Recherche avancée

Sur d’autres sites (6016)

• FFMpeg - Split Window RTMP - Delay on Second Stream

  22 février 2016, par Nick Smit

  I’m trying to combine two live RTMP sources into one split screen output with combined audio. The output is then sent on to a receiving RTMP server.

  Using the following command, which uses the same RTMP input for both feeds, I’ve managed to get the above working, however the input on the left is delayed by about 2 seconds from the one on the right.

  ffmpeg -re -i rtmp://myserver.tld/live/stream_key -re -i rtmp://myserver.tld/live/stream_key \
  
  -filter_complex "\
  
  nullsrc=size=1152x720 [base];\
  
  [0:v] crop=576:720 [upperleft];\
  
  [1:v] crop=576:720 [upperright];\
  
  [base][upperleft] overlay=shortest=1 [tmp1];\
  
  [tmp1][upperright] overlay=shortest=1:x=576;\
  
  [0:a][1:a]amix \
  
  " -c:a libfdk_aac -ar 44100 -threads 32 -c:v libx264 -g 50 -preset ultrafast -tune zerolatency -f flv rtmp://myserver.tld/live/new_stream_key

  Output :

  ffmpeg version N-76137-gb0bb1dc Copyright (c) 2000-2015 the FFmpeg developers
  
  built with gcc 4.8 (Ubuntu 4.8.4-2ubuntu1~14.04)
  
  configuration: --prefix=/home/ubuntu/ffmpeg_build --pkg-config-flags=--static --extra-cflags=-I/home/ubuntu/ffmpeg_build/include --extra-ldflags=-L/home/ubuntu/ffmpeg_build/lib --bindir=/home/ubuntu/bin --enable-gpl --enable-libass --enable-libfdk-aac --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libx265 --enable-nonfree
  
  libavutil      55.  4.100 / 55.  4.100
  
  libavcodec     57.  7.100 / 57.  7.100
  
  libavformat    57.  8.102 / 57.  8.102
  
  libavdevice    57.  0.100 / 57.  0.100
  
  libavfilter     6. 12.100 /  6. 12.100
  
  libswscale      4.  0.100 /  4.  0.100
  
  libswresample   2.  0.100 /  2.  0.100
  
  libpostproc    54.  0.100 / 54.  0.100
  
  [flv @ 0x3a0e940] video stream discovered after head already parsed
  
  Input #0, flv, from 'rtmp://myserver.tld/live/stream_key':
  
  Metadata:
  
  Server          : NGINX RTMP (github.com/arut/nginx-rtmp-module)
  
  displayWidth    : 1152
  
  displayHeight   : 720
  
  fps             : 29
  
  profile         :
  
  level           :
  
  Duration: 00:00:00.00, start: 5.717000, bitrate: N/A
  
  Stream #0:0: Audio: aac (LC), 48000 Hz, stereo, fltp, 163 kb/s
  
  Stream #0:1: Video: h264 (High), yuv420p, 1152x720, 30.30 fps, 29.97 tbr, 1k tbn, 59.94 tbc
  
  [flv @ 0x3a49e00] video stream discovered after head already parsed
  
  Input #1, flv, from 'rtmp://myserver.tld/live/stream_key':
  
  Metadata:
  
  Server          : NGINX RTMP (github.com/arut/nginx-rtmp-module)
  
  displayWidth    : 1152
  
  displayHeight   : 720
  
  fps             : 29
  
  profile         :
  
  level           :
  
  Duration: 00:00:00.00, start: 9.685000, bitrate: N/A
  
  Stream #1:0: Audio: aac (LC), 48000 Hz, stereo, fltp, 163 kb/s
  
  Stream #1:1: Video: h264 (High), yuv420p, 1152x720, 30.30 fps, 29.97 tbr, 1k tbn, 59.94 tbc
  
  [libx264 @ 0x3a9cd60] Application has requested 32 threads. Using a thread count greater than 16 is not recommended.
  
