Upscaling videos using artificial intelligence (AI) can elevate the quality of your content significantly. In this detailed guide, we’ll walk you through the process of free AI video upscaling which is cross-platform compatible for Windows, Mac, and Linux. You’ll need to ensure your computer can run the Vulkan graphics system. This step-by-step tutorial will help you leverage the power of Waifu2x, an open-source project, to upscale your videos. All the website links mentioned are provided below.
First off, navigate to the official Waifu2x project page. Familiarise yourself with the project details and functionalities. For this guide, we will use the Waifu2x-ncnn Vulkan project from GitHub. This specific project offers cross-platform compatibility.
For Windows:
For Linux:
sudo apt install ffmpeg
Windows users should download, extract, and copy the ffmpeg
and ffprobe
executables to the Waifu2x directory.
Create the following three directories within your main Waifu2x directory:
video
: to store your input and output videos.frames
: for storing extracted video frames.scaled
: for storing upscaled video frames.You can use this command to navigate within your terminal:
cd path/to/waifu2x_folder
Before diving into the AI processing, we must extract frames from your video file. Use this command in the terminal:
ffmpeg -i video/video.mp4 frames/frame%06d.png
This command will extract all frames and save them in the frames
directory as PNG files.
To keep the audio from the original video:
ffmpeg -i video/video.mp4 -q:a 0 -map a video/audio.mp3
Apply the Waifu2x upscaling to the extracted frames:
waifu2x-ncnn-vulkan.exe -i frames -o scaled -n 3 -s 2
In this command:
-i frames
: specifies the input directory.-o scaled
: specifies the output directory.-n 3
: sets the noise level, adjust this value as desired.-s 2
: sets the scaling factor. Here, it's set to 2x.Reassemble the upscaled frames into a video and add the extracted audio back in:
ffmpeg -r 29.97 -i scaled/frame%06d.png -i video/audio.mp3 -c:v libx264 -pix_fmt yuv420p -c:a aac video/upscaled.mp4
Here:
-r 29.97
: sets the frame rate, make sure it matches the original video's frame rate.-i scaled/frame%06d.png
: specifies the input directory of the upscaled frames.-i video/audio.mp3
: adds the extracted audio back into the video.-c:v libx264 -pix_fmt yuv420p -c:a aac
: specifies the video and audio codecs.You've successfully upscaled your video using AI! This process can be adjusted and optimized to fit various needs. Experiment with different noise levels, scaling factors, and frame rates to get the best results for your projects.
Q1: What software do I need for AI video upscaling?
Q2: How do I install ffmpeg on Linux?
sudo apt install ffmpeg
.Q3: Can I change the scaling factor in Waifu2x?
-s
parameter in the command.Q4: Is it necessary to extract audio from the video?
Q5: What if my frame rate is different from the examples provided?
-r
) in the commands to match your original video’s frame rate.Q6: Can I use a different noise reduction parameter?
-n 1
, -n 2
, etc.) to achieve different results.By following these detailed steps, you can achieve high-quality AI upscaling for your videos on multiple platforms. Happy upscaling!
In addition to the incredible tools mentioned above, for those looking to elevate their video creation process even further, Topview.ai stands out as a revolutionary online AI video editor.
TopView.ai provides two powerful tools to help you make ads video in one click.
Materials to Video: you can upload your raw footage or pictures, TopView.ai will edit video based on media you uploaded for you.
Link to Video: you can paste an E-Commerce product link, TopView.ai will generate a video for you.