Recherche avancée

Sur d’autres sites (9523)

• How to create movie screenshot by ffmpeg in an amazon S3 path

  5 décembre 2019, par user2004082

  I tried to create using ffmpeg a video screenshot from a remote video url in heroku console. Below is how I generated a movie instance and can see also an empty ready to be written file at S3. But the last line movie.screenshot is not working and generates this error :

  FFMPEG::Error: Failed encoding.Errors: no output file created

  Here is the code

  s3 = Aws::S3::Resource.new(region: 'us-west-1')
  
  bucket = s3.bucket("ruby-sample-kb-#{SecureRandom.uuid}")
  
  bucket.create
  
  object = bucket.object('ex-vid-test-kb.jpg')
  
  object.put(acl: "public-read-write")
  
  path = object.public_url
  
  movie = FFMPEG::Movie.new("https://www.googleapis.com/download/storage/v1/b/seppoav/o/3606137_51447286560__56BAF29C-05CB-4223-BAE6-655DF2236321.MOV?generation=1492780072394755&alt=media")
  
  movie.screenshot(path, :seek_time => 2)

  I also tried the following line just if it should be written via put. What am I missing here ?

  object.put(acl: "public-read", body: movie.screenshot(path, :seek_time => 2))

• lavf/ftp: check for truncation in snprintf

  25 novembre 2021, par Anton Khirnov

  lavf/ftp: check for truncation in snprintf
  Silences e.g. the following warning in gcc 10 :
  
  src/libavformat/ftp.c : In function ‘ftp_move’ :
  
  src/libavformat/ftp.c:1122:46 : warning : ‘%s’ directive output may be truncated writing up to 4095 bytes into a region of size 4091 [-Wformat-truncation=]
  
   1122 |     snprintf(command, sizeof(command), "RNTO %s\r\n", path) ;
  
        |                                              ^  
  
  src/libavformat/ftp.c:1122:5 : note : ‘snprintf’ output between 8 and 4103 bytes into a destination of size 4096
  
   1122 |     snprintf(command, sizeof(command), "RNTO %s\r\n", path) ;
  
        |     ^ 
  

  • [DH] libavformat/ftp.c

• Low performance when extracting frames from video

  25 février 2016, par Rakatan

  I am writing a video processing application and have hit a bit of a snag.

  The idea is to extract frames (images) from a video file, process them on the GPU and then write them back to a new video file.

  So far i have been successful in doing this using JavaCV.

  The only problem is that extracting the frames takes a very long time (in the 200ms region). This is way to much, as added to the GPU processing time and the time it takes to write to a new file, it results in a total process duration of about 700ms per frame.

  This is how i grab the frames :

   FFmpegFrameGrabber frameGrabber = new FFmpegFrameGrabber(videoLocation);
  
   frameGrabber.start();
  
  
  
   frameGrabber.setFrameNumber(frameNumber);
  
   frame = frameGrabber.grabImage();

  Pretty standard, nothing special, and it works. I am able to process the frame, convert it to a bitmap, etc...

  As i have said before, the bottom two instructions (setting the frame number and grabbing the image) take a very very long time. Writing the frames also takes a fair bit.

  I would greatly appreciate if anyone has any input regarding this issue. Maybe i’m not setting up something correctly, maybe there is another solution that is faster.