Recherche avancée

Sur d’autres sites (6976)

• Getting Intra-Frames from a video

  2 juin 2017, par Thomas

  First, let me preface this with saying that I have to process a massive amount of videos and shaving even 1 second per video is very significant.

  I’m trying to extract the position of all intra-frames in a video.

  FFProbe works, but it’s amazingly slow.

  FFMpeg works much faster, but it’s still quite slow.

  So, my question is if there is a lib, where I could do something like this pseudo code :

  var frame = first_frame;
  
  do
  
  {
  
      if (frame.type == intra_frame) log(frame.time);
  
      frame = frame.next;
  
  } while (frame != null);

  and scan through the video very fast, not looking at anything inside the frames, but just the headers to find the type of the frame and move on to the next.

  The end goal is to generate filmstrips, for preview, as fast as possible. I need to generate 48 thumbnails roughly evenly spread through the movie and covering the whole movie span (excluding 5 sec on each end). I am hoping to speed the process up by extracting only intra-frames. Since I need a fixed number of thumbnails, I can only do this by knowing the total amount of intra-frames.

  This scenario works only if the time to find all the intra-frame indices and extract 48 intra frames is less than getting frames, in a single pass, at regular intervals, 48 times.

  ---

  Edit : This is the solution I have found so far ; in this example I take one frame every minute

  ffmpeg -hide_banner -skip_frame nokey -i _a.mp4 -vf "select='eq(pict_type\,PICT_TYPE_I)*(lt(abs(t\-30)\,2)+lt(abs(t\-90)\,2)+lt(abs(t\-150)\,2)+lt(abs(t\-180)\,2)+lt(abs(t\-240)\,2)+lt(abs(t\-300)\,2)+lt(abs(t\-360)\,2)+lt(abs(t\-420)\,2))',showinfo,scale=640:480,tile=6x8" -vsync 0 frames.jpg

  I found that if I specify the frames directly, I don’t really get a speed difference.

  Short of going through the API, is there anything faster with FFMPEG ?

• Anomalie #4363 (Nouveau) : mot de passe vide bloque le formulaire de réinitialisation du mot de pa...

  1er août 2019, par cy_altern -

  Scénario type :

  • un administrateur crée des comptes pour de futurs utilisateurs en renseignant nom, mail, login mais SANS mot de passe
  • il envoie un mail à ces utilisateurs en leur donnant l’URL de la page "J’ai perdu mon mot de passe" (.../spip.php ?page=spip_pass) afin qu’ils génèrent leur mot de passe
  • l’utilisateur saisit son mail dans le formulaire de cette page squelettes-dist/formulaires/oubli.html et il se prend en retour "Erreur : vous n’avez plus accès à ce site."
    Pour l’instant la seule alternative est de créer un mot de passe quelconque lors de la création du compte de l’utilisateur.

  Y a t’il une raison valide pour expliquer cette restriction de mot de passe vide à la réinitialisation ?

  Sinon, le patch semble rapide :
  ligne 85 de squelettes-dist/formulaires/oubli.php
  remplacer :
  

  <span class="CodeRay">} elseif ($r[1]['statut'] == '5poubelle' or $r[1]['pass'] == '') {
  </span>
  Télécharger

  
  par
  

  <span class="CodeRay">} elseif ($r[1]['statut'] == '5poubelle') {
  </span>
  Télécharger

• Python get Audio/Video frames separately from video file

  3 juillet 2016, par AlvaroAV

  What i’m trying to do :

  Hi !

  I’m trying to store the Video and Audio information from a video file. I would like to store video frames and audio frames separately in different variables.

  My intention is to manage video/files and do some actions with the audio and video frame list, but to do what I’m plannign to do I need to store this audio/video frames separately. I’ve read a lot of questions in StackOverflow about python and audio/video managing.

  Most people recommend to use OpenCV or ffmpeg to manage videos. I saw some scripts using these libraries to get video(only video) frames, but none of them are getting audio, most of them are just getting video frames and save them as RGB images. I also check some scripts where people get audio frames from a mp3 file, but I’m not sure if you can do that in a video file

  Most important thing to me is to know the best way to manage video and audio separately. I’m not looking for people to do my code, just asking to point me in a good direction.

  One of the things I’m trying to do is to send this information via socket, but as I said I need the audio and video frames to be in separated variables (yes, i’m wondering about an stream app, but that’s not the only thing I’m trying to do)

  I know I should give more information, and maybe show some code, but I don’t have any concret code I tried some things, but I’ve never been capable to separate audio and video. I know that each format has his own encryption, and at the end I decided to use "mp4" as video format but I don’t know neither if this is the best format for what I’m trying to do.

  Resume :

  Is openCV the best way to manage video and audio separately ?

  Wich is the easiest way to separate video and audio frames ? Is it possible ?

  Wich is the best documentation I should read to learn about video/audio management ?

  I would like to do the things with my own code, and use in the less way possible openCV or other libraries.

  My "basic" idea is to get a "list" of audio and video frames, and then I would like to do some operations, but right now I can’t find the best way for me to manage a vide using python. I even wonder if could be possible to manage a video as raw data

  I need to know wich is the best library to manage videos using python, for me the best library, will be the one that allows me to manage the videos more "freely"

  I’ve already checked :

  I’ve read too many questions on this theme, the most recent are :

  How to extract audio from video file

  Split audio video separately from given video using MLT

  Embed audio video in python gui