Victoria Beckham’s Post

It's finally here!!!! Runway Image to Video!! Here are some tests & my honest opinion at the end of the thread.

Learn how to seamlessly loop Director Mode videos in today's Runway Academy.

Moodboard teasers are an enticing sneak peek of what’s to come… and most interestingly, a glimpse into the designer’s mind, the inspiration and meaning that underpin a Collection… How about a quick branding exploration… respond below 🤗 What does this piece suggest? What feelings, thoughts, impressions does it generate? How does it define Loewe in one word? #creativeconcepting #Fashionbranding #visualness #brandconnection #designprocess #creativepaths #communicationstrategy

#runwayml Gen2 ups the ante for anyone currently working with #genai, by providing a desperately needed #compositing workflow. Why is this important? Most of the GenAI models have difficulty creating #alphachannels, a lot of the type of visuals many would like to create with GenAI, up until now, was impossible or incredibly laborious. One would have to run a separate passs for depth to do deep compositing, even if it was just to stack a single layer atop another. Or, one would have to run a batch using something like “segment anything,” to get the matte passes, or just outright do it by hand, as #rotoscopy. Yet, even that process is sub-optimal when dealing with hair, smoke, glass, etc., because the alpha channels for a *perfect* alpha channel down to the pixel and opacity levels, including reflections and refractions, wasn’t possible without tons of detailed work that most people don’t have the time, tools, or patience to do. Without compositing, modern VFX as we know them today, would simply not be possible. It’s that simple. Anyone working in #vfx, knows that shots can live and die by how good the #compositor is, since they’re the last person to touch a shot from what is often many dozens or even hundreds of other artists and/or studios. They can sometimes save critical shots that weren’t shot or rendered the proper way, and that could be millions. Compositors take the insanity of dozens, hundreds, to even thousands of layers and #cgi passes, in 2D, 2.5D, and 3D and are the ones who get shots “finaled” to ship. That compositing is now offered as part of the workflow for GenAI means that instead of trying to render the entirety of a scene and try to get everything right in one shot, the first time, or via multiple iterations, means *saving time and money;* basically the same thing. It means fewer tokens can be spent to get certain things right, and the “spare/ofher” tokens on getting something more difficult right, separately. One does this until all the disparate elements are as close, or precise, as the way one wants or needs them to be. Then, these can be composited. This means we’ll start seeing more videos with fewer hallucinations and greater temporal consistency. Individual items can now be rendered at maximum resolution, then scaled to preserve quality, while further enabling shots with parallax and/or dolly shots at the compositing stage, not the latent diffusion stage, which is often difficult to control. It also means that additional details from stock or other content, be they footage or 3D models, can be brought in to guide the GenAI, or added to the shot to aid in creating greater integration of elements, so they appear more cohesive. Basically, this offers a substantial quality jump for GenAI shots, as they now have a chance to be done faster, at greater quality, while also integrating live-action and/or stylized elements, to provide greater depth and detail, without having to do them as all as GenAI.

Runway, I have been waiting for this feature for a long time on #texttoimage platforms, but you have done it first. A big THANK YOU. #midjourney, #stablediffusion ehat’s up? I would love to compositnreal images into #textoimage created scenes. #amazon kinda already have a tool to do this for there ecommerce vendors. I want it too. #texttovideo #imagetovideo #videoediting #aivideo

Learn how to use the Gen-2 Motion Slider in today's Runway Academy.

Hey everyone! Excited to share my latest article on Runway Incursions. Dive into the depths of Runway Incursions with me and discover new insights by clicking https://lnkd.in/ecPKCEmw Don't forget to hit that *follow button* to stay updated on future articles. Let's embark on this journey together! #StayInformed #FollowForMore"

Use Multi Motion Brush to add realistic motion throughout your scenes. Learn how with today's Runway Academy.

Runway took it to the next level with Multi Motion Brush feature.⠀⠀⠀ ⠀ Controlling multiple areas of videos, generating them with independent motion is a real game changer. 👍 Though, identity theft in facial biometric verification comes to mind.⠀⠀⠀ ⠀ QoL features brings a lot to the table, but some of them are also fraud risks. With cyber insurance on the rise I wonder how insurance companies will cope with the scope and pace of these features. #genai #cyberinsurance 🤔

Gen-ai enhanced video animation appears to have some core functionality in place for explosive growth. If you like animation and gen-so, maybe creating natural language tools for rigging animation in Runway would be useful, fun, and lucrative. Imagine: “and then the character smiles a big grin, with a thievish look in their eyes.” You can map the utterance to segmented controls on specific features in the image, and voila - no-touch multi-motion brush. 🗣️👨🎨🎥🍿