Recherche avancée

Sur d’autres sites (6776)

• Replicating C metadata example from ffmpeg in C# using ffmpeg-autogen - av_dict_get fails

  2 octobre 2017, par Adam Baxter

  I am trying to replicate https://ffmpeg.org/doxygen/3.0/doc_2examples_2metadata_8c_source.html in C# using the wrapper from https://github.com/Ruslan-B/FFmpeg.AutoGen

  I can open and read some properties of the file just fine, however tag is always null, even after the call to av_dict_get

  My code is as follows

  using System;
  
  using System.IO;
  
  using System.Reflection;
  
  using System.Runtime.InteropServices;
  
  using System.Text;
  
  using FFmpeg.AutoGen;
  
  
  
  namespace ffmpeg_test
  
  {
  
      class Program
  
      {
  
          static void Main(string[] args)
  
          {
  
              var currentPath = Path.GetDirectoryName(Assembly.GetEntryAssembly().Location);
  
              var libPath = Path.Combine(currentPath, "lib");
  
              SetDllDirectory(libPath);
  
              ffmpeg.av_register_all();
  
              ffmpeg.avcodec_register_all();
  
  
  
              DoRiskyThings();
  
  
  
          }
  
  
  
          private static unsafe void DoRiskyThings()
  
          {
  
              var pFormatContext = ffmpeg.avformat_alloc_context();
  
  
  
              if (ffmpeg.avformat_open_input(&pFormatContext, "01 - 999,999.opus", null, null) != 0)
  
                  throw new ApplicationException(@"Could not open file.");
  
  
  
              ffmpeg.avformat_find_stream_info(pFormatContext, null);
  
  
  
              AVStream* pStream = null;
  
              pStream = pFormatContext->streams[0];
  
              var codecContext = *pStream->codec;
  
  
  
              Console.WriteLine($"codec name: {ffmpeg.avcodec_get_name(codecContext.codec_id)}");
  
              Console.WriteLine($"number of streams: {pFormatContext->nb_streams}");
  
  
  
              //attempting to replicate https://ffmpeg.org/doxygen/3.0/doc_2examples_2metadata_8c_source.html
  
              AVDictionaryEntry* tag = null;
  
              tag = ffmpeg.av_dict_get(pFormatContext->metadata, "", null, 2);
  
              while (tag != null)
  
              {
  
                  tag = ffmpeg.av_dict_get(pFormatContext->metadata, "", tag, 2);
  
                  Console.WriteLine(Encoding.UTF8.GetString(tag->key,100));
  
                  //tag->key and //tag->value are byte pointers
  
              }
  
          }
  
  
  
          [DllImport("kernel32", SetLastError = true)]
  
          private static extern bool SetDllDirectory(string lpPathName);
  
      }
  
  }

• MediaRecorder - FFMPEG - RTMP issue

  5 octobre 2017, par Bear10

  Currently I’m working on streaming Audio and Video from a video captured in Chrome 61.0.3163.100 with the MediaRecorder API.

  The backend is running on node and it spawns a child process which runs ffmpeg :

  ffmpeg -threads 4 -i udp://127.0.0.1:41234 -t 5400 -ar 44100
  
  -b:a 128k -c:v libx264 -c:a libfdk_aac -pix_fmt yuv420p -r 30 -g 60 -vb 2048k -preset slow -minrate 2000k -maxrate 4000k test.mp4

  On the client side I send the raw Blob from Chrome to the Node server via web socket :

   var mediaStream = new MediaStream([videoOutput.getVideoTracks()[0], audioOutput.getAudioTracks()[0]]),
  
              recorder = new MediaRecorder(mediaStream, {
  
                  mimeType: 'video/webm;codecs=vp8'
  
              });
  
  
  
          socket = io('http://localhost:' + port);
  
  
  
          recorder.addEventListener('dataavailable', function (evt) {
  
              socket.emit('video:data', evt.data);
  
          });

  And in Node we simply resend via udp to the ffmpeg process :

  let client = dgram.createSocket('udp4'),
  
  
  
  client.send(chunk, 0, chunk.length, 41234, '127.0.0.1', function (err, bytes) {
  
              if (err) {
  
                  throw err;
  
              }
  
          });

  The resulting mp4 is viewable on VLC without any issues, despite many of the following warnings showing up in the FFMPEG console :

  stderr: frame= 1751 fps= 31 q=17.0 size=   15997kB time=00:00:58.39 bitrate=2244.3kbits/s dup=449 drop=32 speed=1.03x
  
  stderr: [mp4 @ 0x7fefe6019000] Non-monotonous DTS in output stream 0:1; previous: 2573919, current: 2573792; changing to 2573920. This may result in incorrect timestamps in the output file.
  
