https://flathub.org/apps/io.github.tntwise.REAL-Video-Enhancer
REAL Video Enhancer is a redesigned and enhanced version of the original Rife ESRGAN App for Linux. This program offers convenient access to frame interpolation and upscaling functionalities on Linux, comparable to Flowframes or enhancr on Windows.
- NEW! CUDA support.
- NEW! MacOS support.
- Support for Ubuntu 20.04+ on AppImage and Flatpak.
- Discord RPC support for Discord system package and Discord flatpak.
- Scene change detection to preserve sharp transitions.
- Preview that shows latest frame that has been rendered.
- Enhancing straight from a YouTube link or a video URL.
Benchmarks done with 1920x1080 video, default settings using JPG image extraction.
| RX 6650 XT | Ensemble False | Ensemble True |
|---|---|---|
| rife-v2.0 - v2.4 | 12.341 fps | - |
| rife-v3.0 - v3.1 | 10.646 fps | - |
| rife-v4.0 - v4.5 | 32.504 fps | - |
| rife-v4.6 | 31.154 fps | 18.078 fps |
| rife-v4.7 - v4.9 | 27.924 fps | 15.463 fps |
| rife-v4.10 - v4.15 | 22.801 fps | 12.981 fps |
| rife-v4.16-lite | 31.205 fps | 19.381 fps |
- Visit the WIKI
https://github.com/TNTwise/REAL-Video-Enhancer/releases/tag/prerelease
- FFMpeg
- rife-ncnn-vulkan
- Real-ESRGAN
- yt-dlp
- VapourSynth-RIFE-ncnn-Vulkan (For RIFE models [4.1-4.5],[4.7-4.12-lite])
- ifrnet-ncnn-vulkan
- RIFE
- PySceneDetect
- realcugan-ncnn-vulkan
- REAL-Cugan
- TheAnimeScripter (For CUDA implementation code and TRT upscaling code)
- Spandrel (For CUDA upscaling model arch support)
- RealESRGAN NCNN python
- cx_Freeze
- Chainner's NCNN Implementation (For scale detection with NCNN models)
- vs-rife (For rife TRT inference code)
