Henry Daubrez 🌸💀
@henrydaubrez • 59,826 subscribers
Artist, Resident Filmmaker and Creative Director at @googlelabs and @flowbygoogle. In Progress... "Junkyard King and The Light Within"
Shorts
Videos
KITSUNE 🦊 💫 When I embarked on this project a month ago, I didn’t expect it to consume holidays, evenings, and far too many nights—but here we are. From the first scenes, I knew I had something special, and I don’t want audiences to watch an “AI film”—I just want them to watch a film, and hopefully, a good one at that. ( Sound On 🔈) 👇 KITSUNE is a tale of love between two souls separated by everything except their shared feelings of loneliness. I grew up in front of beautiful cartoons, from timeless treasures like those of Don Bluth, which I watched again and again to the point of damaging my VHS tapes, to early 90s anime, and later, of course, plenty of Studio Ghibli. And yes, before you ask—I know Hayao Miyazaki would disapprove of this film 100%, but then again… I’m not (only?) seeking approval. I’ve had goosebumps many times while reviewing the evolving states of this film, and I hope at least some of you will feel the same. Another famous director (Guillermo del Toro , I see you) recently said AI could create “semi-compelling screensavers,” and I see this as a step toward proving him wrong. Because you’ll ask: under the hood, there’s been tons of writing, re-writing, and switching directions mid-way. All shots were generated with Google’s text-to-video hashtag#VEO2. I faced countless challenges and hoops to bring my vision to life, finding ways to prompt and structure within the limitations of text-to-video despite VEO’s excellent prompt adherence. So, is VEO magic? No, not really—and the 1,700+ curated sequences on my hard drive (out of an estimated 5,000–7,000 total generations) are proof of that. What impressed me most was the global consistency, adherence, and how I could achieve tweaks by simply adjusting a few words. But what mattered most to me was creating something warm, nostalgic, and full of heart, avoiding the cold, clinical feel of so many films leveraging AI. Also, I’m a 40-year-old kid who grew up in front of the TV, has been creative his entire life, and has been designing professionally for nearly two decades. The more time passes, the more I know I can relate to what Nick Rubin said in that now-famous interview, where he mentions having no technical knowledge but trusting and building his own taste. If you like this film, this isn't just "Oh, AI is magic." You need to steer the damn ship. Then there’s MMAudio for sound effects, regular good old stock sound libraries, music on Udio for this version (yes, there’s a second version—more on that later), and tons and tons (and tons!) of editing, sound design, and small post-processing touches. Is this exposing risks for animators? Perhaps. Or it could also be their greatest companion, because once again, this is the worst it will ever be, yada yada yada.... No, it isn’t perfect, and if you look close enough, you’ll find defects and variations, but this is a film I’m proud of, not just an AI one... Enjoy. Wanna see a clean uncompressed version?
Henry Daubrez 🌸💀1,000,246 görüntüleme • 1 yıl önce
THE LIGHT WITHIN A concept trailer powered by Nano Banana Pro 🍌 I haven’t been impressed by new AI tools lately but this one swept me off my feet. The kicker? I used Google Flow and a few wild tricks to get perfect consistency. Here's a step-by-step full breakdown 🧵 #1/15
Henry Daubrez 🌸💀146,624 görüntüleme • 6 ay önce
I rewatched Home Alone with my son today and it was funny to notice how many inconsistencies, raccord issues, and tiny broken details are in a film considered a gem of the 90s. We never cared. We were in the story, so our brains let the imperfections slide. What struck me is how different the standard feels when using AI. I obsess over every tiny continuity mistake, every aesthetic shift, every micro detail, because I know any small flaw will immediately be used by the anti-AI crowd to discredit the whole thing. Putting both side by side, it is insane. Traditional filmmaking is full of human errors and nobody cares because the story carries it. Meanwhile, I am pushing for an even higher level of technical perfection simply to avoid giving people ammunition to dismiss the work. And here is the spoiler. Even if the story was flawless, even if the rhythm was right, even if the continuity was perfect, the conclusion would still be “it is stolen content” or “I don’t like AI.” At the end of the day, the story is what matters. The rest is noise.
Henry Daubrez 🌸💀131,082 görüntüleme • 6 ay önce
WHITE RABBIT A glimpse into a new world 1-minute screentest. How this came together 👇
Henry Daubrez 🌸💀45,718 görüntüleme • 2 ay önce
Junkyard King - Ep. 0 | MAKING OF ⚙️ HERE’S HOW I MADE THIS, STEP BY STEP 🧵👇 1/14
Henry Daubrez 🌸💀41,364 görüntüleme • 2 ay önce
⚡️🌸 JUNKYARD KING Episode 0 ⚔️ Studio quality is getting closer than people think...