  
  
  stderr: frame= 1767 fps= 31 q=17.0 size=   16166kB time=00:00:58.90 bitrate=2248.4kbits/s dup=452 drop=32 speed=1.03x
  
  stderr: [libmp3lame @ 0x7fefe7004c00] Queue input is backward in time
  
  
  
  stderr: [mp4 @ 0x7fefe6019000] Non-monotonous DTS in output stream 0:1; previous: 2614768, current: 2614686; changing to 2614769. This may result in incorrect timestamps in the output file.

  However when I try to re stream this specific video on Youtube or any other RTMP platform (ex : Facebook) the audio is choppy, the command I use :

  ffmpeg -i ./test.mp4 -f flv rtmp://a.rtmp.youtube.com/live2/MYAPIKEY

  On the other hand any other "good" video not made from the aforementioned process re streams just fine and I suspect it might be from the warnings I’m getting.

  As an added bit of information streaming directly from my PC with avfoundation to youtube also works fine, and writing from avfoundation to a file and then streaming to youtube also works ok.

  The goal is to stream from my browser to the node server and directly to the Youtube RTMP without the choppy audio issue.

  If someone knows how to get rid of the warning to ensure ffmpeg isn’t the issue or can point me in the direction to achieve the desired result that’d be great.

• Matplotlib : ValueError : I/O operation on closed file [duplicate]

  28 septembre 2017, par gopi

  This question is an exact duplicate of :

  • animation.FuncAnimation ValueError : I/O operation on closed file

  %matplotlib inline
  
  import math,os,sys,numpy as np
  
  from numpy.random import random
  
  from matplotlib import pyplot as plt, rcParams, animation, rc
  
  from __future__ import print_function, division
  
  from ipywidgets import interact, interactive, fixed 
  
  from ipywidgets.widgets import *
  
  rc('animation', html='html5')
  
  rcParams['figure.figsize'] = 3,3
  
  %precision 4
  
  np.set_printoptions(precision=4, linewidth=100)
  
  
  
  def lin(a,x,b): return a * x + b
  
  
  
  a = 3.
  
  b = 8.
  
  n = 30
  
  
  
  x = random(n)
  
  y = lin(a,x,b)
  
  
  
  plt.scatter(x,y)
  
  
  
  def sse(y, y_pred): return ((y-y_pred)**2).sum()
  
  def loss(y, a, x, b): return sse(y, lin(a, x, b))
  
  def avg_loss(y, a, x, b): return np.sqrt(loss(y, a, x, b)/n)
  
  
  
  a_guess = -1
  
  b_guess = 1
  
  avg_loss(y, a_guess, x, b_guess)
  
  
  
  lr = 0.01
  
  #d[(y-(a*x+b))**2, b] = 2 (y_pred - y)
  
  #d[(y -(a*x+b)) **2, a] = x * dy/db
  
  
  
  def upd():
  
      global a_guess, b_guess
  
      y_pred = lin(a_guess, x, b_guess)
  
      dydb = 2 * (y_pred - y)
  
      dyda = x * dydb
  
      a_guess -= lr*dyda.mean()
  
      b_guess -= lr*dydb.mean()
  
  
  
  
  
  fig = plt.figure(dpi=100, figsize=(5,5))
  
  plt.scatter(x,y)
  
  line, = plt.plot(x, lin(a_guess, x, b_guess))
  
  plt.close()
  
  
  
  def animate(i):
  
      line.set_ydata(lin(a_guess, x, b_guess))
  
      for i in range(10): upd()
  
      return line,
  
  
  
