Runway Gen-3 Video AI - In depth Test and Review

Olivio Sarikas
4 Jul 202418:36

TLDRThe Runway Gen-3 Video AI is tested in this review, showcasing its capabilities and limitations. It excels in creating detailed, cinematic landscapes and drone flight scenes, particularly with fire and smoke effects. However, it struggles with complex human movements and close-up details. The alpha version offers limited settings but promises more features post-testing. Pricing is on the higher side, with an unlimited plan for heavy users. The potential for unique content creation is highlighted, despite the current imperfections.

Takeaways

  • πŸ†• Runway Gen-3 is an alpha version video AI model currently accessible to paid customers only.
  • πŸ” The alpha version offers limited settings compared to the previous CH 2 model, including no camera control or style selection.
  • πŸŽ₯ Gen-3 excels at creating timelapse videos and landscape scenes, especially with drone flight perspectives.
  • 🚫 The model struggles with complex human movements, such as fast limb movements, often resulting in unrealistic animations.
  • πŸ’¬ A unique feature of Gen-3 is its ability to perform lip sync, which was demonstrated in a test video.
  • 🎨 The model shows potential in creating artistic and cinematic scenes, with good consistency in character details.
  • πŸ”₯ Gen-3 handles fire and smoke effects well, making it suitable for creating dramatic and cinematic shots.
  • 🎡 The model has issues with body movement, often resulting in morphing or inconsistent anatomy, especially in dance and gymnastics.
  • 🌟 It delivers a dreamlike quality in its videos, which some may see as a feature rather than a flaw, adding to the unique appeal.
  • πŸ’° The pricing for Runway's service seems high, especially considering the alpha status and limitations of the Gen-3 model.
  • πŸ“ˆ There's potential for Gen-3 to become a powerful tool once the alpha testing phase is over and more features are enabled.

Q & A

  • What is the current status of Runway Gen-3 Video AI?

    -Runway Gen-3 Video AI is currently in its alpha version and is available for use by paid customers only.

  • What are some limitations of the Runway Gen-3 Alpha version compared to the previous model?

    -The Runway Gen-3 Alpha version has fewer settings compared to the previous model, such as limited resolution options (only 720p), fewer settings for prompt length, and no advanced camera controls or motion brush features.

  • What is one feature that Runway Gen-3 Video AI can do exceptionally well?

    -Runway Gen-3 Video AI does exceptionally well with timelapse videos, especially with landscapes and drone flying scenes.

  • Can Runway Gen-3 Video AI perform lip sync?

    -Yes, Runway Gen-3 Video AI can perform lip sync, and the reviewer demonstrated this with a video example.

  • What are some areas where Runway Gen-3 Video AI struggles?

    -Runway Gen-3 Video AI struggles with complex human motion, especially when arms and legs are moving quickly, and it can sometimes produce unrealistic or morphing movements.

  • What is the reviewer's opinion on the consistency of the AI-generated videos?

    -The reviewer appreciates the consistency of the AI-generated videos, noting that objects and scenes stay in place and do not flicker or morph excessively, which is impressive for an AI model.

  • What kind of scenes does the reviewer find particularly stunning in the Gen-3 Model?

    -The reviewer finds landscape shots, drone flights, and scenes with fire and smoke particularly stunning in the Gen-3 Model.

  • How does the reviewer feel about the potential of Runway Gen-3 Video AI for creative projects?

    -The reviewer sees huge potential and creative opportunities with Runway Gen-3 Video AI, especially for those who have a plan for using the generated videos, like the creator of the Sora music video.

  • What are the pricing plans for Runway Gen-3 Video AI?

    -There is a standard plan offering 625 credits per month, an unlimited plan for $76 per month, and a pro version with 2250 credits per month.

  • What is the reviewer's final verdict on Runway Gen-3 Video AI after testing?

    -The reviewer finds Runway Gen-3 Video AI capable of creating beautiful videos, especially for certain types of scenes, but notes that it makes too many mistakes for them to consider paying for it at this stage.

Outlines

00:00

πŸš€ Introduction to Runway Gen 3 Alpha Version

The speaker introduces the new Runway Gen 3 model, an alpha version currently accessible only to paid customers. They discuss their extensive testing of the model and provide an overview of the website's features, which are limited in the alpha version compared to the previous Gen 2 model. The speaker highlights the ability to input prompts, adjust video length, and select settings such as seat type and watermark removal. They also mention the lack of high-resolution options and the absence of advanced settings found in Gen 2, suggesting that these may be added post-testing.

05:01

πŸŒ„ Gen 3's Strengths in Landscape and Timelapse Videos

The speaker praises Gen 3's capabilities in creating landscape and timelapse videos, noting the model's consistency and detail. They share examples of drone flight through various scenes, such as a neon city and a tropical rainforest, and commend the model's ability to handle complex transitions and maintain video quality. However, they also point out issues with certain elements, like reflections in mirrors and the accuracy of celestial bodies, while acknowledging the overall impressive results.

