Upscaling via NVENC takes too long

[SGpro]

I recorded a lot of vertical videos on iPhone; they're in 1920x1080 with a rotation of 90. (Landscape video flipped 90 degrees)
I uploaded these videos to YouTube and intended to keep it there for storage. Still, when 1080p vertical videos are uploaded to YouTube, they're automatically downscaled to 480p, with the exception of the video being 4k.

So I decided to upscale the video to 4K, then uploading them.
I've been using Format Factory to upscale and converting the video to 2160x3840 in HEVC, but there is an apparent degradation in quality.

Knowing that 4K videos on YouTube uses VP9, I decided to give Handbrake a try.
With all the default presets on VP9 2160p, although my video did convert to the right format, it's a lot slower, and I later found out that HandBrake GUI's do not allow upscaling in dimensions, which defeats the whole purpose of converting in the first place.

[Woodstock]

It does not matter what encoding you use to upload - YouTube is going to re-encode it. It is best to minimize the number of times compression is applied. Upscaling is still applying compression after distorting the picture (each pixel becomes 4 in your case), so that defeats wanting to maximize quality. Using NVENC to encode it is using a "less good" encoding to boot...

I have to ask, are you judging what YT is doing to the video over the last month or so, or long term? YT and other video services are currently down-grading what they display by default because of the COVID19 panic, so you might just be seeing that.

[Deleted User 11865]

Your title says NVENC and then your post says VP9… which is it?

[mduell]

None of it makes any sense and there's no log.

[Deleted User 13735]

Please don't upscale -- for the children.