  [libx264 @ 0x3a9cd60] using SAR=1/1
  
  [libx264 @ 0x3a9cd60] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX
  
  [libx264 @ 0x3a9cd60] profile Constrained Baseline, level 3.1
  
  [libx264 @ 0x3a9cd60] 264 - core 142 r2389 956c8d8 - H.264/MPEG-4 AVC codec - Copyleft 2003-2014 - http://www.videolan.org/x264.html - options: cabac=0 ref=1 deblock=0:0:0 analyse=0:0 me=dia subme=0 psy=1 psy_rd=1.00:0.00 mixed_ref=0 me_range=16 chroma_me=1 trellis=0 8x8dct=0 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=0 threads=11 lookahead_threads=11 sliced_threads=1 slices=11 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=0 weightp=0 keyint=50 keyint_min=5 scenecut=0 intra_refresh=0 rc=crf mbtree=0 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=0
  
  Output #0, flv, to 'rtmp://myserver.tld/live/new_stream_key':
  
  Metadata:
  
  Server          : NGINX RTMP (github.com/arut/nginx-rtmp-module)
  
  displayWidth    : 1152
  
  displayHeight   : 720
  
  fps             : 29
  
  profile         :
  
  level           :
  
  encoder         : Lavf57.8.102
  
  Stream #0:0: Video: h264 (libx264) ([7][0][0][0] / 0x0007), yuv420p, 1152x720 [SAR 1:1 DAR 8:5], q=-1--1, 25 fps, 1k tbn, 25 tbc (default)
  
  Metadata:
  
  encoder         : Lavc57.7.100 libx264
  
  Stream #0:1: Audio: aac (libfdk_aac) ([10][0][0][0] / 0x000A), 44100 Hz, stereo, s16, 128 kb/s (default)
  
  Metadata:
  
  encoder         : Lavc57.7.100 libfdk_aac
  
  Stream mapping:
  
  Stream #0:0 (aac) -> amix:input0
  
  Stream #0:1 (h264) -> crop
  
  Stream #1:0 (aac) -> amix:input1
  
  Stream #1:1 (h264) -> crop
  
  overlay -> Stream #0:0 (libx264)
  
  amix -> Stream #0:1 (libfdk_aac)
  
  Press [q] to stop, [?] for help
  
  [flv @ 0x3a0e940] Thread message queue blocking; consider raising the thread_queue_size option (current value: 512)
  
  frame=   81 fps= 20 q=15.0 size=     674kB time=00:00:03.24 bitrate=1703.3kbits/frame=  102 fps= 22 q=22.0 size=     945kB time=00:00:04.08 bitrate=1896.4kbits/

  Is there any way to force FFMpeg to read both RTMP inputs at the same time ?

• Video rendering in OpenGL on Qt5

  28 janvier 2016, par Bobnono

  I’m trying to render my decoded frame from FFMPEG to an QOpenGLWindow.
  FFMPEG give me NV12 AVFrame or YUV420p or RGB.
  I choose the simplest RGB.

  I created an c_gl_video_yuv class inherit from QOpenGLWindow and QOpenGLFunctions_3_3.

  I want to use shader to draw my rectangle and texture it with the video frame. (for YUV I want the sader to convert it in RGB and apply texture)

  my c_gl_video_yuv class is define as below :

  class c_gl_video_yuv : public QOpenGLWindow,protected QOpenGLFunctions_3_3_Core
  
  {
  
  public:
  
     c_gl_video_yuv();
  
     ~c_gl_video_yuv();
  
  ---
  
     void update_texture(AVFrame *frame, int w, int h);
  
  
  
  protected:
  
     void initializeGL();
  
     void paintGL();
  
     void resizeGL(int width, int height);
  
  
  
     void paintEvent(QPaintEvent *);
  
  
  
  private:
  
  
  
     ---
  
     GLuint textures[2];
  
     --- 
  
     // Shader program
  
     QOpenGLShaderProgram *m_program;
  
  
  
     GLint locVertices;
  
     GLint locTexcoord;
  
  
  
  };

  I initialise the OpenGL :

     void c_gl_video_yuv::initializeGL()
  
      {
  
        // Init shader program
  
          initializeOpenGLFunctions();
  
          glGenTextures(2, textures);
  
          /* Apply some filter on the texture */
  
          glBindTexture(GL_TEXTURE_2D, textures[0]);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
  
  
  
          glBindTexture(GL_TEXTURE_2D, textures[1]);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
  
