Recherche avancée

Sur d’autres sites (6413)

• FFMPeg generated video : Audio has 'glitches' when uploaded to YouTube

  7 octobre 2023, par CularBytes

  I've generated a voice from Azure AI Speech at 48KHz and 96K Bit Rate, generated a video of some stock footages and I'm trying to combine all of that with a background music.
The voice-over is generated per setence, so that I know how long each setence is and to include relevant video footage.

  


  I'm using FFMpeg through the FFMpegCore nuget package.

  


  The problem

  


  After the video is complete with background music, I play it on my computer and it's perfect (no audio glitches, music keeps playing). But when uploaded to youtube it has 'breaks' in the music inbetween sentences (basically everytime a new voice-fragment is starting).

  


  Example : https://www.youtube.com/watch?v=ieNvQ2TNq44

  


  The code

  


  All of the footage is combined with mostly FFMpeg.Join(string output, string[] videos). These video files also contain the voice-overs (per sentance).

  


  After that I try to add the music like this :

  


     string outputTimelineWithMusicPath = _workingDir + $@"\{videoTitle}_withmusic.mp4";
    FFMpegArguments
        .FromFileInput(inputVideoPath)
        .AddFileInput(musicPath)
        .OutputToFile(outputPath, true, options => options
            .CopyChannel()
            .WithAudioCodec(AudioCodec.Aac)
            .WithAudioBitrate(AudioQuality.Good)
            .UsingShortest(true)
            .WithCustomArgument("-filter_complex \"[0:a]aformat=fltp:44100:stereo,apad[0a];[1]aformat=fltp:44100:stereo,volume=0.05[1a];[0a][1a]amerge[a]\" -map 0:v -map \"[a]\" -ac 2"))
        .ProcessSynchronously();


  


  I've tried to mess around with the CustomArgument, but so far no success.

  


  For example, I thought removing apad from the argument so no 'blank spots' are added, should perhaps fix the issue. Also tried to use amix instead of amerge.

  


  Last try

  


  I've tried to first make sure both files had the same sample rate, in the hope to fix the issue. So far, no success

  


      string outputVideoVoicePath = _workingDir + $@"\{title}_voiceonly_formatting.mp4";
    string musicReplacePath = _workingDir + $@"\{title}_music_formatted.aac";
    FFMpegArguments
    .FromFileInput(inputVideoPath)
    .OutputToFile(outputVideoVoicePath, true, options => options
        .WithAudioCodec(AudioCodec.Aac)
        .WithAudioBitrate(128)
        .WithAudioSamplingRate(44100)
    )
    .ProcessSynchronously();
    
    FFMpegArguments
        .FromFileInput(music.FilePath)
        .OutputToFile(musicReplacePath, true, options => options
            .WithAudioCodec(AudioCodec.Aac)
            .WithAudioBitrate(256) //also tried 96 (which is original format)
            .WithAudioSamplingRate(44100)
        )
        .ProcessSynchronously();
    
    
    Console.WriteLine("Add music...");
    var videoTitle = Regex.Replace(title, "[^a-zA-Z]+", "");
    string outputTimelineWithMusicPath = _workingDir + $@"\{videoTitle}_withmusic.mp4";
    FFMpegArguments
        .FromFileInput(outputVideoVoicePath)
        .AddFileInput(musicReplacePath)
        .OutputToFile(outputTimelineWithMusicPath, true, options => options
            .CopyChannel()
            .WithAudioCodec(AudioCodec.Aac)
            .WithAudioBitrate(AudioQuality.Good)
            .UsingShortest(true)
            .WithCustomArgument("-filter_complex \"[0:a]aformat=fltp:44100:stereo[0a];[1]aformat=fltp:44100:stereo,volume=0.05[1a];[0a][1a]amix=inputs=2[a]\" -map 0:v -map \"[a]\" -ac 2"))
        .ProcessSynchronously();
    return outputTimelineWithMusicPath;


  


  I'm not much of an expert when it comes to audio/video codecs. I do scale each stock video to 24fps, 1920x1080 and the music has a original bitrate of 256Kbps / 44100 sample rate (so I probably don't even have to convert the audio file).

  


• lpc : add ff_lpc_calc_ref_coefs_f() function

  1er septembre 2015, par Rostislav Pehlivanov

  lpc : add ff_lpc_calc_ref_coefs_f() function
  This commit adds a function to get the reflection coefficients on
  
  floating point samples. It’s functionally identical to
  
  ff_lpc_calc_ref_coefs() except it works on float samples and will
  
  return the global prediction gain. The Welch window implementation
  
  which is more optimized works only on int32_t samples so a slower
  
  generic expression was used.
  Signed-off-by : Rostislav Pehlivanov <atomnuker@gmail.com>
  

  • [DH] libavcodec/lpc.c
  • [DH] libavcodec/lpc.h

• ffmpeg for video transcription [closed]

  22 juin 2023, par TSz

  Does ffmpeg have capabilities to generate a transcription from a video or an audio file in .srt format or other ? Task is to create subtitles for YT videos (30-60min long). Timestamps and duration would be important so the text is synchronized to the audio. Languages english and hungarian

  


  no code was tried yet