Recherche avancée

Sur d’autres sites (6709)

• FFmpeg remux rtp to mpegts [on hold]

  16 décembre 2013, par Ardoramor

  I am trying to remux rtp stream into mptegts format. I have an SDP file with the following contents :

  v=0
  
  o=- 0 0 IN IP4 127.0.0.1
  
  s=Unnamed
  
  i=N/A
  
  c=IN IP4 192.168.17.44
  
  t=0 0
  
  a=recvonly
  
  a=orient:portrait
  
  m=video 8202 RTP/AVP 96
  
  a=rtpmap:96 H264/90000
  
  a=fmtp:96 packetization-mode=1;profile-level-id=428028;sprop-parameter-sets=Z0KAKJWgKA9E,aM48gA==;
  
  a=control:trackID=1

  I execute the following ffmpeg command :

  ffmpeg -i "test.sdp" -f mpegts -vcodec copy "/tmp/test.ts"

  And I get the following information :

  Input #0, sdp, from 'test.sdp':
  
    Metadata:
  
      title           : Unnamed
  
      comment         : N/A
  
    Duration: N/A, start: 0.066622, bitrate: N/A
  
      Stream #0.0: Video: h264 (Baseline), yuv420p, 640x480, 90k tbr, 90k tbn, 180k tbc
  
  [mpegts @ 0x1101d4c0] muxrate VBR, pcr every 9000 pkts, sdt every 200, pat/pmt every 40 pkts
  
  Output #0, mpegts, to '/tmp/test.ts':
  
    Metadata:
  
      title           : Unnamed
  
      comment         : N/A
  
      encoder         : Lavf53.4.0
  
      Stream #0.0: Video: libx264, yuv420p, 640x480, q=2-31, 90k tbn, 90k tbc
  
  Stream mapping:
  
    Stream #0.0 -> #0.0

  I receive the following error :

  [mpegts @ 0x1c85f960] h264 bitstream malformated, no startcode found, use -vbsf h264_mp4toannexb
  
  av_interleaved_write_frame(): Operation not permitted

  So I add the suggested bitstream filter :

  ffmpeg -i "test.sdp" -f mpegts -vbsf h264_mp4toannexb "/tmp/test.ts"

  But the h264 encoding now becomes h262 (mpeg2video) :

  ~$ffprobe /tmp/test.ts
  
  Input #0, mpegts, from '/tmp/test.ts':
  
    Duration: 00:00:04.13, start: 1.400000, bitrate: 640 kb/s
  
    Program 1
  
      Metadata:
  
        service_name    : Unnamed
  
        service_provider: FFmpeg
  
      Stream #0.0[0x100]: Video: mpeg2video (Main), yuv420p, 640x480 [PAR 1:1 DAR 4:3], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 120 tbc

  Is there any way to keep the h264 codec without re-encoding it ? Doing so becomes very CPU intensive.

  Update

  Hopefully this will clear up the issue and remove the off-topic stamp.

  I'm writing an Android app that is based off of SpyDroids streaming architecture. The app communicates with the server, providing it the SDP. The server spawns an ffmpeg process to remux the incoming video stream into mpegts and broadcasts it on multicast (right now just file).

  SpyDroid performs streaming by sending recorded mp4 file through localsocket, received h264 packets, supposedly (according to code removed mp4 h264 prefix [annexb]), wraps it with rtp headrs and sends it on its way. Thus, the RPT stream I get is clearly not originally generated as such.

  As @Wagner Patriota has mentioned, I should add '-vcodec copy'. I had run the remuxing with it before as well but the error is still present (full output) :

  ~$ffmpeg -i "test.sdp" -f mpegts -vcodec copy -vbsf h264_mp4toannexb "/tmp/test.ts"
  
  ffmpeg version 0.8.6, Copyright (c) 2000-2011 the FFmpeg developers
  
    built on Jan 30 2012 17:17:54 with gcc 4.5.2
  
    configuration: --prefix=/usr --libdir=/usr/lib64 --shlibdir=/usr/lib64 --mandir=/usr/share/man --enable-shared --enable-gpl --enable-version3 --enable-nonfree --enable-postproc --enable-avfilter --enable-pthreads --enable-x11grab --disable-avisynth --enable-libdc1394 --enable-libfaac --enable-libgsm --enable-libmp3lame --enable-libx264 --enable-libxvid --extra-cflags='-O2 -g -m64 -mtune=generic -fPIC' --disable-stripping --disable-demuxer=v4l --disable-demuxer=v4l2 --disable-indev=v4l --disable-indev=v4l2
  
    libavutil    51.  9. 1 / 51.  9. 1
  
    libavcodec   53.  7. 0 / 53.  7. 0
  
    libavformat  53.  4. 0 / 53.  4. 0
  
    libavdevice  53.  1. 1 / 53.  1. 1
  
    libavfilter   2. 23. 0 /  2. 23. 0
  
    libswscale    2.  0. 0 /  2.  0. 0
  
    libpostproc  51.  2. 0 / 51.  2. 0
  
  [h264 @ 0x16b4b1c0] concealing 232 DC, 232 AC, 232 MV errors
  
  [h264 @ 0x16b4b1c0] concealing 63 DC, 63 AC, 63 MV errors
  
  [h264 @ 0x16b4b1c0] concealing 25 DC, 25 AC, 25 MV errors
  
  [h264 @ 0x16b4b1c0] concealing 138 DC, 138 AC, 138 MV errors
  
  [h264 @ 0x16b4b1c0] concealing 69 DC, 69 AC, 69 MV errors
  
  [sdp @ 0x16b43400] Estimating duration from bitrate, this may be inaccurate
  
  
  
