Recherche avancée

Sur d’autres sites (11947)

• Streaming ffmpeg from fifo file starts only when i close the fifo file

  23 juin 2022, par tamirg

  Im starting an ffmpeg process, where the input is a FIFO file i created.
Im writing some data in a loop to the FIFO file, but the ffmpeg process doesn't start streaming until one of the two happens :

  


  

  1. i'm closing the file

  


  2. iv'e written a certain amount of data. after a while of writing, the ffmpeg process starts streaming. The more data i write, the faster it starts running. (im writing a chunk of data on each loop, if i just duplicate those chunks times 100, it starts much faster).

  


  


  What can be the reason for that ? Is there a minimum of data required for the ffmpeg process to start streaming ? How can i "force" it to start, without closing the FIFO file after writing ?

  


• How can I convert WebM file to WebP file with transparency ?

  24 août 2020, par c-an

  I tried it with ffmpeg.

  


  ffmpeg input.webm output.webp


  


  input.webm contains transparent background and But the alpha channel becomes white in webp. I think that means alpha channel doesn't come together.

  


  I extracted frames with this command :

  


  ffmpeg -i input.xxx -c:v libwebp output_%03d.webp


  


  And it also gives me webp files with white background.

  


  How can I convert it properly with alpha channel ? OR should I convert it from other format(extension) ?

  


• Speech recognition with python-telegram-bot without downloading an audio file

  25 juin 2022, par linz

  I'm developing a telegram bot in which the user sends a voice message, the bot transcribes it and sends back what was said in text.
For that I am using the python-telegram-bot library and the speech_recognition library with the google engine.
My problem is, the voice messages sent by the users are .mp3, however in order to transcribe them i need to convert them to .wav. In order to do that I have to download the file sent to the bot.
Is there a way to avoid that ? I understand this is not an efficient and a safe way to do this since many active users at once will result in race conditions and takes a lot of space.

  


  
def voice_handler(update, context):
    bot = context.bot
    file = bot.getFile(update.message.voice.file_id)
    file.download('voice.mp3')
    filename = "voice.wav"
    
    # convert mp3 to wav file
    subprocess.call(['ffmpeg', '-i', 'voice.mp3',
                         'voice.wav', '-y'])

    # initialize the recognizer
    r = sr.Recognizer()
    
    # open the file
    with sr.AudioFile(filename) as source:
    
        # listen for the data (load audio to memory)
        audio_data = r.record(source)
        # recognize (convert from speech to text)
        text = r.recognize_google(audio_data, language='ar-AR')
        
        
def main() -> None:
    updater.dispatcher.add_handler(MessageHandler(Filters.voice, voice_handler))