0
0
Kling AI Video Generator Challenges OpenAI’s Sora
OpenAI’s upcoming Sora AI video generation tool has sparked both excitement and concern among creatives. However, a new contender, Kling, developed by Kuaishou, a Chinese tech giant and TikTok competitor, is generating buzz in the AI community.
Kling boasts several key advantages over its competitors. It can produce videos up to two minutes long in 1080p resolution at 30 frames per second, a significant improvement over Pika and RunwayML. This outperforms what OpenAI claimed Sora would achieve.
Kling vs. Vidu: The Chinese AI Video Battle
Kling’s launch follows Vidu, another Chinese AI video tool created by Shengshu Technology and Tsinghua University. While Vidu specializes in shorter videos, it excels in creating realistic content using Universal Vision Transformer (U-ViT) technology, surpassing Runway and Pika.
On the other hand, Kling utilizes advanced 3D face and body reconstruction to generate lifelike movements based on a single full-body image. It addresses common issues in AI-generated videos, such as distorted limbs.
Realism and Physics: Kling’s Strong Points
Similar to Sora, Kling excels in simulating real-world physics for accurate object interactions. It can also follow prompts effectively, creating multi-view shot sequences for dynamic videos.
Kuaishou showcased Kling’s capabilities through demo videos, demonstrating superior realism and consistency compared to Runway and Pika, even rivaling Sora in some aspects.
Global Release and Accessibility
Although currently available as a public demo in China with a waitlist, Kuaishou plans a broader global release for Kling. To access the demo, users must register with a Chinese phone number on the Mandarin-language Kling website and follow steps on the official site.
For those eager to explore AI video generation now, options like Pika Labs and RunwayML, or locally Stable Video Diffusion, remain viable choices.
Image/Photo credit: source url
