I am also thinking from that side indeed: if super resolution for video becomes commonplace, then I can well imagine that video platforms will serve lower resolutions than was possible, because the end user will boost it on their side anyway. That saves the platform holder a little bit of computing power and especially bandwidth. But as you point out: these algorithms are not a cat piss; this can cost significant power at the receiving end. And that again per viewer, which is enormously inefficient compared to the one-off performance of the video on the supply side. That is always an option because either there is already a source video in the desired resolution or the provider can apply super resolution itself. That would be a one-off and not per end user again.
Another side to this: the biggest benefit seems to me to be bandwidth savings, for the publisher and the receiver. But for recipients that would be desirable at most on mobile, because we pay per GB there. But this feature is not available on the smartphone. And when it does, it probably just drains the battery. So where I would still like to save bandwidth (on the road), it doesn’t help quickly either. It mainly helps to reduce bandwidth costs for the provider at the expense of end users.
At home I would rather just receive a video in the desired resolution, without burdening my system with super resolution + the associated electricity costs. Let’s hope that video platforms do not embrace these techniques too much, because then we will lose out on balance.
[Reactie gewijzigd door geert1 op 6 maart 2023 12:41]
