Recherche avancée

Sur d’autres sites (11220)

• Troll spirit

  15 juin 2013, par Mans — Law and liberty

  Last week’s announcements from the White House of steps being taken to begin fighting back against patent trolls, along with legislation passed in Vermont for the same purpose, are worthy of praise. Whether they prove effective or not, they are a sign of the problem finally having been recognised by the highest authorities. That said, only one aspect of the issue is addressed, that of non-practising entities or trolls. Little effort is being made to stymie troll-like behaviour from otherwise legitimate actors. While a stake is driven through the heart of the troll, its spirit remains free to roam the corporate world, and like a demon of darkness it possesses companies, compelling them to engage in the very practices we seek to eradicate.

  The demon

  The most damaging, when wielded by a troll, are those patents with vague or overly broad claims. These can easily be asserted against large numbers of alleged infringers, many of which likely choose to settle out of court rather than risk an expensive litigation process with uncertain outcome. Such negotiations are frequently subject to non-disclosure agreements prohibiting publication of details in any deals, or even the existence thereof. As a result, an accused has no way of assessing a fair price for a licence (assuming the patent is in fact valid), and the patent holder can thus extract from each would-be infringer precisely as much as they are willing or able to pay to avoid a lawsuit.

  At the root of this problem is the ease with which applications for the patents in question are granted. Given the volume of patent applications, it is hardly reasonable to demand a hugely more extensive examination process than currently takes place (although some improvements here are no doubt possible) ; after all, a speedy decision is in the best interest of all parties. The solution must evidently be found elsewhere.

  The exorcism

  An obvious cure to the problem is the abolishment of the patent system. As this is clearly not feasible today, more practical, albeit less effective, remedies must be sought. A few ideas follow.

  Make patent validity all or nothing

  Change the rules such that any claim being found invalid cancels the patent its entirety. With the full patent at stake in this manner, companies would be discouraged from gambling on frivolous claims and encouraged to conduct a more thorough background investigation before filing.

  Maintain a registry of licences

  Require that all patent licence agreements be filed in an open, easily searchable registry. This would hopefully increase fairness in licensing deals.

  Mandate reimbursement of licence fees for invalidated patents

  If a patent is challenged and found invalid, require that the owner reimburse any licence fees previously collected for the patent in question. Apart from being morally right, this could act as a deterrent to over-charging. The amount requested for a licence would likely be balanced against the risk of being made to pay it all back later, resulting in lower licence fees for low-confidence patents.

  These suggestions, alone or together, will not completely eradicate the problems of patent abuse. They are but small steps towards a more thorough overhaul of a system increasingly ill-suited to the nature and pace of modern technological development.

• Does framerate influence compressing videos on ffmpeg ?

  27 novembre 2020, par caiomarqs

  When I try to compress a video on ffmpeg with two pass, I see there is some disparity between the number I targeted in the command line and the output's final video bitrate. I'd like to know if there are some other factors that I'm not considering. Maybe framerate ?

  


  Example :

  


  Windows command :

  


  ffmpeg -y -i input.mp4 -c:v libx264 -b:v 313k -pass 1 -an -f null NUL && ^ 
ffmpeg -i input -c:v libx264 -b:v 313k -pass 2 -c:a aac -b:a 96k output.mp4


  


  Linux command :

  


  ffmpeg -y -i input.mp4 -c:v libx264 -b:v 313k -pass 1 -an -f null /dev/null && \
ffmpeg -i input -c:v libx264 -b:v 313k -pass 2 -c:a aac -b:a 96k output.mp4


  


  As you can see below, although my video bitrate input was 313kbps, the output's video bitrate was slightly higher :

  


  print of output's data

  


  What exactly is it ?

  


• checkasm : h264dsp : Move the x and y variables into the randomize_buffer macro

  27 juin 2016, par Martin Storsjö

  checkasm : h264dsp : Move the x and y variables into the randomize_buffer macro
  This avoids the risk of accidentally clobbering such variables outside
  
  of the macro if the same variables are used there.
  Signed-off-by : Martin Storsjö <martin@martin.st>
  

  • [DBH] tests/checkasm/h264dsp.c