Model Details
Enhance your low-resolution or low-framerate footage with professional-grade video upscaling and frame interpolation. This model utilizes advanced enhancement technology to intelligently upscale videos up to 4x their original size while recovering lost details, reducing compression noise, and dramatically improving overall visual clarity.
Whether you need to revive legacy archival footage, convert standard 1080p high-definition videos into crisp 4K, or smooth out choppy motion, this video-to-video upscaler delivers reliable, high-fidelity results. You can easily supply your video through our app interface, run the enhancement process directly in the UI, and download the resulting high-resolution output directly to your device.
### Key Capabilities - **Resolution Upscaling**: Multiply your video dimensions by setting an `upscale_factor` ranging from 1.0 to 4.0 (defaults to 2.0). - **Frame Interpolation**: Seamlessly increase the framerate of your video by setting a `target_fps` (between 16 and 60). The model intelligently generates accurate intermediate frames, resulting in smooth, lifelike motion. - **Codec Control**: By default, the output utilizes the highly efficient H.265 codec for optimal quality-to-size ratio. You can toggle `H264_output` to true if you require broader playback compatibility for older systems.
### How to Get the Best Results - **Source Material**: While the model excels at enhancement, providing the highest possible quality source file yields the best results. Avoid highly compressed files, as extreme artifacts may be sharpened. - **Target Framerates**: Only use frame interpolation (`target_fps`) if your original video suffers from noticeable stuttering or if you need to conform a cinematic framerate (e.g., 24fps) to a higher broadcast standard (e.g., 60fps). - **Scaling Factors**: A 2x upscale is typically the sweet spot for balancing processing speed and visual improvement, but pushing it to 4x is excellent for older 480p or 720p content.
### Example Usage
To run this model via the ModelRunner JavaScript client, use the following code:
```javascript import { modelrunner } from "@modelrunner/client";
const result = await modelrunner.subscribe("topazlabs/upscale/video", { input: { "video_url": "https://example.com/your-source-video.mp4", "upscale_factor": 2.0, "target_fps": 60, "H264_output": false } });
console.log(result.data); ```
