HEY-GEN

Context We are an escape room (Blackout-Room, https://blackout-room.com/) with two universes: Le Sorcier Maudit — inspired by the Harry Potter universe (without direct elements from the movie) Mécanisation — steampunk/post-apocalyptic universe A few years ago, we ordered promotional videos (classic editing based on professional photos, simple effects: glow on the wand, igniting torch, light parallax). These videos worked well, but have become too familiar to our audience — we showed them too often, fresh content is needed. Video 1: https://www.youtube.com/watch?v=DSy6YgB_gdQ Video 2: https://www.youtube.com/watch?v=hfSDDeMLjEU Video 3: https://www.youtube.com/watch?v=2numswdTBv0 Project Goal To recreate promotional videos in the same spirit (dynamic, atmospheric, for advertising on AdWords/Instagram/Facebook), but using modern video generation/animation tools based on AI, to go much further than simple static effects: for example, bringing photos to life into a full video scene (a wizard casting a spell on another, steam/smoke erupting from a steampunk mechanism, a character moving in the scenery, etc.) based on our existing photos. Required Results 3 videos in the first stage (the video about the gift certificate will be added later if necessary): Video for the universe "Le Sorcier Maudit" Video for the universe "Mécanisation" General video presentation of the escape room (both universes) Each video in 3 formats: square 1:1 (1080x1080), vertical 9:16 (Stories/Reels), horizontal 16:9 (feed/YouTube) Short duration (~20-30 seconds) MP4/MOV, H.264, audio stereo AAC 128kbps+ What Style We Are Looking For Smooth, well-crafted transitions between scenes (without abrupt cuts) Animation of decor/character elements, not just Ken Burns/basic parallax Compliance with the brand's visual identity (logo and brand book will be provided) Decreasing rhythm: the first scene is calmer to allow reading the text, subsequent scenes are shorter What We Provide Professional photos of both universes (high resolution) Logo in vector format (AI) Old videos for reference (to show the tone/style that needs to be recognized but surpassed — not copied) Script text + storyboard of scenes (parallax), which I will prepare myself What We Expect from the Performer Proficiency in one or more modern AI video generation/animation tools (image-to-video, motion, etc.) Classic editing skills (transitions, typography, sound design, export according to specified characteristics) Portfolio with examples of AI animation on static images (not just classic editing) Ability to quickly make adjustments based on feedback (in a ticket format: one brief per video, adjustments collected in one place) Budget / Timeline To be discussed depending on the number of videos and the level of animation — please indicate a price range per video and an estimated delivery time in your response. How to Respond Please send: Portfolio (preferably with an example of AI animation based on a static photo) AI tools you usually use Approximate price and timeline estimate

Project Description to create a camera path recreation. That is: same camera height; same focal length; same speed; same rotations; same FOV. Options for how this can be done: Option 1. AI Camera Match (best) ⭐⭐⭐⭐⭐ There are models that can analyze the movement of the first camera and generate or adapt the second video so that the camera moves almost identically. This is done through: Runway Gen-4 Google Flow (Veo) Luma AI Dream Machine Kling AI 2.1 The essence is: “Here is the video BEFORE the repair. Here is the video AFTER the repair. Make the camera of the second video the same as in the first.” This already works, although sometimes it requires several iterations. Option 2. Optical Flow + Warp ⭐⭐⭐⭐☆ If the difference is small, the editor can use AI in DaVinci Resolve or After Effects. AI: stabilizes; slightly changes perspective; adjusts speed; creates a smooth morph between frames. This is much cheaper. Option 3. Image-to-Video ⭐⭐⭐⭐⭐ If there are many photos or good video after the repair, AI can almost “fly” the camera through the room again. This is already being done by: Veo Kling Hailuo Luma Then it doesn't matter how the operator filmed at all.

Need an intro for a YouTube channel on the topic of fishing https://www.youtube.com/@%D0%A0%D0%B8%D0%B1%D0%B0%D0%BB%D0%BA%D0%B0%D0%B2%D0%A0%D0%B0%D0%B4%D0%BE%D0%BC%D0%B8%D1%88%D0%BB%D1%96. I envision it with an animated background in water where a perch fish swims in and out in reverse direction with appropriate text design for the channel name or something similar.

Generate a video clip from the rendering of a building using a photo of the object according to the reference and with the impact of the script. A developed test prompt needs to be refined. Possible neural networks for generation: King AI, Runway, Luma, Google AI Pro, Google AI Ultra. But this is not certain; you can suggest your own)

There is an author and songwriter using AI tools. There are 15 songs. They have been published through a distributor on all platforms, including social media. It is necessary to create video clips for social media posts based on the songs, specifically for TikTok and YouTube Shorts. For each song, two clips should be made featuring the best moments. AI video generators can be used for visualization, but their quality must be realistic to avoid negative effects from the publication. It might be better to create a video clip and then overlay the audio from the song that has been officially uploaded to social media during the publication. I will listen to your proposal. However, some clips can also be presented with good visual representation of neutral elements, such as a player interface, non-trivial audio waves, etc. What appeals to women can also be played around, looking at niche trends and how women are already overlaying the song on their creative works. The author's website with its own style, which would be appropriate to transfer into recognizability in the clips: https://nefaktalex.com/