So recently I joined somebody's Patreon and he posts upscaled anime episodes and I really like it. Was thinking of dabbing in it and buying an upscaler program myself to try to upscale 480p-720p videos to 1080p-2k. Does anybody have any suggestions on what to buy? So far, I'm, thinking of Gigapixel AI or DVDFab Enlarger AI. I don't really know the difference in them, but any tip, suggestions, and recommendation would be nice.
+ Reply to Thread
Results 1 to 7 of 7
-
-
You can try my free Up-Scale.
See here for more info and download:
https://forum.videohelp.com/threads/396482-up-scale-a-new-small-and-simple-Upscaler -
for Vapoursynth:
Waifu2x (https://github.com/Nlzy/vapoursynth-waifu2x-ncnn-vulkan)
Anime4KCPP (https://github.com/TianZerL/Anime4KCPP)
might be interessting.users currently on my ignore list: deadrats, Stears555, marcorocchini -
VSGAN and Digiframes. There are many upscaling models available on https://upscale.wiki/wiki/Model_Database#Anime
-
This thread on restoring sailor moon might be helpful https://forum.videohelp.com/threads/403809-Anime-re-mastering-how-to-achieve-a-particu...ok#post2638093
-
There is an option to use Clarity Anime upscaler (https://clarityai.co/anime-image-upscaling) for free, with A1111. Full tutorial:
How to Upscale Images for Free with Clarity Upscaler:
1. Install A1111 WebUI (three options from hardest to easiest):
a) Follow the official tutorial: https://github.com/AUTOMATIC1111/stable-diffusion-webui#installation-and-running
b) Watch the YouTube tutorial: https://youtube.com/watch?v=kqXpAKVQDNU
c) Use One-Click start on
@runpod_io (not free, but easy)
2. Install the extension:
Open A1111 WebUl -> Click Tab "Extensions" -> Click Tab "Available" -> Find "[Tiled Diffusion with Tiled
VAE]" -> Click Install
3. Download the models and place them in the correct folder:
- https://huggingface.co/dantea1118/juggernaut_reborn/resolve/main/juggernaut_reborn.saf...?download=true -> stable-diffusion-webui/models/Stable-diffusion
- https://huggingface.co/philz1337x/upscaler/resolve/main/4x-UltraSharp.pth?download=true -> models/ESRGAN
- https://huggingface.co/philz1337x/embeddings/resolve/main/JuggernautNegative-neg.pt?download=true -> embeddings
- https://huggingface.co/philz1337x/loras/resolve/main/SDXLrender_v2.0.safetensors?download=true -> models/Lora
- https://huggingface.co/philz1337x/loras/resolve/main/more_details.safetensors?download=true -> models/Lora
- https://huggingface.co/lllyasviel/ControlNet-v1-1/resolve/main/control_v11f1e_sd15_til...?download=true -> models/ControlNet
4. Go to the img2img tab and apply the parameters:
Model: juggernaut_reborn,
Prompt: masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>
Negative prompt: (worst quality, low quality, normal quality:2) JuggernautNegative-neg
Image: The image you want to upscale
Sampler: DPM++ 3M SDE Karras,
Steps: 18,
CFG scale: 6.0,
Denoising strength: 0.35,
Tiled Diffusion:
"Enable Tiled Diffusion": true,
"Keep input size": true,
"Method": "MultiDiffusion",
"Tile tile width": 112,
"Tile tile height": 144,
"Tile Overlap": 4,
"Tile batch size": 8,
"Upscaler": 4x-UltraSharp,
"Scale factor": 2,
ControlNet 0:
"Enable": true,
"Low VRAM": false,
"Pixel Perfect": true,
"Control Type": tile
"Preprocessor: tile_resample,
"Model": control_v11f1e_sd15_tile,
"Control weight": 0.6,
"Starting control step": 0.0,
"Ending control step": 1.0,
"Control mode": 1,
5. Upscale:
Click Generate to upscale your image
Similar Threads
-
up-scale, a new small and simple Upscaler
By ProWo in forum Video ConversionReplies: 14Last Post: 30th Nov 2021, 11:12 -
S-VHS + HDMI upscaler + HDMI capture?
By Jupit0r in forum Capturing and VCRReplies: 2Last Post: 22nd Aug 2020, 02:00 -
Best choice of composite upscaler for HD capture
By nsputnik in forum Capturing and VCRReplies: 3Last Post: 16th Apr 2020, 15:32 -
Video upscaler help please
By hdfills in forum Capturing and VCRReplies: 1Last Post: 27th May 2017, 19:46 -
Looking for talented upscaler for commision work!
By focus187 in forum Video ConversionReplies: 4Last Post: 30th May 2016, 18:20