Topaz Video Upscaler

Topaz video upscaling on fal.ai with explicit enhancement-model selection, output FPS control, and post-run billing correction.

Model Overview

Topaz Video Upscaler restores and enlarges an existing video while preserving motion and structure. This Nodespell node keeps the input surface clean while still exposing the Topaz-specific controls that materially affect output quality and billing.

Best At

• Increasing resolution on existing footage while keeping the original shot intact.
• Choosing different Topaz restoration models for cleaner, noisier, or lower-quality source material.
• Optional frame-rate conversion up to 60 FPS alongside the upscale.

Limitations / Not Good At

• Pricing depends on output resolution tier, output FPS, and the selected model tier, so larger clips can get expensive quickly.
• This is not a generative video model; it improves and enlarges supplied footage rather than inventing new scenes.
• The default estimate is only a starting point and can still be corrected after the run from fal's billed amount.

Ideal Use Cases

• Upscaling archived clips, trailers, or marketing footage for higher-resolution delivery.
• Restoring compressed or noisy source material with a more suitable Topaz enhancement model.
• Simple graph workflows that need one video input and one improved video output.

Input & Output Format

• Input: required video_url, plus optional model, upscale_factor, target_fps, restoration sliders, and H264_output.
• Output: upscaled video returned on response.

Performance Notes

• fal currently prices this model per output second, with separate tiers for up to 720p, 720p to 1080p, and above 1080p output.
• 60 FPS output doubles the base price, while Gaia 2 costs half of the normal rate.
• Nodespell starts from measured source-video stats, then corrects the final credits after completion using fal billing events when that hidden endpoint is available.