Quality Estimation of YouTube Video Service

Detta är en Magister-uppsats från Blekinge Tekniska Högskola/Sektionen för datavetenskap och kommunikation

Sammanfattning: YouTube is today one of the most popular video sharing web sites. It is using Flash Player technology together with progressive download to deliver videos in a variety of qualities adaptable to the clients connection speed and requirements. Recently YouTube has got much more user attention with the introduction of HD (High Definition) video content on its web site. This work started with analyzing different aspects of YouTube videos and finding a way to detect re-buffering events during playback. In an effort to do this, 38 videos were uploaded to YouTube and analyzed with respect to codec, container format, encoded bitrate and resolution. An application called YouTube Player API was used to detect the re-buffering events and some more useful information for the investigation. The greater part of the work concentrated on estimating the effect of re-buffering on the end user perceived quality of YouTube videos. Finally, conclusions were made by presenting a way to estimate the effect of re-buffering on the perceptual quality of YouTube videos and stating that the maximum quality available on YouTube for HD-720P and HD-1080P is approx 3.91 and 3.86 (on a scale from 1 to 5) respectively.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)