Recherche avancée

Sur d’autres sites (7767)

• Anomalie #3557 : Header composed-by tronqué

  25 décembre 2017

  C’’est surtout que j’ai du rajouter https://zone.spip.org/trac/spip-zone/browser/_dev_/univers_spip/inc/univers_analyser.php#L135 à 139 pour que ça marche aussi dans le cas de mutu "assez" classiques.
  Mais que ça reste du spécifique bricolé.

  Je relance donc l’idée :

  Peut-être que pour améliorer les perfs en ayant un entête plus concis, le Composed-by pourrait ne contenir que :
  SPIP 3.1.0-beta @ www.spip.net + URL du fichier config.txt

  qui permettrait :

  • d’être générique
  • de gagner pas mal d’octets dans ce qui est envoyé en entête HTTP

• Stop encoding a video from piped video and directshow audio data

  9 avril 2023, par obscure

  Via node.js I'm spawning a FFmpeg process which generates a video using jpeg images received from a node.js stream. Additionally it records audio from a directshow device.

  


  const child = require("child_process");
const stream = require("stream");
let inputStream = new stream.PassThrough();
let ffmpeg = child.spawn("ffmpeg.exe", [
    "-f", "image2pipe", "-r", "24", "-fflags", "nobuffer",
    "-i", "pipe:0",
    "-f", "dshow",
    "-sample_rate",
    "48000",
    "-i", 'audio="Microphone"',
    "-acodec",
    "pcm_s16le",
    "-ac", "2",
    "-ar", "48000",
    "-vcodec", "libx264",
    "-crf", "10",
    "-preset", "veryfast",
    "-framerate", "1",
    "output.mkv"
]);

inputStream.pipe(ffmpeg.stdin);


  


  The problem is, if I later stop sending images down the pipe

  


  inputStream.end();


  


  to ultimately stop FFmpeg from encoding the video, it does not stop - I guess because it's still grabbing audio data from the directshow device.

  


  After digging through the FFmpeg manual I've found the -shortest output options which should :

  


  


  ...Finish encoding when the shortest output stream ends...

  


  


  So upon adding the option e.g.

  


  ...
    "-preset", "veryfast",
    "-framerate", "1",
    "-shortest"
    "output.mkv"
]);


  


  it indeed stops encoding but it's happening too fast.

  


  So a video worth 10 seconds of data just has 5 seconds of audio and from second 5 - 10 there's just silence. I assume it's happening because it doesn't drain the rest of the audio buffer before stopping.

  


  How can I stop encoding the output from the video and the audio data, while keeping all the data up to the point I've triggered the 'stop' ?

  


• avcodec/mediacodecdec : refactor to take advantage of new decoding api

  16 février 2018, par Aman Gupta

  avcodec/mediacodecdec : refactor to take advantage of new decoding api
  This refactor splits up the main mediacodec decode loop into two
  
  send/receive helpers, which are then used to rewrite the receive_frame
  
  callback and take full advantage of the new decoding api. Since we
  
  can now request packets on demand with ff_decode_get_packet(), the
  
  fifo buffer is no longer necessary and has been removed.
  This change was motivated by behavior observed on certain Android TV
  
  devices, featuring hardware mpeg2/h264 decoders which also deinterlace
  
  content (to produce multiple frames per field). Previously, this code
  
  caused buffering issues because queueInputBuffer() was always invoked
  
  before each dequeueOutputBuffer(), even though twice as many output
  
  buffers were being generated.
  With this patch, the decoder will always attempt to drain new frames
  
  first before sending more data into the underlying codec.
  Signed-off-by : Matthieu Bouron <matthieu.bouron@gmail.com>
  

  • [DH] libavcodec/mediacodecdec.c
  • [DH] libavcodec/mediacodecdec_common.c
  • [DH] libavcodec/mediacodecdec_common.h