at a time. The down side is that this uses up A LOT of the server processing power. :(

  Is there a way, or a program, that can process several video at a time and generate me screenshots on my Desktop ? If this is possible then I can just use my spare computer here to process everything then upload the screenshots/video to my server.

  This is what I basically have running now on the server. kayweb.com.au/blogs/Web-Development/Generating-screenshots-using-FFmpeg

  Something like this, But this thumbnail generator puts everything into one image. I need to be able to choose with thumbnail I want to use.
  http://www.tothepc.com/archives/make-movie-caps-screenshots-with-free-video-thumbnails-maker/

  Anyone have any suggestions ?

• Moviepy/FFMPEG error while writing file : Permission denied

  26 février 2024, par rodinaxcvi

  I am using Power Automate to try to run this powershell script "python3 %Directory%%FileName%" :

  


  What my code is supposed to do :

  


  

  1. Download a video from youtube

  


  2. Put small clips from a folder in bottom left

  


  3. Overwrite audio from downloaded Youtube video with new audio from a folder in the same directory (I am replacing the video's audio with an .mp3 of someone narrating something)

  


  


  ERROR CODE :

  


  OSError: \[Errno 32\] Broken pipe

MoviePy error: FFMPEG encountered the following error while writing file combined_video_1TEMP_MPY_wvf_snd.mp3:

b'combined_video_1TEMP_MPY_wvf_snd.mp3: Permission denied\\r\\n'


  


  Code Snippet that seems to be throwing up error (NOT FULL CODE) :

  


  # Combine ad clips with main video footage
def combine_videos(main_video_path, ad_video_paths, output_path, bottom_videos_folder, custom_audio_path):
    main_clip = VideoFileClip(main_video_path).subclip(0, 60)  # Keep the original aspect ratio

    main_clip = main_clip.resize(width=1920)  # Resize based on width
    if main_clip.size[1] > 1080:
        main_clip = main_clip.resize(height=1080)  # Resize based on height if it's still too tall

    main_clip_x = (1920 - main_clip.size[0]) / 2
    main_clip_y = (1080 - main_clip.size[1]) / 2
    main_clip = main_clip.set_position((main_clip_x, main_clip_y))

    ad_clip = concatenate_to_duration(ad_video_paths, 60, bottom_videos_folder)  # Assuming 60 seconds for the ad clip
    ad_clip = ad_clip.resize(width=250)  # Resize the ad clip
    ad_clip = ad_clip.set_position(("left", "bottom"))  # Position the ad clip at the bottom right

    custom_audio = AudioFileClip(custom_audio_path)

    # Specify the output directory for the audio file
    output_audio_directory = "C:/Temp/AIContentCreator"
    output_audio_path = os.path.join(output_audio_directory, "custom_audio.mp3")

    # Write the custom audio to the specified directory
    custom_audio.write_audiofile(output_audio_path)

    main_clip = main_clip.set_audio(custom_audio)

    final_clip = CompositeVideoClip([main_clip, ad_clip], size=(1920, 1080))
    final_clip.write_videofile(output_path, fps=30)

    main_clip.close()
    ad_clip.close()
    custom_audio.close()

## SAVE PATHS ##
# Video Folder Paths
save_path = r"C:\Temp\Videos\YoutubeVideo"
bottom_videos_folder = r"C:\Temp\Videos\BottomVideo"
custom_audio_path = r"C:\Temp\AIContentCreator\GPTContentCreation\output.mp3"  # Existing audio file path

# Download videos from a playlist
playlist_url1 = ''
videos_playlist1 = download_playlist(playlist_url1, save_path)

# Prepare bottom videos
bottom_videos = list_videos_in_folder(bottom_videos_folder)

# Combine each pair of videos and export 
combined_video_counter = 1  
for main_video in videos_playlist1:
    shuffled_ads = random.sample(bottom_videos, len(bottom_videos))
    ad_video_paths = [os.path.join(bottom_videos_folder, ad) for ad in shuffled_ads[:5]]  # Adjust number as needed

    combined_video_path = r'C:\Temp\Videos\TiktokVideos\combined_video_{combined_video_counter}.mp4'
    
    # Call the combine_videos function with the existing custom_audio_path
    combine_videos(main_video, ad_video_paths, combined_video_path, bottom_videos_folder, custom_audio_path)
    
    combined_video_counter += 1
`


  


  What I've tried :

  


  

  • Setting all related folders to full access

  


  • Setting full directory paths in my code

  


  • Running Power Automate and CMD as administrator

  


  • Moving files to a different folder like /Temp to see if permission error still consists

  


  


  Reoccurring problems :

  


  


  • I am not sure where MoviePy is writing "b'combined_video_1TEMP_MPY_wvf_snd.mp3". Could it be trying to place that file in another directory with out proper permissions ?

    


  


  • Whenever I try to de-select read-only on TiktokVideos folder and refresh it, "Read-only" is still on. Keeps doing this.