Early Alpha Google Colab Notebook
- AnimationKit Colab Notebook for Real-ESRGAN
.
Features Real-ESRGAN video upscaling, RIFE interpolation/motion smoothing, (improvised) deflickering, and FFMPEG hevc_nvenc (h265) compactor
#ALPHA 2: Released 9/4/21
Feature additions
- ✅ Deflickering (P3)
- ✅ Upscaling from individual frames (P2)
- ✅ Target length (in seconds) for RIFE interpolation - Replaces old length_multiplier option
Credits: Motion smoothing conceived from "Zoom animation processing and motion interpolation" added by https://twitter.com/unltd_dream_co. This part of the script uses RIFE real-time video interpolation to smooth out the resulting video.
Upscaling uses Real-ESRGAN (https://github.com/xinntao/Real-ESRGAN). A demo notebook for static images can be found here: https://colab.research.google.com/drive/1k2Zod6kSHEvraybHl50Lys0LerhyTMCo?usp=sharing. The demo was based on the following paper: of our paper ''Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data''.
#Alpha 1 release: All major bugs have (hopefully) been patched. New functions will be released in the testing branch until debugged.
Feature additions & bugfixes:
- ✅ Anime model for realesrgan
- ✅ 2x model
- ✅ target output path
- ✅ omission of outscaling (note, may need to check defaults in inference.py)
- ✅ target fps no longer causes duplicates prior to ffmpeg end-phase (prevents upscaler from wasting gpu cycles on duplicate frames; major speed increase)
- ✅ notebook set to high memory - necessary for RIFE
Planned additions:
- More deflicker options (am and pm deflicker flags may be optimal)
- --skip option for RIFE
- --tiles option for RIFE/Real-ESRGAN (?)
- face detection for real-esrgan
- h264 option (for people with older rigs)
- Preview video option for people who don't want to download the full video yet
Difficulties (may take longer to figure out)
- target output path for RIFE
- target duration for RIFE (need to understand how exp works, then I'll just have it divide)
Feel free to report bugs!