Veo-3 Gets a BIG Upgrade & Moonvalley First Look!

Theoretically Media
8 Jul 202515:15

TLDRGoogle's Veo 3 has received a significant upgrade, introducing image to video with audio capabilities. The new feature allows users to create videos from images, with impressive results, especially for '90s film-style content. The update also includes a frames-to-video option, available in both standard and fast modes, with the latter being more cost-effective. Additionally, the video explores Moon Valley's new video model, which focuses on clean, non-copyrighted data. Moon Valley offers features like video-to-video, trajectory control, and shot extensions, aimed at professional use. The presenter shares tips on optimizing V3 usage and highlights Moon Valley's potential for filmmakers.

Takeaways

  • 🚀 Google's V3 has received a significant update, introducing an image-to-video feature with audio capabilities.
  • 🎥 The new feature allows users to create videos from images, with options to set first and last frames, and it works in both V3 Standard and V3 Fast modes.
  • 💰 V3 Fast is a more cost-effective option, costing only 20 credits per generation compared to 100 credits for V3 Standard.
  • 🌟 The script demonstrates impressive character animations and voice consistency, though there are still some limitations in action sequences.
  • 💡 A useful tip is to plan ahead and combine multiple lines of dialogue in a single generation to save credits.
  • 🎨 The script highlights the ability to retexture and stylize images using tools like Midjourney for different looks.
  • 🎉 Moon Valley, an AI video generator aimed at professionals, has released a new model with a focus on clean, non-copyrighted training data.
  • 🎥 Moon Valley offers features like shot extensions, camera control, and trajectory control, allowing for more creative video generation.
  • 🎬 The video-to-video feature in Moon Valley shows promise, with the ability to restyle existing video content and control character movements.
  • 📈 Pro subscribers on the $20 plan can now top off their credits, making it more affordable to use V3's features.
  • 🔗 The script encourages further exploration of Moon Valley's features and suggests it could be a valuable tool for professional use.

Q & A

  • What is the main feature of Google's V3 update?

    -The main feature of Google's Veo 3 update is the introduction of image-to-video with audio capabilities, allowing users to create videos from images with synchronized audio.

  • How does the 'frames to video' feature in V3 compare to previous versions?

    -The 'frames to video' feature in V3 is an improvement over previous versions, which were more hit-or-miss. It now offers better quality and more reliable results.

  • What is the difference between V3 Fast and V3 Standard?

    -V3 Fast is a cheaper version that costs 20 credits per generation, while V3 Standard costs 100 credits per generation. The differences in quality between the two modes are fairly negligible, making V3 Fast a more cost-effective option.

  • What is the significance of the '90s film look' mentioned in the script?

    -The '90s film look' refers to a specific visual style that the midjourney images were designed to emulate. It adds a nostalgic and stylistic element to the generated video content.

  • Why did V3 refuse to generate certain images?

    -V3 refused to generate images that closely resembled celebrities, such as Tom Cruise and Daniel Craig, to avoid issues with celebrity likenesses.

  • What is Moon Valley, and what is its unique approach?

    -Moon Valley is an AI video generator aimed at professionals and Hollywood. Its unique approach is to use a clean model with no copyrighted material in its training data.

  • What are some of the new features launched by Moon Valley?

    -Moon Valley has launched several new features, including shot extensions, camera control from image, video trajectory control, and the ability to keyframe up to 128 frames.

  • How does Moon Valley's video-to-video feature perform?

    -Moon Valley's video-to-video feature shows promising results, especially in restyling existing video content. However, there are still some issues, such as inaccurate lip sync.

  • What is the advantage of using Moon Valley for professional usage?

    -Moon Valley's advantage for professional usage lies in its clean model, high-quality output at 1080p and 16-bit, and advanced features like trajectory control and extensive keyframing capabilities.

  • What is the 'Starship Troopers test' mentioned in the script?

    -The 'Starship Troopers test' is a test run on the channel using a scene from the 1990s CGI television show 'Starship Troopers Rough Necks.' It is used to evaluate the performance of Moon Valley's video-to-video feature.

  • What is the pricing update for V3?