  
  
          qDebug() << "c_gl_video_yuv::initializeGL() initialise shader"<< endl;
  
          m_program = new QOpenGLShaderProgram(this);
  
          m_program->addShaderFromSourceFile(QOpenGLShader::Vertex, ":/shader/Ressources/vertex_yuv.glsl");
  
          m_program->addShaderFromSourceFile(QOpenGLShader::Fragment, ":/shader/Ressources/rgb_to_rgb_shader .glsl");
  
          m_program->link();
  
  
  
      //    /* Grab location of shader attributes. */
  
          locVertices = m_program->attributeLocation("position");
  
          locTexcoord = m_program->attributeLocation("texpos");
  
  
  
          /* Enable vertex arrays to push the data. */
  
          glEnableVertexAttribArray(locVertices);
  
          glEnableVertexAttribArray(locTexcoord);
  
  
  
          /* set data in the arrays. */
  
              glVertexAttribPointer(locVertices, 2, GL_FLOAT, GL_FALSE, 0,
  
                            &vertices[0][0]);
  
              glVertexAttribPointer(locTexcoord, 2, GL_FLOAT, GL_FALSE, 0,
  
                            &texcoords[0][0]);
  
  
  
          // GL options
  
          glEnable(GL_DEPTH_TEST);    
  
      }

  And I render

  void c_gl_video_yuv::paintGL()
  
      {
  
  
  
          qDebug() <<  "paintGL()   set viewport "<* Clear background. */
  
          glClearColor(0.5f,0.5f,0.5f,1.0f);
  
          glClear(GL_COLOR_BUFFER_BIT);
  
  
  
          if(first_frame)
  
          {
  
              qDebug() <<  "paintGL() Bind shader" <<  endl;
  
              m_program->bind();
  
  
  
              /* Get Ytex attribute to associate to TEXTURE0 */
  
                  m_program->bindAttributeLocation("Ytex",0);
  
                  m_program->bindAttributeLocation("UVtex",1);
  
  
  
                  qDebug() <<  "paintGL() Bind texture" <<  endl;
  
              if(!is_init)
  
              {
  
                  glActiveTexture(GL_TEXTURE0);
  
                  glBindTexture(GL_TEXTURE_2D, textures[0]);
  
                  glTexImage2D( GL_TEXTURE_2D, 0, GL_RGB, frame_width, frame_height, 0, GL_RGB, GL_UNSIGNED_BYTE, frame_yuv->data[0] );
  
  
  
                  glActiveTexture(GL_TEXTURE1);
  
                  glBindTexture(GL_TEXTURE_2D, textures[1]);
  
                  glTexImage2D( GL_TEXTURE_2D, 0, GL_RGB, frame_width, frame_height, 0, GL_RGB, GL_UNSIGNED_BYTE, frame_yuv->data[0] );
  
                  is_init = true;
  
              }
  
              else
  
              {
  
                  glActiveTexture(GL_TEXTURE0);
  
                  glBindTexture(GL_TEXTURE_2D, textures[0]);
  
                 glTexSubImage2D(GL_TEXTURE_2D, 0, 0,0, frame_width, frame_height, GL_RGB, GL_UNSIGNED_BYTE, frame_yuv->data[0]);
  
  
  
                 glActiveTexture(GL_TEXTURE1);
  
                 glBindTexture(GL_TEXTURE_2D, textures[1]);
  
                 glTexSubImage2D(GL_TEXTURE_2D, 0, 0,0, frame_width, frame_height, GL_RGB, GL_UNSIGNED_BYTE, frame_yuv->data[0]);
  
              }
  
  
  
  
  
  
  
              glVertexAttribPointer(locVertices, 2, GL_FLOAT, GL_FALSE, 0, vertices);
  
              glVertexAttribPointer(locTexcoord, 2, GL_FLOAT, GL_FALSE, 0, texcoords);
  
  
  
              glEnableVertexAttribArray(0);
  
              glEnableVertexAttribArray(1);
  
  
  
              glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
  
  
  
              glDisableVertexAttribArray(1);
  
              glDisableVertexAttribArray(0);
  
              m_program->release();
  
          }
  
      }

  the vertex shader is :

