Recherche avancée

Sur d’autres sites (4450)

• Save FFMpeg conversion to PHP variable vs. File System for use with Whisper API ?

  13 avril 2023, par SScotti

  I just started working on a little demo to transalte audio captured from the front-end as audio/webm using JS and then sent the back-end in a Laravel App. I guess there are JS libraries that can handle the conversion, but I'd rather use a server side solution with FFMPEG, which I am doing.

  


  The backend code is below. It seems to be working after playing around with the PHP composer package that I'm using vs. one for Laravel that is also there. I'd rather use this one because I have other PHP apps that are not Laravel.

  


  Questions :

  


  


  1. With the FFMpeg library, is there a way to capture the converted .mp3 file to a PHP variable in the script rather than saving it to the file system and then reading it back in later ?

     


  


  2. For the OpenAI call, I'd like to catch exceptions there also. I just sort of have a placeholder there for now.

     


     protected function whisper(Request $request) {

    $yourApiKey = getenv('OPENAI_API_KEY');
    $client = OpenAI::client($yourApiKey);

    $file = $request->file('file');
    $mimeType = $request->file('file')->getMimeType();
    $audioContents = $file->getContent();

    try {

        FFMpeg::open($file)
        ->export()
        ->toDisk('public')
        ->inFormat(new \FFMpeg\Format\Audio\Mp3)
        ->save('song_converted.mp3');
    }
    catch (EncodingException $exception) {
        $command = $exception->getCommand();
        $errorLog = $exception->getErrorOutput();
    }

    $mp3 = Storage::disk('public')->path('song_converted.mp3');
    try {
    $response = $client->audio()->transcribe([
    'model' => 'whisper-1',
    'file' =>  fopen($mp3, 'r'),
    'response_format' => 'verbose_json',
    ]);
    }
    catch (EncodingException $exception) {
        $command = $exception->getCommand();
        $errorLog = $exception->getErrorOutput();
    }

 echo json_encode($response);

}


     


  


  


• How to send RTP stream to Janus from NGINX RTMP module ? [closed]

  25 novembre 2024, par Matéo

  I'm trying to create a stream and display it in a browser. I have already configured NGINX with the rtmp module and my stream works very well with HLS (between 5 and 10 seconds of latency).

  


  Now I would like to set up a low-latency stream and that's why I have installed the janus-gateway webRTC server that allows to take in input an RTP stream and provide in output a webRTC stream.

  


  Here's the schema I'd like to follow :

  


  OBS -> RTMP -> Nginx-rtmp-module -> ffmpeg -> RTP -> Janus -> webRTC -> Browser

  


  But I have a problem with this part : "nginx-rtmp-module -> ffmpeg -> janus"

  


  In fact, my janus's server is running and demos streaming works very well in localhost, but when i try to provide an RTP stream, Janus don't detect the stream in the demos (it shows "No remote video available").

  


  Anyone can help me, please ?

  


  Ressources :

  


  

  • My janus.plugin.streaming.jcfg configuration :

  


  


  rtp-sample: {
        type = "rtp"
        id = 1
        description = "Opus/VP8 live stream coming from external source"
        metadata = "You can use this metadata section to put any info you want!"
        audio = true
        video = true
        audioport = 5002
        audiopt = 111
        audiortpmap = "opus/48000/2"
        videoport = 5004
        videopt = 100
        videortpmap = "VP8/90000"
        secret = "adminpwd"
}



  


  

  • My nginx.conf application :

  


  


  application test {

        deny play all;

        live on;
        on_publish http://localhost/test/backend/sec/live_auth.php;

        exec ffmpeg -i rtmp://localhost/test/$name -an -c:v copy -flags global_header -bsf dump_extra -f rtp rtp://localhost:5004;

}


  


• libavfilter/dnn_bakcend_openvino : Add automatic input/output detection

  17 janvier 2024, par Wenbin Chen

  libavfilter/dnn_bakcend_openvino : Add automatic input/output detection
  Now when using openvino backend, user doesn't need to set input/output
  
  names in command line. Model ports will be automatically detected.
  For example :
  
  ffmpeg -i input.png -vf \
  
  dnn_detect=dnn_backend=openvino:model=model.xml:input=image :\
  
  output=detection_out -y output.png
  can be simplified to :
  
  ffmpeg -i input.png -vf dnn_detect=dnn_backend=openvino:model=model.xml\
  
   -y output.png
  Signed-off-by : Wenbin Chen <wenbin.chen@intel.com>
  
  Reviewed-by : Guo Yejun <yejun.guo@intel.com>
  

  • [DH] libavfilter/dnn/dnn_backend_openvino.c
  • [DH] libavfilter/dnn_filter_common.c