Tag: OpenAI Sora

  • Shengshu Technology Launches AI Video Generator to Compete with OpenAI

    Shengshu Technology Launches AI Video Generator to Compete with OpenAI

    Shengshu Technology has introduced its newest AI creation, the Vidu 1.5. This company is stepping into the area of text-to-video and image-to-video generation, adding a new player to the mix to rival OpenAI’s Sora, which made its debut earlier this year. For those unfamiliar with Shengshu Technology, it is a budding AI company that started up in March 2023.

    Features and Capabilities

    The new AI model closely resembles OpenAI’s Sora, with a few modifications. The Vidu 1.5 can generate videos that are up to eight seconds long from both images and text prompts. It employs its own feature called multiple-entity consistency to smoothly combine people, objects, and settings when creating a video based on user inputs.

    Unique Functionality

    Another interesting aspect of Vidu is its multiple-angle consistency feature. This allows users to create videos either from any uploaded images or by submitting three photos of the same subject. The company claims that this AI model has advanced control capabilities that enhance motion and detail in the backgrounds of the videos it generates. According to the developers, users can produce an eight-second video in less than 30 seconds.

    Video Quality

    Vidu has showcased several AI-generated videos on its official website, which appear quite realistic. However, the limitation is that users can only create videos of eight seconds, while the Sora model lets users make videos that last up to a minute.

    When it comes to video quality, the Vidu 1.5 can generate videos at a maximum resolution of 1080p. However, it does have some shortcomings. For instance, a few AI-generated videos miss minor details and show some unrealistic actions, like a scene where flames seem to pass through the middle of a car. These are not significant issues, but they might be noticeable if watched closely.

    Pricing and Availability

    The Vidu 1.5 operates on a freemium model. The free version allows users to create a 4-second clip with decent resolution, while the premium version costs $9.99 per month and enables the generation of eight-second videos at 1080p. The Vidu 1.5 is now accessible to everyone through its official website.

    Source: Link


    Image 1
    Image 1
    Image 1
  • MediaTek AI-Powered Video Creation on Dimensity 9300 & 8300

    MediaTek AI-Powered Video Creation on Dimensity 9300 & 8300

    AI is rapidly evolving, enabling the creation of videos from stable images, a process termed image-to-video creation. Utilizing stable diffusion, AI can transform still images into videos. Advanced AI models have now made it possible to generate text-to-video content based on user prompts. For example, OpenAI’s Sora can produce highly realistic videos. For a closer look, check out OpenAI’s Sora and computer-generated videos using the provided link.

    Text to Video – AI Generated Image by OpenAI

    AI-powered video creation is revolutionary, but many existing AI-based media generation tools restrict the number of videos users can create. Due to the substantial processing power required, running these AI models locally on your device might be more efficient.

    MediaTek & Kwai

    A recent ITHome news article suggests that MediaTek may enable on-device AI video creation with its Dimensity 9300 and 8300 chipsets. These chipsets’ NPUs will facilitate video generation from stable images. MediaTek was rumored to introduce this feature at MWC 2024, and now a partnership with Kwai has been unveiled. Kwai, a video-sharing platform and editor akin to TikTok and CapCut, will collaborate with MediaTek.


    MediaTek AI-Powered Video Creation on Dimensity 9300 & 8300