Practical guide
2× or 4× video upscaling? Dimensions, limits, and quality checks
2× doubles each pixel dimension; 4× quadruples each dimension. That means 2× produces four times as many pixels per frame, while 4× produces sixteen times as many—before an app’s output cap applies. Pick the size your viewing or editing task needs, then inspect the result.
By David Balderston, developer of Local Video Upscaler. Updated October 10, 2026. The examples below are resolution arithmetic, not processing-time measurements or evidence of recovered detail.
Worked examples for landscape clips
| Source dimensions | Nominal 2× output | Nominal 4× output | Practical interpretation |
|---|---|---|---|
| 640 × 360 | 1280 × 720 | 2560 × 1440 | 2× reaches a 720p frame; 4× reaches a 1440p frame |
| 1280 × 720 | 2560 × 1440 | 5120 × 2880 | Nominal 4× exceeds a 4K cap; check the actual exported dimensions |
| 1920 × 1080 | 3840 × 2160 | 7680 × 4320 | 2× already reaches a 4K frame; 4× cannot exceed the app’s 4K cap |
These are 16:9 examples. Portrait and other aspect ratios have different dimensions. Local Video Upscaler caps output at 4K and skips inputs already at 4K. A scale selection is not a promise that every source exports at exactly that multiple.
Start with the smallest useful enlargement
If you want a 640 × 360 family clip to look more comfortable on a phone or small display, try 2× first. Compare a short section at the same display size. If the destination needs a larger frame, try 4× and inspect whether it improves the viewing experience enough to justify the larger output and processing work.
For a 1920 × 1080 source, nominal 2× already reaches 3840 × 2160. Selecting 4× does not bypass Local Video Upscaler’s output cap. Larger dimensions do not guarantee a sharper face, readable text, or restored motion detail.
Pixel count is different from processing time
For a source with width W and height H, a nominal 2× frame has (2W) × (2H) = 4WH pixels. A nominal 4× frame has 16WH pixels. Processing time also depends on the model, source length, device, temperature, frame rate, and output cap. The arithmetic does not establish a four-times timing difference between settings.
Test a short clip on your actual device before starting a long export. Keep the device supplied with power when practical, allow ventilation, and leave space for the source and output. No universal minutes-per-minute benchmark is claimed here.
Inspect movement, not just one flattering frame
Pause on a face and inspect edges, eyes, teeth, hair, and text. Then play the same section and look for detail that changes unnaturally from frame to frame. Compare the original and output at the same viewing size. An enlarged frame can look smoother because of how it is displayed, without containing more reliable information.
Try Natural and Clean on the same short source if you are choosing between modes. Record your choice and the app version. The best setting for one clip can be a poor fit for another.
Check the export before archiving it
Local Video Upscaler exports HEVC video with the source frame rate and synchronized audio. HDR sources become SDR. Check dimensions, duration, orientation, playback, audio synchronization, and color in the app you intend to use next. Keep the original and mark the enhanced copy so you can distinguish them later.
See the complete facts and limits, step-by-step old-video workflow, and video-upscaling tool comparison. The Apple listing supplies current price and device requirements.
View Local Video Upscaler on the App Store
Scan with your iPhone