ExpedusTM DVA Subjective Video Quality Assessment License – MOS Scoring
Efficient Video QoE Quality Assessment
Subjective video quality assessment is a common method for evaluating video processing and services. The traditional Absolute Category Rating (ACR) method requires a large number of viewer surveys to be effective. To address the inefficiencies of the ACR method, the industry has sought to develop algorithms that can achieve measurement metrics similar to ACR results. The MOS-V used by Expedus DVA is one widely applied in the industry. Building on this, Expedus DVA offers the capability for real-time parallel MOS-V testing of hundreds of video streams.
The main reasons for the decline in video quality during the video encoding and transcoding process are:
· Coarse Quantization
· Excessively long GOP structure
· Low frame rate
· Frame information loss
The MOS-V algorithm detects the following information and achieves subjective evaluation results similar to traditional ACR through high-performance computation:
· GOP Structure Analysis: GOP Length, Frame Information Loss Distribution
· Frame-by-frame quality assessment: Frame type, frame size, frame rate, degree of damage to previous I/P frames
· Bandwidth estimation: Program type, quantization level
· Video content: Degree of action change
Unlike the commonly used active video quality detection algorithms in the industry, MOS-V is a passive algorithm. Its advantages include lower computational load, ease of application in operational environments, and the ability to perform parallel detection of multiple video streams. According to experimental data, MOS-V has a certain correspondence with active detection algorithms.
Video resolution is an important factor that leads to differences in subjective quality. To more accurately assess subjective video quality, especially in the context of mobile video applications, MOS-V is subdivided into Absolute MOS-V and Relative MOS-V. The calculation of Absolute MOS-V considers changes in video resolution as a factor contributing to video quality degradation; Relative MOS-V calculates video quality at the same resolution. In low bitrate scenarios, Relative MOS-V aligns more closely with the actual quality perception of users.