Text this: S2A: Scale-Attention-Aware Networks for Video Super-Resolution