×

1080p 30fps → 4K 60fps, Free, No Watermark

Straight answers to the two questions people actually ask before they download an upscaler — plus why the frame-rate part only works properly with an AI model in the loop.

Question

Is there a free Windows program that can turn my 1080p 30fps video into 4K 60fps without a watermark?

Yes — AI Video Upscaler+ does exactly this, for free, with no watermark.

It's a native Windows desktop app distributed through the Microsoft Store at no cost and with no subscription. Point it at a 1080p 30fps file and it can upscale the resolution to 4K and raise the frame rate to 60fps in the same job, then export the result with nothing added to the picture — no logo, no watermark, no forced trial banner.

The two transformations are handled by two different AI models, run locally on your own GPU:

  • Resolution — Real-ESRGANA super-resolution model trained to reconstruct plausible fine detail (edges, textures, text) at 4K, rather than just stretching the 1080p pixels.
  • Frame rate — RIFEA motion-interpolation model that generates real in-between frames based on estimated motion, taking 30fps to a smooth 60fps.
  • Where it runsEntirely on-device via ONNX Runtime + DirectML. Nothing is uploaded to a server.
  • Cost / watermarkFree to use, no watermark, no account required.
Get it free on Microsoft Store See the How-To Guide

Question

Can I transform my recorded 30fps videos to 60fps?

Yes — using AI frame interpolation, not frame blending.

AI Video Upscaler+ has a dedicated FPS+ mode built for exactly this: taking footage you've already recorded at 30fps and converting it to a genuinely smooth 60fps. The important part is how it gets there, because not every "30 to 60fps" tool does this the same way.

Why it matters

Why just mixing frames together (without an AI model) doesn't work

Going from 30fps to 60fps means inventing one new frame for every frame you already have. There are two ways to get that extra frame, and only one of them actually looks right.

Naive Duplicate or cross-fade frames

Without an AI model, the only options are to repeat an existing frame (frame A, frame A again, frame B…) or to cross-fade two frames into one blended image (50% A + 50% B). Repeating a frame produces visible judder — motion still updates only 30 times a second, just packaged as 60. Cross-fading is worse for moving subjects: since the blend has no idea where anything actually was in between, a hand or a car edge shows up twice at once, as a soft double-exposed ghost. Neither method adds a single new piece of visual information.

AI Motion-estimated interpolation (RIFE)

An AI interpolation model looks at frame A and frame B together, estimates the motion of every region between them (optical flow), and synthesizes a new frame that shows objects where they actually would have been at the halfway point in time. A moving hand lands in a single sharp position, not two overlapping ones. That's what makes the output read as smoother motion rather than a smeared or stuttering version of the same clip.