10:03

🎭 Mixed Results in Animation and Body Movement

The speaker explores Gen 3's performance in animating human figures, noting mixed results. They discuss attempts at creating a spaghetti-eating animation and a gymnast's exercise, highlighting the model's struggles with complex movements and anatomical accuracy. Despite these challenges, they find some success in animations like a violin player and a drummer, where the model captures the essence of the action despite minor inconsistencies. The speaker appreciates the model's ability to create a dreamlike, artistic quality in its animations.

15:03

πŸ’Έ Pricing and Potential of Gen 3 for Creative Projects

The speaker discusses the pricing plans for Runway Gen 3, mentioning the limited credits in the standard plan and the higher cost of the unlimited plan. They consider the potential of the model for creative projects, drawing a comparison with the success of the Sora music video. While acknowledging the current limitations and errors in the alpha version, they express optimism about the model's future capabilities and its potential to revolutionize video creation. The speaker concludes with a call to action for viewers to like and subscribe for more content.

Mindmap

Keywords

πŸ’‘Runway Gen-3 Video AI

Runway Gen-3 Video AI refers to the third generation of video artificial intelligence technology developed by Runway. It is an advanced tool that uses AI to generate videos based on user prompts. In the context of the video, the reviewer is testing the capabilities of this new model, exploring its strengths in creating various video scenes and pointing out areas where it falls short.

πŸ’‘Alpha version

An 'Alpha version' is a term used in software development to describe a product that is still in the early stages of testing and not yet ready for public release. The video script mentions that the Runway Gen-3 is an alpha version, indicating that it is currently available only to paid customers for testing purposes and may contain bugs or limited features.

πŸ’‘Prompt

In the context of AI video generation, a 'prompt' is a text input provided by the user that guides the AI in creating a specific scene or video. The script describes how the user can enter a prompt and adjust settings to influence the output of the Runway Gen-3 AI.

πŸ’‘Lip sync

Lip sync, short for lip synchronization, is the process of matching an actor's lip movements to pre-recorded or dubbed speech or song. The video script highlights a feature of the Runway Gen-3 AI that allows for lip-syncing in generated videos, showcasing the AI's ability to create more realistic and dynamic content.

πŸ’‘Timelapse

A 'timelapse' is a video technique where time is condensed, showing hours, days, or even years of action in a short video. The reviewer notes that the Runway Gen-3 AI excels at creating timelapse videos, particularly with landscapes and slow-moving scenes, demonstrating the AI's capability to handle complex visual transformations.

πŸ’‘Drone flight

A 'drone flight' refers to the movement of a camera as if it were mounted on a drone, providing aerial or bird's-eye views. The script mentions that the Runway Gen-3 AI does particularly well with drone flight simulations, creating immersive and visually appealing scenes.

πŸ’‘Harry Potter effect

The 'Harry Potter effect' in the video script refers to a magical or fantastical scene reminiscent of the Harry Potter series, such as a small entrance leading to a large, expansive space. The reviewer tests the AI's ability to create such imaginative scenes, which is a testament to the AI's creative potential.

πŸ’‘Anatomy

In the context of the video, 'anatomy' refers to the accurate representation of the human body's structure in the generated videos. The reviewer points out that the Runway Gen-3 AI sometimes struggles with accurate anatomy, particularly in scenes with complex human movements.

πŸ’‘Cinematic light

Cinematic light refers to the use of lighting in a way that enhances the visual storytelling, creating mood and atmosphere. The script praises the Runway Gen-3 AI for its ability to create cinematic light effects, contributing to the overall aesthetic and emotional impact of the generated videos.

πŸ’‘Text effects

Text effects in video production involve the use of visual styles and animations to make text stand out. The reviewer is impressed by the Runway Gen-3 AI's ability to create consistent and visually appealing text effects, adding another layer of creativity to the generated content.

πŸ’‘Dreamlike quality

The 'dreamlike quality' mentioned in the script refers to the surreal and fantastical nature of the scenes generated by the Runway Gen-3 AI. Despite some inaccuracies, the AI's output has a unique, otherworldly appeal that the reviewer finds enjoyable and engaging.

Highlights

Runway Gen-3 Video AI is currently in alpha version and accessible only to paid customers.

The alpha version allows users to enter prompts and adjust video length and seat settings.

720p is the only resolution available for Gen-3, which is lower than other Runway models.

Compared to Gen-2 (Chen 2 model), Gen-3 has fewer settings and capabilities.

Gen-3 excels in creating timelapse videos and drone flight scenes.

The model has issues with complex human movements, especially when limbs move quickly.

Lip-syncing is a cool feature that Gen-3 can perform, demonstrated in a video example.

The model shows impressive consistency in scenes with fire and smoke.

Gen-3 struggles with anatomical accuracy in certain body movements, like gymnastics.

The model produces cinematic lighting and detailed close-ups of faces.

Text effects in videos are consistent and create visually appealing results.

The model has a dreamlike quality, which can be seen as a feature rather than a flaw.

Pricing for the standard version starts at $76 per month with 625 credits included.

An unlimited plan is available for $76 per month, similar to Mid Journey's offering.

The Pro version with 2250 credits per month is more expensive due to the high GPU processing time.

The reviewer believes the potential for creating unique content with Gen-3 is huge, as seen with the Sora music video.

The reviewer suggests that the current pricing is reasonable for those with a clear plan for using the AI-generated videos.