Recherche avancée

Sur d’autres sites (9313)

• Piwik 2.10.0 – Release Candidate

  22 décembre 2014, par Piwik Core Team — Community

  We are proud to announce that the release candidate for Piwik 2.10.0 is now available !

  How do I upgrade to the release candidate ?

  You can upgrade to the release candidate in one click, by following instructions in this FAQ.

  Think you’ve found a bug ?

  Please create a bug report in our issue tracker.

  What’s new in Piwik 2.10.0 ?

  Since our last release Piwik 2.9.1 one month ago, over 100 issues have been closed. We’ve focused on fixing bugs, improving performance, and we created a new plugin that will let you better scale Piwik to very high traffic websites using Redis.

  Much improved Log Analytics

  Log Analytics is the powerful little-known feature of Piwik that lets you import dozens of different server logs into Piwik. In Piwik 2.10.0 you can now import Netscaler logs, IIS Advanced Logging Module logs, W3C extended logs and AWS CloudFront logs. Piwik will also automatically track the username as the User ID and/or the Page Generation Time when it is found in the server logs.

  Better scalability using Redis (advanced users)

  At Piwik PRO we are working on making Piwik scale when tracking millions of requests per month. In this release we have revamped the Tracking API. By using the new QueuedTracking plugin you can now queue tracking requests in a Redis database, which lets you scale the Piwik tracking service. The plugin is included as Free/libre software in the core Piwik platform. More information in the QueuedTracking user guide.

  Better performance

  A few performance challenges have been fixed in this release.

  The Visitor Log and the Live API will render much faster on very high traffic websites. Any custom date ranges that you have selected as default in your User Settings (eg. ‘Last 7 days’ or ‘Previous 30 days’) will now be pre-processed so that your analytics dashboard will always load quickly.

  For users on shared hosting, the real time widgets could be use a lot of server resource as they are refreshed every ten seconds. We’ve improved this by only requesting data when the Browser Tab containing the Real time widgets is active.

  Other changes

  We packed in many other changes in this release such as compatibility with Mysql 5.6 and Geo location support for IPv6 addresses. A community member made Piwik compatible with Internet Explorer 9 when running in compatibility mode (which is still used in several companies).

  The Tracker algorithm has been updated : when an existing visit uses a new Campaign then it will force creating a new visit (same behavior as Google Analytics).

  If you need professional support or guidance, get in touch with Piwik PRO.

  Changelog for Piwik 2.10.0 – we plan to release Piwik 2.10.0 around 2015 Jan 5th.

  Happy Analytics, and we wish you a nice holiday season !

• FFMpeg is writing a matplotlib animation with diminishing quality

  3 mars 2017, par Bryan Stafford

  I am trying to create a bunch of animations that follow the paths National Hockey League teams travel during the current season. Currently, my animation looks like this.

  As you can see, the animation begins decently sharp but within 2 seconds the quality bottoms out and everything is grainy.

  The relevant code is this :

  ani = animation.FuncAnimation(fig, update, frames=len(all_x + 20), 
  
                                interval=75, repeat=False)
  
  filepath =  map_team + '.mp4'
  
  ani.save(filepath, writer='ffmpeg', fps=30)
  
  plt.tight_layout()
  
  plt.show()

  I have chosen FFMpeg as my writer because I hope to put about 30 graphics online and the MP4 format keeps each file around 1.5MB. If there’s a way to keep the graphics sharp, with low file sizes, and able to be uploaded onto Imgur, I would love to do that with my project.

  Thanks.

• HEVC File bigger after converting from h264

  26 janvier 2019, par Aaroknight

  I’m currently working an an automated Python script for indexing and converting all my movies and episodes with ffmpeg. I use subprocess.call() for running the ffmpeg command and tested this command with some movies. As expected the big h264 files were converted to merely one third of what they used to have.

  But now that I was testing the method I found that a converted episode (about 400MB in h264) had over 1,6GB in hevc. I have absolutely no idea why the new file would be that much bigger in hevc.
  This is my code :

  def convert(path):
  
      outvid = path.strip(".mkv") + "h265.mkv"
  
  
  
      cmd = ["ffmpeg", "-i", path, "-map", "0", "-map_metadata", "0", "-map_chapters", "0", "-c:v", "libx265",
  
             "-c:a", "copy", "-c:s", "copy", "-preset", "ultrafast", "-x265-params", "lossless=1", outvid]
  
      subprocess.call(cmd)
  
  
  
  convert("/Volumes/2TB/Black Butler/Season 1/Black Butler S01E01.mkv")

  I don’t have that much experience with ffmpeg, nor with subprocess. This is one of my first bigger projects. I hope someone can tell me what the problem might be.

  UPDATE
  Problem applies only for small video files. I now just check for the file size and skip the small files. Wouldn’t have made much of a difference anyway.