==I am not the creator of this model, not contributor to this model.==
This model is showing great results for batch processing for chaiNNer upscaling (x3 or x4 mode), using this with a conservative unsharp mask you can get SD video frames up to a reasonably good HD video.
(Workflow: Extract the frames, load the image frames, upscale and save w/ sequence no. To reassemble, load the images and load the video, use the images for "Save video" and connect the original video AUDIO connection to Save Video. )