  
  
  ani = animation.FuncAnimation(fig, animate, np.arange(0, 40), interval=100)
  
  ani

  But when I run it I get the following error,

      ValueError Traceback (most recent call last)
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/IPython/core/formatters.pyc in __call__(self, obj)
  
      309             method = get_real_method(obj, self.print_method)
  
      310             if method is not None:
  
  --> 311                 return method()
  
      312             return None
  
      313         else:
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in _repr_html_(self)
  
     1233         fmt = rcParams['animation.html']
  
     1234         if fmt == 'html5':
  
  -> 1235             return self.to_html5_video()
  
     1236 
  
     1237 
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in to_html5_video(self)
  
     1207                                 bitrate=rcParams['animation.bitrate'],
  
     1208                                 fps=1000. / self._interval)
  
  -> 1209                 self.save(f.name, writer=writer)
  
     1210 
  
     1211             # Now open and base64 encode
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in save(self, filename, writer, fps, dpi, codec, bitrate, extra_args, metadata, extra_anim, savefig_kwargs)
  
     1061                         # TODO: See if turning off blit is really necessary
  
     1062                         anim._draw_next_frame(d, blit=False)
  
  -> 1063                     writer.grab_frame(**savefig_kwargs)
  
     1064 
  
     1065         # Reconnect signal for first draw if necessary
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/contextlib.pyc in __exit__(self, type, value, traceback)
  
       33                 value = type()
  
       34             try:
  
  ---> 35                 self.gen.throw(type, value, traceback)
  
       36                 raise RuntimeError("generator didn't stop after throw()")
  
       37             except StopIteration, exc:
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in saving(self, *args, **kw)
  
      287             yield self
  
      288         finally:
  
  --> 289             self.finish()
  
      290 
  
      291     def _run(self):
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in finish(self)
  
      307     def finish(self):
  
      308         'Finish any processing for writing the movie.'
  
  --> 309         self.cleanup()
  
      310 
  
      311     def grab_frame(self, **savefig_kwargs):
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/matplotlib/animation.pyc in cleanup(self)
  
      346     def cleanup(self):
  
      347         'Clean-up and collect the process used to write the movie file.'
  
  --> 348         out, err = self._proc.communicate()
  
      349         self._frame_sink().close()
  
      350         verbose.report('MovieWriter -- '
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/subprocess32.pyc in communicate(self, input, timeout)
  
      925 
  
      926         try:
  
  --> 927             stdout, stderr = self._communicate(input, endtime, timeout)
  
      928         finally:
  
      929             self._communication_started = True
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/subprocess32.pyc in _communicate(self, input, endtime, orig_timeout)
  
     1711             if _has_poll:
  
     1712                 stdout, stderr = self._communicate_with_poll(input, endtime,
  
  -> 1713                                                              orig_timeout)
  
     1714             else:
  
     1715                 stdout, stderr = self._communicate_with_select(input, endtime,
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/subprocess32.pyc in _communicate_with_poll(self, input, endtime, orig_timeout)
  
     1767             select_POLLIN_POLLPRI = select.POLLIN | select.POLLPRI
  
     1768             if self.stdout:
  
  -> 1769                 register_and_append(self.stdout, select_POLLIN_POLLPRI)
  
     1770                 stdout = self._fd2output[self.stdout.fileno()]
  
     1771             if self.stderr:
  
  
  
  /home/gopinath/Workspace/anaconda3/envs/py27/lib/python2.7/site-packages/subprocess32.pyc in register_and_append(file_obj, eventmask)
  
     1746             poller = select.poll()
  
     1747             def register_and_append(file_obj, eventmask):
  
  -> 1748                 poller.register(file_obj.fileno(), eventmask)
  
     1749                 self._fd2file[file_obj.fileno()] = file_obj
  
     1750 
  
  
  
  ValueError: I/O operation on closed file

  I almost tried everywhere and I couldn’t find a solution, I tried reinstalling ffmpeg from the sources, moved this file to the folder where I installed ffmpeg and gave a try.

  I’m using python 3.5,
  Ubuntu 16.04
  Anaconda 4.4
  Matplotlib 2.02

  actually this particular course was taught using python 2.7 so I created a virenv installed 2.7 along with all related dependencies. So will it be the problem ?? I mean using python 2.7 on top of 3.5 ?