Recherche avancée

Sur d’autres sites (4385)

• ffmpeg / avconv slideshow with transitions and Python

  24 janvier 2015, par KravAn

  As i understand ffmpeg / avconv is good for creating slideshow(video).
  But how do I create a slideshow with different effects with Python (Django possibly)..
  ie : fading of image, fading image one into another, maybe other effects between transfer from one to another.

• drawtext with ffmpeg python

  22 mai 2024, par Wolf Wolf

  I am trying to add a text to a video using ffmpeg and python.
I tried to do this in the following ways, but it didn't work.

  


  first

  


  (
         ffmpeg
         .input(in_video)
         .filter('drawtext',
                 fontsize=30,
                 fontfile=r"D:\projects\python\editor_bot\downloads\Candara.ttf",
                 text='test test test.',
                 x='if (eq(mod(t\\, 15)\\, 0)\\, rand(0\\, (w-text_w))\\, x)',
                 y='if (eq(mod(t\\, 10)\\, 0)\\, rand(0\\, (h-text_h))\\, y)')
         .output(f'output-final.mp4')
         .run()
 )


  


  second

  


  fil = fr"drawtext=text='test test test':fontsize=30:fontfile=':\projects\python\editor_bot\downloads\Candara.ttf':x='if (eq(mod(t\, 15)\, 0)\, rand(0\, (w-text_w))\, x)':y='if (eq(mod(t\, 10)\, 0)\, rand(0\, (h-text_h))\, y)'"
    (
        ffmpeg
        .input(in_video)
        .output(f'output-final.mkv', filter_complex=fil)
        .run()
    )


  


  But by running this command

  


  ffmpeg -i v1.mp4 -filter:v drawtext="fontsize=30:fontfile=candara.ttf:text='testtest test.':x=if(eq(mod(t\,15)\,0)\,rand(0\,(w-text_w))\,x):y=if(eq(mod(t\,10)\,0)\,rand(0\,(h-text_h))\,y)" -c:a copy -c:v libx264 -preset slow -crf 18 V13.mkv


  


  In the terminal, what I want is done exactly

  


  thanks

  


• Preventing the python and ffmpeg heroku buildpacks from overwriting LD_LIBRARY_PATH

  3 mars 2014, par Simon

  I'm deployng a Django app to heroku, which requires ffmpeg. To accomplish this I am using heroku-buildpack-multi to install both heroku-buildpack-ffmpeg and heroku-buildpack-python, and all of that works fine. The problem is my that app also depends on django-pylibmc-sasl, python-memcached, pylibmc et al. which, as per usual, heroku senses and automatically installs libmemcached for me.

  Here's where something goes a little wrong. If I remove the custom buildpack everything runs fine (except for ffmpeg obviously). As soon as I add it in, however, while I can run ffmpeg, python fails on import pylibmc (or rather on import _pylibmc inside the module itself). After a amount of head-scratching I decided to have a look at the environment variables, here's what I got :

  With only the Python buildpack enabled :

  LD_LIBRARY_PATH=/app/.heroku/vendor/lib

  With both the Python and the ffmpeg buildpacks enabled :

  LD_LIBRARY_PATH=:vendor/ffmpeg/lib

  It looks like one or both of the buildpacks simply overwrites the other, or avoids setting the variable should it be already set. The ffmpeg buildpack seems to set LD_LIBRARY_PATH in a way that looks kosher to me, while the Python buildpack does a few things that I don't really understand the reason for.

  Solution

  Anyway, after manually overriding the library path using heroku config:set LD_LIBRARY_PATH=/app/.heroku/vendor/lib:vendor/ffmpeg/lib I am able to use both libmemcached and ffmpeg, but it doesn't feel too robust. What if something changes in one of the buildpacks path settings, or I add another buildpack - then I would have to manually edit the library path variable.

  Better solution ?

  So, while this is not really an urgent question at all, I simply would like to know :

  • Is there a better way of solving this issue ?
  • Might I have made some configuration error leading up to the path conflict ?
  • Should this be considered a bug in either of the buildpacks ?