    -V3 now allows pro subscribers on the $20 plan to top off their credits, which was previously only available on the $250 Ultra plan. This allows users to save money by purchasing additional credits at a lower cost.

Outlines

00:00

😀 Google V3 Update and Image-to-Video Features

This paragraph discusses the latest update to Google's V3, which introduces an image-to-video feature with audio capabilities. The author highlights the improvements over previous versions, noting that the new feature is more reliable and offers better quality. They demonstrate the feature using a short piece they created, showcasing the ability to generate video from images with a specific aesthetic, such as a '90s film look. The paragraph also covers the challenges faced, like issues with celebrity likenesses, and the cost implications of using different modes (standard vs. fast). The author provides tips on how to optimize the use of this feature, including planning to get multiple outputs from a single generation to save costs. Additionally, they mention a fix for a glitch related to refunding credits and compare the quality of outputs from V3 fast and V3 standard, concluding that the differences are negligible, making the cheaper fast mode a good option.

05:01

😎 Character Work and Pricing Updates in V3

The focus of this paragraph is on the character work and voice consistency in V3. The author praises the improved character animations and voice cadences, demonstrating with examples like a flamethrower girl and an orc barbarian queen. They note that consistent voices across different generations are achievable, though there are still some limitations, such as occasional changes in voice for certain characters. The paragraph also covers the use of a prompt builder tool to enhance the quality of dialogue and character descriptions. The author provides a pricing update, highlighting the new ability for pro subscribers on the $20 plan to purchase additional credits, which was previously only available on the more expensive $250 Ultra plan. This change allows users to save money while still accessing a significant number of generations per month.

10:02

🚀 Moon Valley's New Video Model and Features

This paragraph introduces Moon Valley's new video model, emphasizing its clean model approach, which avoids copyrighted material in its training data. The author reviews the model's capabilities, noting that while text-to-video is currently a weaker aspect, image-to-video and video-to-video features show promise. They demonstrate the model's ability to generate high-quality videos from images and its impressive video-to-video features, such as restyling a scene from the 1990s CGI television show Starship Troopers. The paragraph also highlights Moon Valley's advanced keyframe control, allowing users to keyframe up to 128 frames, and its trajectory control feature, which offers significant control over camera movements and character actions. The author expresses excitement about the potential of these features for professional use and mentions plans to explore them further.

15:03

🎉 Conclusion and Future Plans

In this final paragraph, the author wraps up the discussion by acknowledging that there is still much more to explore and cover. They announce plans to do a rapid-fire review on Thursday to delve into additional features and updates. The author thanks the audience for watching and signs off, expressing enthusiasm for the ongoing developments in AI video generation and the potential of tools like Moon Valley and Google V3.

Mindmap

Keywords

💡Veo-3

Veo-3 refers to an AI tool or platform that has recently undergone a significant upgrade. In the context of the video, Veo-3 is the central focus, particularly its new feature of converting images to videos with audio. This upgrade is a major step forward, as it allows users to create more dynamic and engaging content compared to previous versions that had limitations in image-to-video conversion. The speaker demonstrates how Veo-3 can now generate coherent video sequences from still images, complete with synchronized audio, which is a game-changer for content creation.

💡Image to video with audio

This is a feature of Veo-3 that enables the conversion of static images into animated videos with accompanying audio. It is a key innovation discussed in the video, allowing users to create more immersive and dynamic content. For example, the speaker shows how three midjourney images were used to create a short video sequence with dialogue, demonstrating the potential for storytelling and content creation. This feature is particularly useful for generating videos quickly and efficiently, especially for those who may not have access to traditional video production tools.

💡Midjourney

Midjourney is an AI-based platform mentioned in the script, known for generating images based on text prompts. It is used in conjunction with Veo-3 to enhance the visual quality of the generated videos. The speaker describes how midjourney images were utilized to create a '90s film look in the video sequences, highlighting its role in achieving specific visual styles. This integration shows how different AI tools can be combined to achieve better results, leveraging midjourney's strengths in image generation to complement Veo-3's video capabilities.

