With the increasing demand for streaming media, it is important to study the effects of quality defects in a video. We concentrated on how buffers and interrupts can impact a user's quality of experience by measuring how a user becomes annoyed with their presence. We determined annoyance by having the user watch a series of videos with different buffer times or interrupt count, asking for the perceived annoyance afterward. The results from 37 users show a polynomial relationship between buffer time and annoyance and a logarithmic relationship between interrupt count and annoyance. We were able to separate our videos into different levels of motion, and, while we found no relationship between annoyance caused by buffers and the motion level, we found an unexplained relationship between annoyance caused by interrupts and the motion level; the lowest motion level seemed to produce the highest annoyance rating. These results can be used by streaming services to understand how buffers and interrupts affect users, which can help find a better balance that minimizes annoyance.
Download:
![[WPI Homepage]](/images/seal.gif){kind=small}
![[Mark Claypool's Homepage]](http://www.cs.wpi.edu/~claypool/gif/mark_thumb.gif){kind=small}