   #version 330 core
  
      attribute vec3 position;
  
      attribute vec2 texpos;
  
      varying vec2 opos;
  
      void main(void)
  
      {
  
          opos = texpos;
  
          gl_Position = vec4(position, 1.0);
  
      }

  and the Fragment shader is :

  version 330 core
  
  in vec2 TexCoords;
  
  out vec4 color;
  
  uniform sampler2D YImage;
  
  uniform sampler2D UVImage;
  
  const vec3 r_c = vec3(1.164383,  0.000000,  1.596027);
  
  const vec3 g_c = vec3(1.164383, -0.391762, -0.812968);
  
  const vec3 b_c = vec3(1.164383,  2.017232,  0.000000);
  
  const vec3 offset = vec3(-0.0625, -0.5, -0.5);
  
  void main()
  
  {
  
     float y_val = texture(YImage, TexCoords).r;
  
     float u_val = texture(UVImage, TexCoords).r;
  
     float v_val = texture(UVImage, TexCoords).g;
  
     vec3 yuv = vec3(y_val, u_val, v_val);
  
     yuv += offset;
  
     color.r = dot(yuv, r_c);
  
     color.g = dot(yuv, g_c);
  
     color.b = dot(yuv, b_c);
  
     color.a = 1.0;
  
  };

  (for RGB frame a replace thevec3(1.164383, 0.000000, 1.596027) ; by vec3(1.0, 1.000000, 1.0) ;

  So before i receive a frame it render nothing, just a grey window - noarmal
  After I receive a frame, the textures are uploaded and the shaders normaly create the player. But nothing appear not even a black rectangle nothing, just plain grey.

  What is wrong ?
  It is not the right way to upload texture, or my vertex are not created ?
  of course i declarde my vetrtices eand texture coordonate

  const GLfloat vertices[][2] = {
  
      {-1.f, -1.f},
  
      {1.f, -1.f},
  
      {-1.f, 1.f},
  
      {1.f, 1.f}
  
  };
  
  const GLfloat texcoords[][2] = {
  
      {0.0f, 1.0f},
  
      {1.0f, 1.0f},
  
      {0.0f, 0.0f},
  
      {1.0f, 0.0f}
  
  };

  I am very new to OpenGL so it is quite fuzzy in my head, but I thought it is not really hard to draw a rectangle with a streaming texture.
  Maybe i should use vbo or fbo but I still don’t really understand this.

  If someone can help me, I will appreciate !
  Thanks

• How can combine two separate scripts being piped together to make one script instead of two ?

  27 mars 2016, par user556068

  For the past couple hours I’ve been banging my head against the wall trying to figure out something I thought would be simple. Maybe it is but it’s beyond me at the moment. So I have now two scripts. Originallly they were part of the same but I could never make it work how it should. So the first part uses curl to download a file from a site. Then using grep and sedto filter out the text I need which is then put into a plain text file as a long list of website urls ; one per line. The last part of the 1st script calls on youtube -dl to read the batch file in order to obtain the web addresses where the actual content is located. I hope that makes sense.

  youtube-dl reads the batch file and outputs a new list urls into the terminal. This second list is not saved to file because it doesn’t need to be. These urls change from day to day or hour to hour. Using the read command, these urls are then passed to ffmpeg using a predetermined set of arguments for the input and output. Ffmpeg is executed on every url it receives and runs quietly in the background.

  The first paragraph describes script1.sh and paragraph 2 obviously describes script2.sh. When I pipe them together like script1.sh | script2.sh it works better than I ever thought possible. Maybe i’m nitpicking at this point but the idea is to have 1 unified script. For the moment I have simplified it by adding an alias to my .bash_profile.

  Here are the last two commands of script1.

  sed 's/\"\,/\//g' > "$HOME/file2.txt";
  
  cat $HOME/file2.txt | youtube-dl --ignore-config -iga -

  The trailing - allows youtube-dl to read from stdin.

  The second part of the script ; what I’m calling script2 at this point begins with

  while read -r input
  
  do
  
  ffmpeg [arg] [input] [arg2] [output]

  What am i not seeing that is causing the script to hang when the two halves are combined yet work perfectly if one is piped into the other ?