💡Frames to video

Frames to video is a specific function within Veo-3 that allows users to convert a series of images (or frames) into a continuous video sequence. This is a core feature discussed in the video, as it is a significant improvement over previous versions of Veo-3. The speaker explains that users can now easily upload images or generate them within the platform and then use frames to video to create a smooth, animated sequence. This feature is particularly useful for creating quick and engaging content without the need for traditional video editing skills.

💡V3 Fast

V3 Fast is a mode within Veo-3 that offers a more cost-effective option for generating videos. It is highlighted in the video as a cheaper alternative to the standard V3 mode, costing only 20 credits per generation compared to 100 credits. The speaker notes that the differences in quality between V3 Fast and the standard mode are negligible, making it a more economical choice for users who want to create multiple video generations without spending too much. This mode is particularly beneficial for those on a budget or who need to generate a large volume of content.

💡Character work

Character work refers to the ability of Veo-3 to generate and animate characters with distinct personalities, voices, and movements. The speaker praises Veo-3's character work, noting how the platform can create convincing character performances, such as the 'flamethrower girl' example. This is a crucial aspect of the video's theme, as it demonstrates Veo-3's potential for creating engaging and believable characters, which is essential for storytelling and content creation. The speaker also mentions how consistent voices can be achieved across different generations, further enhancing the character work.

💡Pricing update

The pricing update is a significant piece of information discussed in the video, particularly regarding the changes in credit top-up options for pro subscribers. Previously, only users on the $250 Ultra plan could top up their credits, but now pro subscribers on the $20 plan can also do so. The speaker explains that this change allows users to save money by purchasing additional credits at a lower cost, making the platform more accessible and cost-effective. This update is important for users who want to maximize their use of Veo-3 without incurring high costs.

💡Moonvalley

Moonvalley is a new AI video generator mentioned in the video, aimed at professionals and the entertainment industry. It is described as having a 'clean model' with no copyrighted material in its training data, making it suitable for professional use. The speaker provides a first look at Moonvalley's capabilities, including its text-to-video and video-to-video features. Moonvalley's ability to generate high-quality videos with control over camera angles and keyframes is highlighted, demonstrating its potential for professional filmmaking and content creation.

💡Video-to-video features

Video-to-video features refer to the capabilities of Moonvalley to take existing video content and transform or restyle it based on user prompts. This is a powerful tool discussed in the video, as it allows users to manipulate and enhance existing footage. For example, the speaker demonstrates how a scene from the 1990s CGI television show Starship Troopers was restyled using Moonvalley. This feature is particularly exciting for filmmakers and content creators who want to experiment with different visual styles or update old footage without starting from scratch.

💡Keyframes

Keyframes are specific points in a video sequence that define the start and end of an animation or transformation. In the context of Moonvalley, keyframes are used to control and manipulate video content with precision. The speaker mentions that Moonvalley allows users to keyframe every frame up to 128 frames, which is an impressive level of control. This feature is crucial for creating complex animations and smooth transitions, as it enables users to define exactly how each frame should look and behave. The example given in the script shows how keyframes were used to create a walking animation cycle, demonstrating the potential for detailed character animation.

Highlights

Google's V3 gets a big update with image to video with audio.

V3's new image to video feature is more reliable than previous options.

V3 now offers 'frames to video' in both standard and fast modes.

V3 fast mode is cheaper at 20 credits per generation compared to 100 for standard.

V3's character work and voice consistency are impressive.

V3 still struggles with certain sequences, like kung fu scenes.

Moon Valley releases a new video model with a focus on clean, non-copyrighted data.

Moon Valley's text to video is currently its weakest feature.

Moon Valley's image to video feature looks decent and supports high resolution.

Moon Valley introduces shot extensions and trajectory control for video editing.

Moon Valley can generate videos in ultra widescreen from 16:9 images.

Moon Valley's video to video features allow for restyling scenes with impressive results.

Moon Valley supports key framing up to 128 frames for detailed control.

Moon Valley integrates well with motion generation platforms like Deep Motion.

Pricing update for V3: Pro subscribers can now top off credits, saving money.

Moon Valley is a powerful tool for professionals needing a clean model.