Recherche avancée

Sur d’autres sites (4342)

• avformat/vobsub : fix several issues.

  29 septembre 2013, par Clément Bœsch

  avformat/vobsub : fix several issues.
  Here is an extract of fate-samples/sub/vobsub.idx, with an additional
  
  text at the end of each line to better identify each bitmap :
      timestamp : 00:04:55:445, filepos : 00001b000 Ace !
  
      timestamp : 00:05:00:049, filepos : 00001b800 Wake up, honey !
  
      timestamp : 00:05:02:018, filepos : 00001c800 I gotta go to work.
  
      timestamp : 00:05:02:035, filepos : 00001d000 < ???>
  
      timestamp : 00:05:04:203, filepos : 00001d800 Look after Clayton, okay ?
  
      timestamp : 00:05:05:947, filepos : 00001e800 I’ll be back tonight.
  
      timestamp : 00:05:07:957, filepos : 00001f800 Bye ! Love you.
  
      timestamp : 00:05:21:295, filepos : 000020800 Hey, Ace ! What’s up ?
  
      timestamp : 00:05:23:356, filepos : 000021800 Hey, how’s it going ?
  
      timestamp : 00:05:24:640, filepos : 000022800 Remember what today is ? The 3rd !
  
      timestamp : 00:05:27:193, filepos : 000023800 Look over there !
  
      timestamp : 00:05:28:369, filepos : 000024800 Where are they going ?
  
      timestamp : 00:05:28:361, filepos : 000025000 < ???>
  
      timestamp : 00:05:29:946, filepos : 000025800 Let’s go see.
  
      timestamp : 00:05:31:230, filepos : 000026000 I can’t, man. I got Clayton.
  Note the two "< ???>" : they are basically split subtitles (with the
  
  previous one), which the dvdsub decoder is now supposed to reconstruct
  
  with a previous commit. But also note that while the first chunk has
  
  increasing timestamps,
      timestamp : 00:05:02:018, filepos : 00001c800
  
      timestamp : 00:05:02:035, filepos : 00001d000
  ...it’s not the case of the second one (and this is not an exception in the
  
  original file) :
      timestamp : 00:05:28:369, filepos : 000024800
  
      timestamp : 00:05:28:361, filepos : 000025000
  For the dvdsub decoder, they need to be "filepos’ed" ordered, but the
  
  FFDemuxSubtitlesQueue is timestamps ordered, which is the reason of the
  
  introduction of a sub sort method in the context, to allow giving
  
  priority to the position, and then the timestamps. With that change, the
  
  dvdsub decoder get fed with ordered packets.
  Now the packet size estimation was also broken : the filepos differences
  
  in the vobsub index defines the full data read between two subtitles
  
  chunks, and it is necessary to take into account what is read by the
  
  mpegps_read_pes_header() function since the length returned by that
  
  function doesn’t count the size of the data it reads. This is fixed with
  
  the introduction of total_read, and old,new_pos. By doing this change,
  
  we can drop the unreliable len16 heuristic and simplify the whole loop.
  
  Note that mpegps_read_pes_header() often read more than one PES packet
  
  (typically in one call it can read 0x1ba and 0x1be chunk along with the
  
  relevant 0x1bd packet), which triggers the "total_read + pkt_size >
  
  psize" check. This is an expected behaviour, which could be avoided by
  
  having a more chunked version of mpegps_read_pes_header().
  The latest change is the extraction of each stream into its own
  
  subtitles queue. If we don’t do this, the maximum size for a subtitle
  
  chunk is broken, and the previous changes can not work. Having each
  
  stream in a different queue requires some little adjustments in the
  
  seek code of the demuxer.
  This commit is only meaningful as a whole change and can not be easily
  
  split. The FATE test changes because it uses the vobsub demuxer.
  

  • [DH] libavformat/mpeg.c
  • [DH] libavformat/subtitles.c
  • [DH] libavformat/subtitles.h
  • [DH] tests/ref/fate/sub2video

• Revision 5ebe94f9f1 : Build fixes to merge vp9-preview into master Various fixups to resolve issues w

  23 décembre 2012, par John Koleszar

  Changed Paths : Add /examples/decode_with_partial_drops.txt Modify /libs.mk Modify /test/test.mk Modify /test/test_libvpx.cc Modify /vp9/common/generic/vp9_systemdependent.c Modify /vp9/common/vp9_alloccommon.c Modify /vp9/common/vp9_blockd.h (...)

• Memory issues when using ffmpeg and gloss to play videos

  21 décembre 2015, par Noughtmare

  I’m trying to make a video player with haskell using ffmpeg-light, JuicyPixels and gloss. I’m now able to play video, but frames that have been played stay in memory. This causes major memory issues. How can I avoid storing all the frames in memory ?

  Here is my code :

  {-# LANGUAGE FlexibleContexts #-}
  
  module Main where
  
  
  
  -- For my own code:
  
  import Graphics.Gloss
  
  import Codec.FFmpeg
  
  import Codec.FFmpeg.Juicy
  
  import Codec.Picture
  
  import Control.Applicative
  
  import Data.Maybe
  
  import Graphics.Gloss.Juicy
  
  import Control.Monad (when, join)
  
  import Codec.FFmpeg.Decode
  
  import Codec.FFmpeg.Enums
  
  import Control.Monad.Error.Class
  
  import Control.Arrow (first)
  
  import Control.Monad.Except (runExceptT)
  
  import Graphics.Gloss.Interface.IO.Animate
  
  import Data.IORef
  
  
  
  -- Temporary hardcoded resolution
  
  resolution :: (Int,Int)
  
  resolution = (640, 360)
  
  
  
  main :: IO ()
  
  main = do
  
      -- First initialize ffmpeg, this needs to be run before other ffmpeg functions
  
      initFFmpeg
  
      -- Open the samplevideo for reading. video :: IO (IO (Maybe (AVFrame, Double)), IO ())
  
      video <- runExceptT $ frameReaderTime' avPixFmtRgb24 "SampleVideo_640x360_1mb.flv"
  
      either
  
            -- This code gets called when the frameReader reports an error
  
            (const $ putStrLn "Can't read file")
  
            -- This opens a new window and plays the video in it on a white background
  
            (animateFixedIO (InWindow "Nice Window" resolution (10, 10)) white . frameAtWait . fst)
  
  
  
            video
  
  
  
  -- This finds the frame at given time
  
  frameAtWait :: IO (Maybe (AVFrame, Double)) -> Float -> IO Picture
  
  frameAtWait getFrame time = do
  
      -- This gets the next frame from the video
  
      (frame, t) <- fromJust <$> getFrame
  
      -- t has to be converted from Double to Float
  
      let t' = realToFrac t
  
          -- The difference between the requested time and the actual frame time
  
          difference = t' - time
  
      -- If the frame is not yet supposed to be shown
  
      if difference > 0 then do
  
          -- Wait until it is
  
          threadDelay . round . (* 1000000) $ difference
  
          -- then return it
  
          fromJust <$> frameToPicture frame
  
      else
  
          -- return it immediately
  
          fromJust <$> frameToPicture frame
  
  
  
  -- This function converts a ffmpeg internal AVFrame to a gloss picture
  
  frameToPicture :: AVFrame -> IO (Maybe Picture)
  
  frameToPicture frame = do
  
      -- convert it to a juicypixels dynamicimage
  
      dynImage <- toJuicy frame
  
      -- then convert it to a gloss picture and return it
  
      return . join $ fmap fromDynamicImage dynImage