  Seems stream 0 codec frame rate differs from container frame rate: 180000.00 (180000/1) -> 90000.00 (180000/2)
  
  Input #0, sdp, from 'test.sdp':
  
    Metadata:
  
      title           : Unnamed
  
      comment         : N/A
  
    Duration: N/A, start: 0.033256, bitrate: N/A
  
      Stream #0.0: Video: h264 (Baseline), yuv420p, 640x480, 90k tbr, 90k tbn, 180k tbc
  
  [mpegts @ 0x16b4a4c0] muxrate VBR, pcr every 9000 pkts, sdt every 200, pat/pmt every 40 pkts
  
  Output #0, mpegts, to '/tmp/test.ts':
  
    Metadata:
  
      title           : Unnamed
  
      comment         : N/A
  
      encoder         : Lavf53.4.0
  
      Stream #0.0: Video: libx264, yuv420p, 640x480, q=2-31, 90k tbn, 90k tbc
  
  Stream mapping:
  
    Stream #0.0 -> #0.0
  
  Press [q] to stop, [?] for help
  
  h264_mp4toannexb failed for stream 0, codec copy: Invalid argument
  
  [mpegts @ 0x16b4a4c0] h264 bitstream malformated, no startcode found, use -vbsf h264_mp4toannexb
  
  av_interleaved_write_frame(): Operation not permitted

  The error reports that the invalid argument has been supplied. Increased loglevel does not give any more information. I know that ffmpeg is sometimes finicky with argument order. However, they seen to be in order of documentation as well as suggested order by @Wagner Patriota.

• Presentation of Piwik’s collaborative translations platform : oTrance [Interview]

  19 avril 2013, par matt — Community, translation

  Piwik enables domain administrators, hobbyists, power users, personal website builders and everyone in between to access enormous amounts of data for website analytics. To support all those users, Piwik needs to be available in a number of different languages. From the start, we made internationalization (i18n) part of Piwik’s DNA. There are now dozens active volunteers who help make sure each language is well represented in the latest official release of Piwik. As of now, Piwik is available in 48 languages.

  Recently a new tool became available that makes the translation of Piwik much easier. The software we are using is an open source platform called oTrance. It has made our translation architecture more robust, and it allows us to expedite the timely delivery of high quality and up-to-date translations to the thousands of people who rely on Piwik every day.

  We’ve met with oTrance creator and lead developer Daniel Schlichtholz who answered a few questions for us.

  1. What is oTrance ?
  2. What sets oTrance apart from the other ways to manage translations ?
  3. Now that oTranCe 1.0 is out, what will you be working on next ?

  What is oTrance ?

  oTranCe is the short form of “Online Translation Center”. It was born because I needed a translation platform for my project MySQLDumper.

  Many languages have been added by the community and manual maintenance became more and more time consuming. I wanted to change that. So I searched for an existing platform I could use and tested a lot of approaches. To put a long story short : none of the given solutions satisfied my needs.

  From the view of a translator maintaining a language should be as easy as possible. In most cases they have to install a program on their local machine or the workflow was too difficult. A translator doesn’t want to struggle with technical things ; he just wants to translate the phrases and wants to know the progress.

  That’s the main goal we want to reach : to make the translation process as easy as possible.

  What sets oTrance apart from the other ways to manage translations ?

  Ease of use is one advantage of oTranCe compared to other solutions. Another advantage is that project administrators can install oTranCe on their own server – so nobody is dependant of a third party provider.

  We love to get feedback from other users. User feedback influences the way oTranCe is developed. We believe that this way oTranCe satisfies the requirements of the real world.

  We also have extensive user documentation, in our “Working with oTranCe” wiki. We try to document use cases in an understandable way. We don’t write down marketing buzz words, but try to explain the use from the view of the user/administrator.

  Now that oTranCe 1.0 is out, what will you be working on next ?

  The language files can be exported to version control and oTranCe can commit changes to the target repository. Currently we support export to Subversion, and we are working on a Git export adapter, which will be released soon.

  Another issue we are trying to solve is the context problem. When your project uses many different phrases the translator often doesn’t know in which context the current phrase is used. Version 1.1.0 (not released yet, but you can grab the latest developer version from GitHub) introduces the oTranCe-connector. The idea behind it : a small plug in grabs the used phrases/keys on the current page, and on click this list is submitted to oTranCe, where the translator can edit the words. This way the translator knows in which context these phrases are used. I wrote a small plug in for OXID eShop. Since it is really easy to implement, my hope is that other plug ins for other applications will be added by the community.

  Matthieu : Congratulations Daniel for having created such an awesome Translation Platform. At Piwik we are really thankful for oTranCe, which has resulted in much better translation process, and happier translators. Keep up the good work !

  If you are a Piwik user, and if you want to participate in translating Piwik, please sign up for an account on oTrance and become part of the team making Piwik available in more languages across the world.

• configure : enable ffnvcodec, nvenc, nvdec for aarch64

  23 avril 2020, par ManojGuptaBonda

  configure : enable ffnvcodec, nvenc, nvdec for aarch64
  Adding the support to build FFMPEG with HW accelerated decode(nvdec) and
  
  encode on aarch64 architecture.
  Signed-off-by : Timo Rothenpieler <timo@rothenpieler.org>
  

  • [DH] configure