Recherche avancée

Sur d’autres sites (5965)

• VBV-CBR and ABR in X264

  11 décembre 2013, par xkfz007

  In x264 VBV-CBR RC mode is based on ABR and vbv-maxrate and vbv-bufsize are set. On the basis of frame QP decided by ABR, the MB row QP is decided by VBV-CBR for each row. I test these two modes and the command lines are listed :
  ABR : -o BasketballDrill_832x480_50_20131210102507.264 BasketballDrill_832x480_50.yuv 832x480 —fps 50 —frames 300 —bitrate 1120 —profile baseline —threads -1 -I 50 -v —psnr
  VBV-CBR : -o BasketballDrill_832x480_50_20131210102507_mb.264 BasketballDrill_832x480_50.yuv 832x480 —fps 50 —frames 300 —bitrate 1120 —vbv-maxrate 1120 —vbv-bufsize 1680 —profile baseline —threads 1 -I 50 -v —psnr

  The final results are :
  ABR : Y-PSNR 31.530dB, bitrate 1146.93kbps
  VBV-CBR : Y-PSNR 31.728dB, bitrate 1230.25kbps
  Compared to ABR, VBV-CBR achieved 4.9% bitrate promotion and 0.198dB PSNR promotion. But the PSNR promotion is lower than bitrate (1% bitrate=0.05dB).

  Having talked so much, I just want to ask what is the advantage of VBV-CBR compared ABR ? If ABR can get the reasonable results, why the VBV-cBR is needed ?
  In my opinion, VBV-CBR maybe can promote the subjective quality owing to deciding the QP for each MB row. But the bitrate also rises. Also I have compared some frames of the two modes, they are almost the same.
  Maybe there are some situations that are fit for VBV-CBR and I don't know. So guys, give me, the new to x264, some professional explanation.

• GDPR : How to fill in the information asset register when using Matomo ?

  4 avril 2018, par InnoCraft

  Disclaimer : this blog post has been written by digital analysts, not lawyers. The purpose of this article is to explain you in details how we filled in the information asset register for Matomo. This work comes from our interpretation of the UK privacy commission resources (ICO). It cannot be considered as professional legal advice. So as GDPR, this information is subject to change.

  The information asset register is for us one of the most important parts of the GDPR implementation process. It consists of an inventory of all information systems you are using to process personal data, exactly like a ledger for an accountant. Note that small and medium-sized organizations could be exempted.

  Filling out this register can be a time-consuming activity. Therefore, we decided to show you a real case sample which we did for Matomo Analytics

• FFmpeg : Cut from live stream RTSP

  10 septembre 2020, par Jax2171

  I cannot solve my problem and I have not found any solution.

  


  I capture an rtsp stream from an ip camera via the command :

  


  ffmpeg -stimeout 2000000 -abort_on empty_output -rtsp_transport tcp -i rtsp://admin:admin@192.168.1.15/av0_0 -c:v copy -c:a aac -f mpegts -y rec.ts


  


  The native video codec is h264 while the audio one is pcm. I capture the video stream without compression and compress the audio in aac to make it compatible with the ts container.

  


  I need to extract portions of video of X duration through during the capture. The command I use is :

  


  ffmpeg -ss X-i rec.ts -c:v copy -c:a copy -t Y -f mp4 -stimeout 60 -y cut.mp4


  


  Everything seems to work fine with no errors. However, the cut file has a slightly longer audio length than the video length. Sometimes the difference can be up to 1 second.

  


  My problem is that I have to chain all the cuts I make, and this difference in duration causes a very annoying lag between one video and another.

  


  It also happens that there is asynchronousness between audio and video in the cuts. The higher the -ss value, the more the audio anticipates the video.

  


  I state that I am not a professional but from what I have read it could be a problem of different PTS between audio and video, but I'm not sure. However, I am sure that the problem arises from the capture of the stream perhaps not in accordance with my needs.

  


  How can I solve ?