Recherche avancée

Sur d’autres sites (5923)

• tensorflow ffmpeg contrib output

  24 juin 2016, par ha_1694

  First, I want to say that I’m completely new to TensorFlow and machine learning in general.
  I’m looking at the ffmpeg contrib section and I see the example of decoding an audio file and I was wondering what is the vector output of decoding an audio file ? The code I’m running looks like :

  import tensorflow as tf
  
  from tensorflow.contrib import ffmpeg
  
  with tf.Session() as sess:
  
      audio_binary = tf.read_file('test.wav')
  
      waveform = ffmpeg.decode_audio(audio_binary, file_format='wav',
  
                                     samples_per_second=16000, channel_count=1)
  
      audio = waveform.eval()

  The audio variable then becomes a numpy.ndarray type that is a list of single element lists (about 620,000 elements total) that are all floats between (in my case) -0.50218207 and 0.39030123. My question is : What do these numbers mean ? Is this some sort of standard audio representation that I’m completely missing ?

• libx265 motion compensation and CU traverse

  17 juin 2016, par Ariana

  I’m trying to play with the H.265 motion compensation and search (HEVC- libx265 implementation from here : https://bitbucket.org/multicoreware/x265/downloads). I need to slightly extend the edge extension search, and fill the macro block with left-most pixels as if my sample videos are like cylinder (rightmost is connected to leftmost).

  What I need to do is basically this :

  

  One way to do that is to modify the edge extension area (which is already in the code, in the frameFilter.cpp), and do that for rightmost and fill parts of blocks which are out with leftmost pixels. I identified the piece of code here which apparently is responsible for that. Can someone help me with implementing this feature ?

  if ((col == 0) | (col == m_frameFilter->m_numCols - 1))
  
      {
  
          // TODO: improve by process on Left or Right only
  
          primitives.extendRowBorder(reconPic->getLumaAddr(m_rowAddr), stride, reconPic->m_picWidth, realH, reconPic->m_lumaMarginX);
  
  
  
          if (m_frameFilter->m_param->internalCsp != X265_CSP_I400)
  
          {
  
              primitives.extendRowBorder(reconPic->getCbAddr(m_rowAddr), strideC, reconPic->m_picWidth >> hChromaShift, realH >> vChromaShift, reconPic->m_chromaMarginX);
  
              primitives.extendRowBorder(reconPic->getCrAddr(m_rowAddr), strideC, reconPic->m_picWidth >> hChromaShift, realH >> vChromaShift, reconPic->m_chromaMarginX);
  
          }
  
      }
  
  
  
      // Extra Left and Right border on first and last CU
  
      if ((col == 0) | (col == m_frameFilter->m_numCols - 1))
  
      {
  
          copySizeY += lumaMarginX;
  
          copySizeC += chromaMarginX;
  
      }
  
  
  
      // First column need extension left padding area and first CU
  
      if (col == 0)
  
      {
  
          pixY -= lumaMarginX;
  
          pixU -= chromaMarginX;
  
          pixV -= chromaMarginX;
  
      }

• ffmpeg scale image with unrelated number in the name

  15 juin 2016, par loctrice

  I’m attempting to scale image(s) that have any given name. I’ve found my script is failing on files that have numbers in the name. "0% financing", "24 Hour", etc. The other files are working fine, so it’s not the script itself. I get :

  [image2 @ 0x7fbce2008000] Could find no file with path ’/path/to/0% image.jpeg’ and index in the range 0-4

  How can I tell ffmpeg that this isn’t a search pattern, or sequential numbered files ? There’s only 1 jpeg in each location, and I do not have control of the file names to change them.

  -update-
  I’ve figured out the command

  ffmpeg -pattern_type none -i /path/to/0%\ image/0%\ image.jpeg -vf scale=320:-1 /path/to/0%\ image/0%\ image.out.jpeg

  gets me past the initial problem, but the output won’t work because I can’t get it now to escape the final argument. If I am in the directory (so no path) and change the output to just out.jpeg it will work, so I’m confident the first error is corrected.

  Now I need to figure out how to use spaces in the path in the output argument ? I’ve tried surrounding it in quotes :

  "0% image.out.jpeg"

  regular escapes :

  0%\ image.out.jpeg

  and surrounding it in quotes and using escapes at the same time :

  "0%\ image.out.jpeg"