When it Comes to 视频质量 Measurements, Average Won't Cut It

Most video-quality metric tools score a video file by computing a score for all frames and then averaging the score. Like the old saw about being comfortable because you have one foot in the oven and the other in an ice bucket, 平均分数可以 be deceiving, mostly because they may hide transient quality issues that can degrade your viewer's quality of experience (QoE). Fortunately, there are several mechanisms to avoid this trap, though they tend to be tool-specific.

例如, SSIMWAVE是SSIMPLUS VOD Monitor产生两个 SSIMPLUS 分数,一个 computed 使用 a simple average and another 使用 an algorithm called the weighted average index, or 围. 这是一个广告 SSIMWAVE 白皮书:“感知视频QoE是 not a static process and the impact of different parts of the video to the overall QoE are different. 人类倾向于记住相对低质量的时刻. 围 … is designed to capture the influence of quality fluctuation to overall QoE.“有什么了不起的围就是你可以 control how much weight to assign these quality fluctuations via a simple numerical control.

Moscow State University (密歇根州立大学) adopted a similar approach with its recent version 12 release of the 视频质量测量工具(VQMT)，它会产生平均分 for each video file and the harmonic mean. Here's an explanation from my contact at 密西根州立大学: "For metrics where higher scores are better, 如 SSIM, PSNR值 [峰值信噪比]，和 VMAF [Video Multimethod Assessment Fusion], viewers may be unsatisfied if 15 一段50秒的视频中有几秒 PSNR值价值 20和35秒 PSNR值值为50. 的 arithmetic mean will be (50*35 + 20*15)/50 = 41 这很好. 的 harmonic mean will be 50/(35/50 + 15/20) = 34.5，也就是接近 reality, because low 价值s have higher priority.“所以，这是相同的原则 SSIMWAVE的围, but there's no flexibility because you can't control how much weight to give the variations. 公式就是公式.

如果你计算一下 VMAF 使用 FFmpeg，你可以在上面学习流媒体学习中心r,你可以选择在 three pooling methods—min, harmonic mean, or mean. 自 VMAF 过滤器可以产生 PSNR值值, SSIM, SSIM女士分数与 VMAF，你可以计算 harmonic mean score for all four metrics. I checked and didn't find the same option in the Netflix vmafossexec.exe工具虽然我可能弄错了. Of course, you could always import the individual frame scores into Google Sheets or Excel and compute the harmonic 意思是，但那需要很多工作.

Another measure of variability is the standard deviation. 那么，如果两个视频文件发布了 VMAF 90分, but one had a standard deviation of 3 and the other 7, you'd k现在 that the latter file had more variability. Hybrik的媒体分析仪是第一种工具 I used that included the standard deviation plus the 价值 and location of the highest and 文件中最低的分数. 毫不奇怪, 除了加上调和均值，密歇根州立大学现在 computes the standard deviation and adds the low and high frame 价值 and location to all 的第12版中的度量计算 VQMT.

请注意，SSIMPLUS VOD监视器， VQMT, 和Hybrik都有结果图 graphically display the metric score over the duration of the file, so you can eyeball the variability yourself, which is useful in a one-off situation but hard to build into an automated 工作流. 对于前两款(但不包括Hybrik)， you can also click the graph and view the actual frames to verify that the low score reflects the 实际主观表象. 这是一个伟大的 feature when comparing encoding techniques or codecs, though again, it's not automatable.

的 bottom line is that if you're working with a simple average of the frame scores, you're missing variations that could impact QoE, 这就是为什么我们在首先. So, find a tool that gives you some of the additional datapoints listed above.