The current magic of AI lies in its ability to create, such as writing summaries, novels, code, and music… it can be said to be all-powerful. One of the amazing technologies among them is AI video, which can generate videos from text or images.
Currently, one of the more famous ones is Runway, which is one of the largest artificial intelligence companies in the world, dedicated to developing image and video AI editing software. It offers more than 30 types of AI Magic Tools, real-time video editing, collaboration, and other features, including video generation from video, text to image, image to image, background replacement, green screen, erasing, and replacing, allowing operations on static images or videos. With the support of artificial intelligence, it can provide perfect video effects, format, and color correction. Runway also allows users to train their own AI models, making it easier for the AI to generate images that meet expectations from text prompts.
-
1. Login
1. Visit the website: https://runwayml.com/, click on “Try for free”.
2. Enter the login and registration interface as shown in the image, where you need to register with an email, or you can log in directly with a Google email. After logging in, a usage prompt will pop up, allowing for a free trial or recharge service. The free trial includes 125 credits, 3 video projects, and 3 editors.
2. Interface and Features
1. Main Working Interface: The left side is the command bar, which includes AI creation tools, AI training, and work library.
The right side is for selecting AI creation tools, the operation area, and the work display area.
2. AI Tools: Including video generation from video (first generation), training your own generator, removing video backgrounds, text to image, image to image, extending images, erasing and replacing, infinite backgrounds, and other 30 AI video and image tools. The various AI tools are shown in the image below.
3. Example Test: Video Generation from Video (First Generation)
This time we only test video generation from video. In the main interface, click the “Video to Video” AI tool to enter the upload video interface and add the video to be processed (such as a person dancing on the grass). After adding, enter the video generation interface, where the left side is the video preview area, and the right side is for function selection and prompt input area.
1. Text to Video
To change the dancing video on the grass into a dancing video in the water. On the right side in the “Style reference” section, select the text prompt method (Prompt), and enter the text prompt for modifying the video: Beauty, exquisite face, dance, flower fall, on the water, keeping other options default, as shown in the image above, then click the “Generate” button to get the generated video.
Enjoy the video:
Note: The generated video has consistent actions, the background matches, but the character image effect is blurry.
2. Preset Material Method
Convert real video images into videos made with different materials. In the right side “Style reference” section, select the preset method (Presets), choose the desired material effect, such as clay, watercolor, origami, pen and ink, charcoal sketch, cloudscape, etc., keeping other options default, as shown in the image, then click the “Generate” button to get the generated video.
Enjoy the Video
Cloudscape
Watercolor
Paper origami(折纸)
Pen and ink(钢笔和墨水)
Charcoal sketch(木炭素描)
Claymation（粘土)
Summary:
-
Experience of Runway’s video generation from video function, the generated video does not exceed 3 seconds.
-
Some AI tools require a recharge upgrade to use.
-
The current test is Gen1 (first generation), and the video effects are still not very mature. Gen2 (second generation) has been launched, and the effects of the second generation will be tested later.
-
Runway’s more than 30 AI tools are quite practical.
-
Runway Gen1 (first generation) mainly modifies existing videos, while Gen2 (second generation) will reflect the creation of new videos.
Welcome to technical exchanges.