LEPIX Ring’s Post

Impressive and very tangible comparison on model Performance. Seems OpenAI' Sora is falling behind and Google's VEO2 showing really strong results. That's actually one of the reasons why we developed at Mabanaft an inhouse Chatbot where we can easily exchange models based on their evolution. The market is moving so fast ( new GenAI models, -tools and the quality of models ) Change is normal (and exciting) and probably one of the reasons why I love what I do at work.

Great comparison examples! The AI video race is heating up, and Google’s Veo 2 has raised the bar. Veo 2 - Features - Resolution: Up to 4K (currently 720p) - Physics Realism: Smooth motion & lighting - Video Sharpness: Sharper, clearer with fewer errors - Prompt Complexity: Handles POV, drone, & wide shots - Creative Accuracy: Better execution on creative shots Excited to see where the AI video race take us next! What do you think? Are we ready for AI-generated 4K videos? Let’s discuss! #AI #Veo2 #Sora #GenerativeAI #Innovation

Ai image to video comparison with Runway, Luma, Kling and Minimax. Prompt was the same on everyone. ”A pigeon flies to a man’s sleeve and poops on it.” 😅🤷🏼♂️ First up Runway Gen3 Turbo that delivered classic credit wasting move of changing the image midway to something else. Next up Runway Gen3 Alpha that brought more ”action” to the shot. Did a few reruns with Luma to get the Pigeon to fly in. Luma also didn’t let me use word poop on the prompt. Kling also needed few tries to get thing flying. And last is my favourite Minimax with surprisingly good interaction with the man and the pigeon. First try also. 👌🏼 The outcomes will change everytime so these are just for you to reference and maybe choose one that you liked to try out. #aivideo #comparison #minimax #lumadreammachine #runway #kling #aicc

Another experiment with nerfs it reacts exactly as I want so the drone and camera images have different tone (deliberate) to see how it affects the visual. Next test will be with day and night data set For those of you who does wants the colors to be continuous it's even more important to do it when processing to nerf.

😍 Check out Runway's new advanced camera controls for Gen 3 Alpha. This lets you control both the direction and intensity of camera movements in AI-generated videos, giving you more precise control over how scenes unfold. Read the article on Maginative - https://lnkd.in/er8NdKid

𝐓𝐡𝐢𝐬 𝐢𝐬 𝐚 𝐟𝐥𝐢𝐠𝐡𝐭 𝐰𝐢𝐭𝐡 𝟒 𝐰𝐚𝐲 𝐩𝐨𝐢𝐧𝐭𝐬. 𝐂𝐨𝐧𝐭𝐢𝐧𝐮𝐞 𝐭𝐡𝐞 𝐭𝐫𝐢𝐚𝐧𝐠𝐥𝐞 𝐚𝐮𝐭𝐨𝐟𝐥𝐢𝐠𝐡𝐭 𝐛𝐲 𝐜𝐫𝐞𝐚𝐭𝐢𝐧𝐠 𝐚𝐧𝐨𝐭𝐡𝐞𝐫 𝐑𝐞𝐜𝐭𝐚𝐧𝐠𝐥𝐞 𝐚𝐮𝐭𝐨𝐟𝐥𝐢𝐠𝐡𝐭. How to connect from the start. https://lnkd.in/gTFdgiZ7 How to connect marvelmind to piwhawk. https://lnkd.in/grsgnmNn How to make Auto Mode and settle parameter. https://lnkd.in/gE5te5CW If you have any questions about the video, please ask me and I hope to get a positive value from the video. Qori Muhammad Febrianto Faiz Albar Risi Ari Kusnandar Alfarizi Ibrahim Fatahillah Hidayat

1 KEYFRAME to create 9 VIDEO CLIPS It all started this morning with the prompt "random image" and testing a combination of 2 newly found SREF STYLES. I was curious to see how Runway would generate different clip ideas into videos based of this sci-fi image. This video shows 9 VIDEO CLIPS generated in Runway Gen-3 Alpha. Some learnings from this AI VIDEO SESSION: - Runway can have problems with fingers ( most shots only 3 or 4 fingers ) - Runway tends to change the KEYFRAME image colors a bit - Runway impresses me in showing a correct "other side" when orbiting around objects / scenes - Runway's Unlimited plan may be the only "way" to explore video generation and generate as many clips you want to choose from the best results - ElevenLabs Webtool ( musictosoundeffects-com ) produces quite useable soundeffects for your scenes Both SREF STYLES used in this KEYFRAME image will be added to my next collection update ( sref.style ). #aivideos #sref #srefstyles #runway #gen3alpha #keyframetovideo #larssx

The Dream Machine - 8264 #AIArt Prompt: Generate a pointillism-style image that invokes a sense of nostalgia with a theme centered around animals. The scene should have a high contrast quality, accentuating the light and the dark aspects to create an image with a depth and rich in detail.

Ray Interactors help me engage things from afar, highlight interactable objects, and handle locomotion (spoilers for the next post...) https://lnkd.in/gGj-me5A

Vista is a generative driving world model by Gao et al. Built on Stable Video Diffusion it can generate driving scenes conditioned on a single input image and optional, additional control inputs. In this example we visualize the latent diffusion steps and the generated, decoded image sequence. Check out the code on the link below, or try the example with your own street images on the Hugging Face space (links in the comments).