Recherche avancée

Sur d’autres sites (3459)

• Bluemix node js build pack to support webm audio conversion

  31 août 2018, par Nimmy Mohandas

  Even after adding ffmpeg to bluemix node js build pack(I tried this https://github.com/BlueChasm/nodejs-buildpack-ffmpeg), it doesn’t support webm audio format conversion.Could anyone please suggest alternate ways to support this issue ?

  Description : I want to do speech to text conversion using Google speech recognition. I am using react native for front-end development. I recorded the voice and generated a webm audio file. Since the google speech recognition doesn’t support webm file format, I converted it into wav format using fluent-ffmpeg. It is working perfectly in my local system, but when I deploy it to the IBM cloud foundry using ffmpeg buildpack(https://github.com/BlueChasm/nodejs-buildpack-ffmpeg), it generates a file without extension. Following are my code,

  router.post('/st', audioUpload.single('audio'), function(req, res, next) {
  
          if (!req.file) {
  
              return res.json({ status: false, error: 'No input given!' });
  
          }
  
  
  
          /**
  
           * convert to .wav
  
           */
  
  
  
  
  
          ffmpeg(req.file.path)
  
              .toFormat('wav')
  
              .inputOptions(['-r 32000',
  
                  '-ac 1'
  
              ])
  
  
  
              .on('error', (err) => {
  
                  console.log('An error occurred: ' + err.message);
  
              })
  
              .save(path.join(__dirname, '..', 'uploads/audio/file.wav'))
  
  
  
          .on('end', () => {
  
  
  
              st.detectAudioFile(path.join(__dirname, '..', 'uploads/audio/file.wav')).then(data => {
  
                      fs.unlink(path.join(__dirname, '..', 'uploads/audio/file.wav'), function(err) {
  
                          if (err) return res.send(err);
  
                          const response = data[0];
  
                          const transcription = response.results
  
                              .map(result => result.alternatives[0].transcript)
  
                              .join('\n');
  
                          return res.json({ transcription });
  
                      }), fs.unlink(req.file.path, function(err) {
  
                          if (err) return res.json(err);
  
                      })
  
  
  
                  })
  
                  .catch(err => {
  
                      res.send({
  
  
  
                          status: false,
  
                          error_code: 400,
  
                          err: err.error || err.message
  
                      });
  
                  });
  
          })
  
      })

  //speech-to-text.js

  const fs = require('fs');
  
  const speech = require('@google-cloud/speech');
  
  
  
  
  
  var detectAudioFile = function(fileName) {
  
  
  
      const Speechclient = new speech.SpeechClient({})
  
  
  
      // Reads a local audio file and converts it to base64
  
      const file = fs.readFileSync(fileName);
  
  
  
      const audioBytes = file.toString('base64');
  
      const audio = {
  
          content: audioBytes,
  
      };
  
      const config = {
  
          // encoding: "FLAC",
  
          // sampleRateHertz: 44100,
  
  
  
          languageCode: 'en-US',
  
      };
  
      const request = {
  
          audio: audio,
  
          config: config,
  
      };
  
      return Speechclient
  
          .recognize(request);
  
  
  
  }
  
  
  
  
  
  
  
  module.exports = {
  
      detectAudioFile
  
  }

  `

• bash variable changes in loop with ffmpeg

  17 septembre 2018, par Mike

  I wrote a skript to quickly create short preview clips from vides I recorded on timestamps that I found worth checking out later for cutting.
  My file with the timestamps is written like this

  FILE_NAME1#MM:SS MM:SS
  
  FILE_NAME2#MM:SS MM:SS MM:SS MM:SS

  example :

  MAH01728#02:47 03:34 03:44 05:00 06:08 06:55

  The script looks like this :

  #!/bin/bash
  
  while read f
  
   do
  
  
  
   file=$(echo $f | cut -d"#" -f1)
  
   filename=${file}".MP4"
  
   timestamps=$(echo $f | cut -d"#" -f2)
  
  
  
  for time in $timestamps 
  
   do 
  
    ffmpeg -ss 00:${time}.0 -i "orig/${filename}" -c copy -t 10 "preview/${file}_${time}.MP4"
  
   done
  
  done < $1

  The script gets half of the previews that I want and on the other the filename is messed up and ffmpeg complains that the file is not found :

  orig/714.MP4: No such file or directory
  
  orig/00:58 01:25.MP4: No such file or directory

  So I modified the script for trouble shooting and just put an echo in front of the ffmpeg command - now all file names are correct. What am I missing ?

  ffmpeg -ss 00:01:47.0 -i orig/MAH01714.MP4 -c copy -t 10 preview/MAH01714_01:47.MP4
  
  ffmpeg -ss 00:02:00.0 -i orig/MAH01713.MP4 -c copy -t 10 preview/MAH01713_02:00.MP4
  
  ffmpeg -ss 00:00:58.0 -i orig/MAH01712.MP4 -c copy -t 10 preview/MAH01712_00:58.MP4
  
  ffmpeg -ss 00:01:25.0 -i orig/MAH01712.MP4 -c copy -t 10 preview/MAH01712_01:25.MP4

• How to get separated audio channel from 5.1 video with ffmpeg ?

  23 septembre 2018, par mr_blond

  I have a video file with audio stream :

   Stream #0:1(eng): Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  
   Stream #0:2(eng): Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s

  And I can do mixdown it to two-channel audio :

  ffmpeg -i vid.mkv -map 0:2 -acodec libmp3lame -y output.mp3

  Update

  It can be splitted by the channelsplit command, but it splits the Stream #0:1 by default and I can’t specify Stream #0:2 channel.

  ./ffmpeg -i vid.mkv -vn -filter_complex 'channelsplit=channel_layout=5.1[FC]' -map '[FC]' front_center.mp3

  How do I specify Stream #0:2 5.1 channel and get separated front center channel from this video ?