We developed a super-resolution (SR) benchmark to analyze SR's capacity to upscale compressed videos. Our dataset employed video codecs based on five compression standards: H.264, H.265, H.266, AV1, and AVS3. We assessed 17 state-ofthe-art SR models using our benchmark and evaluated their ability to preserve scene context and their susceptibility to compression artifacts. To get an accurate perceptual ranking of SR models, we conducted a crowd-sourced side-by-side comparison of their outputs. The benchmark is publicly available at https://videoprocessing.ai/benchmarks/super-resolutionfor-video-compression.html. We also analyzed benchmark results and developed an objective-quality-assessment metric based on the current bestperforming objective metrics. Our metric outperforms others, according to Spearman correlation with subjective scores for compressed video upscaling. It is publicly available at https://github.com/EvgeneyBogatyrev/super-resolution-metric.

翻译：我们构建了一个超分辨率（SR）基准，用于分析SR在提升压缩视频分辨率方面的能力。本数据集采用了基于五种压缩标准的视频编解码器：H.264、H.265、H.266、AV1和AVS3。我们利用该基准评估了17种最新SR模型，检验了它们在保持场景上下文方面的能力以及对压缩伪影的敏感度。为了获得SR模型的准确感知排序，我们开展了众包并排对比实验，比较了各模型的输出结果。该基准已在https://videoprocessing.ai/benchmarks/super-resolution-for-video-compression.html 公开。我们还分析了基准实验结果，并基于当前性能最佳的客观度量开发了一项目标质量评估指标。根据该指标与压缩视频超分辨率主观得分的斯皮尔曼相关性，其表现优于其他方法。该指标已在https://github.com/EvgeneyBogatyrev/super-resolution-metric 公开。