Recherche avancée

Sur d’autres sites (12323)

• How to use Mozilla DeepSpeech to generate subtitle from video file ?

  31 mars 2022, par SingularitySG

  I'm currently trying to generate subtitles file .srt using Mozilla DeepSpeech libraries.

  


  Confusion I'm having :

  


  I'm following this guide but I'm completely lost at the part where he uses pyAudioAnalysis to trim out the silence portion of the .wav files.

  


  Another is the portion where he processes the audio file via calling the ds model & ds scorer
I've downloaded both deepspeech-0.9.3-models.scorer and deepspeech-0.9.3-models.pbmm from https://github.com/mozilla/DeepSpeech/releases Do I just reference the model and scorer to the downloaded file path instead ?

  


  def ds_process_audio(audio_file, file_handle): 
    ds = Model(ds_model)
    ds.enableExternalScorer(ds_scorer)


  


  At the moment, I'm able to extract out .wav files from .mp4, .mkv video format.

  


  I've tried pip install pyAudioAnalysis but I'm not sure how to call the functions that are related to it as per his guide in trimming out the silence portion of the .wav file. Below is the code I'm currently working with as of now.

  


  video_name = "Videos\Gintama_EP342.mkv"
audio_name = video_name + ".wav"


def extractAudio(input_file, audio_file_name):
    # Extract audio from input video file and save to audio/ in root dir
    # Args:
    #    input_file : input video file
    #    audio_file_name : save audio WAV file with same filename as video file

    command = ["ffmpeg", "-hide_banner", "-loglevel", "warning", "-i", input_file, "-ac", "1", "-ar", "16000", "-vn", "-f", "wav", audio_file_name]    
    try:
        ret = sp.call(command, shell=True)
        print("Extracted audio to audio/{}".format(audio_file_name.split("/")[-1]))
    except Exception as e:
        print("Error: ", str(e))
        exit(1)


  


  Appreciate any help given. Thank you

  


• FFMPEG “buffer queue overflow, dropping.” while filter_complex concating

  8 août 2019, par March3April4

  I’m using the below command to concat several videos to a single video.

  ffmpeg -i /storage/emulated/0/Cache/1565235095822.mp4
  
  -i /storage/emulated/0/Cache/1565235097989.mp4
  
  -i /storage/emulated/0/Cache/1565235099167.mp4 
  
  -i /storage/emulated/0/Cache/1565235100276.mp4
  
  -i /storage/emulated/0/Cache/1565235096982.mp4
  
  -filter_complex [0:v:0] [0:a:0] [1:v:0] [1:a:0] [2:v:0] [2:a:0] [3:v:0] [3:a:0] [4:v:0] [4:a:0] concat=n=5:v=1:a=1 [v] [a]
  
  -map [v]
  
  -map [a]
  
  -max_muxing_queue_size 9999
  
  -preset ultrafast
  
  /storage/emulated/0/Cache/1565235101418.mp4

  The input videos are also made by ffmpeg, and are about 1 megabyte big. (Each consists of a single picture and an audio)

  The above command line works well for mainstream devices such as galaxy s10 series, but for models like Lg K10, the below line appears.

  [Parsed_concat_0 @ 0xb9ab05a0] Buffer queue overflow, dropping.

  The encoding does not stop even though the above line shows up, but the result video contains some time skips and missing audio.

  I assumed that this has to do with performance of the device itself, so, I tried to do the job without encoding. I’ve tried the concat demuxer with no luck.

  What else can I try to make this to work ? Any help will be thankful.

  ------------------Edit-------------------------

  The ffmpeg videos are create by the following command.

          String[] array = {
  
                   "-framerate", "1",
  
                  "-i", "targetVideo.png",
  
                  "-i", "targetAudio.mp3",
  
                  "-r", "14",
  
                  "-c:v", "libx264",
  
                  "-c:a", "aac",
  
                  "-pix_fmt", "yuv420p",
  
                  "-video_track_timescale", "90000",
  
                  "-c:a", "copy",
  
                  outputPath.mp4};

• avformat/matroskaenc : Check for failure when writing SeekHead

  29 décembre 2019, par Andreas Rheinhardt

  avformat/matroskaenc : Check for failure when writing SeekHead
  mkv_write_seekhead() would up until now try to seek to the position where
  
  the SeekHead ought to be written, write the SeekHead and seek back. The
  
  first of these seeks was checked as was writing, yet the seek back was
  
  unchecked. Moreover the return value of mkv_write_seekhead() was unchecked
  
  (the ordinary return value was the position where the SeekHead was written).
  This commit changes this : Everything is checked. In the unseekable case
  
  (where the first seek may nevertheless work when it happens in the buffer)
  
  a failure at the first seek is not considered an error. In any case,
  
  failure to seek back is an error.
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
  

  • [DH] libavformat/matroskaenc.c