Recherche avancée

Sur d’autres sites (9502)

• Using avconv (ffmpeg) to concatenate a bunch of .bmps into a mkv/avi video [migrated]

  8 juillet 2012, par user1509246

  Hoi,

  Trying to figure out how to get avconv to concatenate a bunch of .bmps together into a video file.

  Here's what I've got so far :

  avconv -f image2 -i Capture/%d.bmp -vcodec mpeg4 -r 24 -b:v 20M Capture.mkv

  While this does work, the quality is terrible - there are tons of artifacts that are visible, the colours are distorted and everything is blurred.

  I've trawled through the documentation for avconv and ffmpeg, but can't find anything that increases the quality.

  Any ideas as to how I can get the quality as close to the original bmps as possible ?

  Thanks for lending me your brains,
   Alex

• Will ffmpeg transcoding change the media's duration ?

  6 juillet 2012, par MaiTiano

  For example, I have a example flv media : 1775818137_1775828481_10344.flv. Only audio data has contained in it. (No video data)

  The major properties of this flv are :

  _____________________________________________________________________
  
  Format                                   : Flash Video
  
  
  
  File size                                : 87.4 KiB
  
  
  
  Duration                                 : 494 hours
  
  
  
  Overall bit rate mode                    : CBR
  
  
  
  Overall bit rate                         : 0 bps
  
  
  
  Writing application                      : Lavf53.4.0
  
  
  
  Audio
  
  
  
  Format                                   : MPEG Audio
  
  
  
  Format version                           : Version 1
  
  
  
  Format profile                           : Layer 3
  
  
  
  Mode                                     : Joint stereo
  
  
  
  Mode extension                           : MS Stereo
  
  
  
  Codec ID                                 : 2
  
  
  
  Codec ID/Hint                            : MP3
  
  
  
  Duration                                 : 49hours
  
  
  
  Bit rate mode                            : CBR
  
  
  
  Bit rate                                 : 64.0 Kbps
  
  
  
  Channel(s)                               : 2 channels
  
  
  
  Sampling rate                            : 44.1 KHz
  
  
  
  Compression mode                         : Lossy
  
  
  
  Stream size                              : 13.3 GiB
  
  
  
  _____________________________________________________________________

  Because this is the flv piece generated from the Suse10.0 server, the timestamp may be not correctly wrote into the flv metadata. Therefore, the Duration item listed above is not the real time of it.
  I has record the time in its file name.
  So, the beginning time is 1775818137 ms
  the ending time of this flv is 1775828481 ms
  The actual duation of this flv is 10344 ms which is about 10 seconds.

  **HERE IS MY QUESTION -.- **

  When I use ffmpeg to transcode flv fromat into ts format, like this,

  ./ffmpeg -i 1775818137_1775828481_10344.flv -f mpegts -vn -acodec libfaac -ar 44100 -ab 48k 1775818137_1775828481_10344.ts

  Is there possibility that the duration of final gotten ts file is not equal to the duation of original flv file ? In other words, the ts file's duration is not 10344ms.

• FFmpeg seeking in Mpeg4 streams

  4 juillet 2012, par Chris Robinson

  I am currently attempting to develop a player that can perform accurate seeking based on an mpeg4 elementary video stream. I'm in the planning stage and trying to decide how to go about things and I'd like some advice before I start.
  Some things to note are :

  • I will have complete control over the encoding of the file.
  • The original content will be I-frame only
  • FFmpeg is the encoding/decoding library
  • Audio can be disregarded for now. I will only be dealing with the video stream.
  • Frame accurate seeking must be implemented

  So, when I'm encoding the content, can I query what type of frame (I, P, B) has been encoded so I can construct an additional index stream for the seeking operation. If not, I can query the GOP after it has been encoded to find the I-frame.

  As for playback, the user needs to be able to type in a specific time and go to that frame (the nearest I-frame will be suitable for now). We can assume that the GOP is closed and the length is fairly short (e.g. 15 frames). My thoughts are to query the index stream that I created during encode and determine the relevant distance into the stream for the requested time.

  I'm not sure how to seek using the FFMpeg library when playing back files.

  Has anyone done anything similar and if so, can you give a brief explanation of how you did it ?