Topview Logo
  • Create viral videos with
    GPT-4o + Ads library
    Use GPT-4o to edit video empowered by Youtube & Tiktok & Facebook ads library. Turns your links or media assets into viral videos in one click.
    Try it free
    gpt video

    Amazing AI Filmmaking is Here: Gen-2 The Ultimate Cinematic AI (Full Tutorial)

    blog thumbnail

    Amazing AI Filmmaking is Here: Gen-2 The Ultimate Cinematic AI (Full Tutorial)

    Last week, I released a video demonstrating how to transform mid-journey or AI-generated images into cinematic presentation videos. Today, we are taking it a step further and turning our images into actual movies with Gen-2 by RunwayML. Brace yourself for this wild ride as we explore the incredible capabilities of this text-to-video generator.

    In the world of AI, Gen-2 by RunwayML is nothing short of amazing. This tool allows us to generate videos based on a text prompt, and the results are truly mind-blowing. To get started, simply add the Gen-2 bot to your Discord server and type in your desired prompt. For example, I used the prompt "a man on a boat in the ocean near some islands." In just seconds, Gen-2 generated a video of this non-existent scene, complete with a man, a boat, and a stunning ocean view.

    The rendering process is surprisingly quick, with the longest wait time I experienced being around two minutes. The output resolution is set at 768 by 448 in a 16:9 aspect ratio, which currently can't be changed. However, there are commands available to upscale the video to a higher resolution and interpolate the footage to reduce choppiness.

    To demonstrate the capabilities of Gen-2, I generated a video of a busy New York City street using the prompt "New York City street busy people walking." By utilizing the upscale and interpolate commands, I was able to produce a higher resolution video with smoother frames. The result was truly remarkable.

    While Gen-2's output may not be a perfect recreation of the original prompt, it captures the desired tone and aesthetic remarkably well. There are some AI glitches, such as unusual clothing details or awkward hand movements, but these minor issues can easily be overlooked in the grand scheme of what this technology achieves.

    To further enhance the cinematic vision, we can also use reference images as prompts in Gen-2. Uploading your own reference image and adding a text prompt can help refine the output. You can even use a tool called Clip Interrogator to generate a more accurate prompt based on the image. By adjusting the CFG scale command, which determines how much weight the reference image has in the output, you can achieve varying levels of fidelity.

    In a previous tutorial, we explored how to create a Gothic Victorian story using mid-journey. Now, we can take those mid-journey images and use them as reference images in Gen-2. By combining the power of both platforms, we can create in-depth cinematic presentations. In this example, I transformed a sequence of images into a captivating short film, showcasing an alien invasion in New York City.

    The use of reference images with Gen-2 can present some challenges. Hands, for instance, are notoriously difficult for AI to generate accurately. However, with thoughtful adjustments and appropriate prompts, we can achieve impressive results. By storyboarding our ideas and combining Gen-2's output with mid-journey's capabilities, we can create cohesive and visually striking films.

    As an experiment, I wanted to see how Gen-2 handles animation. Inspired by Samurai Jack, I generated a Samurai character and prompted, "Samurai walks, camera autumn forest concept art Samurai Jack." The animated output was not exactly what I had in mind, but it did capture a minimalist Samurai Jack aesthetic that I found appealing. Although Gen-2 currently lacks the ability to animate sword-fighting sequences, we can still create engaging visual narratives by combining its output with other tools.

    Keywords:

    AI-generated videos, Gen-2, cinematic presentation, text-to-video generator, Discord-based, AI glitches, reference images, Clip Interrogator, CGF scale, mid-journey, Gothic Victorian, short film, animation, Samurai Jack aesthetic.

    FAQ:

    1. How can I access Gen-2?

      • Join the Gen-2 Discord community and wait for invitations to be sent out at regular intervals.
    2. Can I change the resolution and aspect ratio of the generated videos?

      • Currently, the resolution is fixed at 768 by 448 in a 16:9 aspect ratio. However, there are commands available to upscale the video and improve resolution.
    3. How accurate is Gen-2 in generating videos based on text prompts and reference images?

      • While Gen-2 may not produce a perfect recreation of the original prompt, it captures the desired tone and aesthetic remarkably well.
    4. Can I combine outputs from Gen-2 with other AI tools to enhance the cinematic presentation?

      • Absolutely! Combining Gen-2's output with tools like mid-journey can result in more cohesive and visually striking films.
    5. Does Gen-2 support animation?

      • Yes, Gen-2 does support animation. However, certain actions like sword-fighting may not be accurately generated. Currently, combining Gen-2 with other tools may be necessary for such sequences.

    One more thing

    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.

    You may also like