Henry Daubrez 🌸💀19,709 görüntüleme • 2 ay önce
Haters gonna hate but: VEO 2 and Google just blew everyone else out of the water. 🔈ON I just came out of a very short night as I just couldn't stop creating and here's the result and a small walkthrough of my experience with the tool: It felt like exploring new worlds, easily blending styles and aesthetics but also being absolutely amazing at understanding small variations. I love my scene but I want snow blowing towards the camera? No problem. I like the video but I would have wanted some claymation? Sure, got you covered. The vast majority of what I threw its way was understood and worked out: night shot, flash light, stop-motion, first person, watercolor, pixel art, etc. I even tried complicated prompts using mirrors and having a person repeatedly reflect in a "bug-in-the matrix-ouch-my-brain" fashion and it worked..you can actually see it in the montage with the man in the blue shirt. Impressive. I have been trying each and every GenAI video tool over the last couple of years and the results for a mere text-to-video are simply in a totally different category than all the others on the market right now...including Sora. I am not even sure where to start: - Resolution in 720p - Prompt adherence is of the highest level, it understands complex scenes and creations, very long prompts with highly detailed explanations of the expectation, camera movements, descriptions, mood, etc. - Format is 8 sec but during those 8 second it understands when you ask for jump cuts or some kind of in-clip editing thanks for the extra angle and scene + character adherence on those cuts VEO. - It is great at creating different styles. It struggled a little with some very specific elements such as asking for an animated oil painting but this is fair... - Generations are fast but it might be because this is an early access - This is the first time I precisely describe the scene I have in my head and actually get it directly or close to it. I had to go for some reroll on some of them but it wasn't because of hallucinations but rather because I felt like I could prompt a little better and fine-tune the output - It is great at filling the gaps in terms of environmental animation - Movements are realistic, some small hallucinations if I try to get a swordfight for example but still lightyears from all the other tools
Henry Daubrez 🌸💀75,618 görüntüleme • 1 yıl önce
Born in a small town in Belgium. Studied computer graphics. Started in games as a 3D artist intern. Graduated with honors. Taught myself digital painting. Then taught it to others, before YouTube was even a thing. Started as a Flash animator. Moved into web design. Became Creative Director at 24. Became partner in a digital studio. Co-founded a graphic design collective. Won design competitions. Built a second-hand clothing startup that gained traction. Became partner in a new studio in Belgium. Opened an office in Chicago. Then Paris. Then Amsterdam. Helped grow it into one of the most awarded studios in the industry. Designed permanent interactive installations seen by millions every year (including one at Navy Pier in Chicago). Became CEO. Led the studio through acquisition. Spoke about design and culture in Tokyo, New York, Barcelona, and beyond. Exhibited work around the world. Sold AI-assisted art at Sotheby’s. Dove into GenAI. Released Kitsune. Opened a new door. Now working in film. Signed with Google Labs as Filmmaker in Residence. ⸻ I’ve been a Flash animator, a web designer, a developer, a 3D generalist, a character modeler, a digital painter, a teacher, a producer, a business developer, a CEO, a head of design… …and I’m still figuring things out. Don’t let where you’re born dictate what’s possible for your life. And don’t let people who know nothing about you, your path, or your work call you talentless or tell you to “pick up a pencil” because you’re using AI. You know who you are.
Henry Daubrez 🌸💀12,852 görüntüleme • 1 ay önce
So, Google said, "Hey, wanna make a film for #GoogleIO ? You have Carte Blanche." 🤯 For a brain that’s been mainlining animation and questionable VHS tapes since forever, that kind of freedom is both a dream and a mild panic attack. To be honest, until recently I never really thought film making was in the cards for me. Life happened, and although I ended up being a designer, film was more of a distant dream of a young version of myself. Following the unexpected attention and warm reception for Kitsune earlier this year, Electric Pink came out as my own self-administered therapy session: a coming-of-age story that’s basically me trying to figure out my own creative wiring, from the fuzzy nostalgia of childhood to the full-on HD chaos of the present. Turns out, the path to making stuff is paved with a lot of self-doubt and sudden left turns. I’m hoping this film can build on the connection Kitsune fostered. For this project, focused on exclusively using a lengthy process involving leveraging Imagen 3 to create the base of still frames (Imagen was amazing at iteratively enabling me to lock art direction) which I would then heavily retouch and prepare to get them as perfect as needed. Then each scene would be run through VEO2 to get animated clips based on those images (and then a very obvious bunch of sound design, voice design, editing, post-production and what not). Now, what's been amazing as I was nearing the end of this project, is to see individual separate tools converge towards becoming "Flow": Google's AI filmmaking tool built on DeepMind's brilliant tools (Veo, Imagen, Gemini). I must say it’s been fascinating seeing how this tech can translate the random noise in my head into something… well, film-like. It felt like a real step forward in allowing creators to iterate and refine their vision without getting bogged down in organization, and for example having an opportunity to use a first-frame to last-frame interpolation proved to be extremely useful as I could fully control the action I was aiming for. Now in its current version, Flow offers even more features such as the scene builder to easily expand storylines, or ingredients to control character, object, and environment consistency...can't wait to dig more into those possibilities...they definitely would have made my life easier during the creation of this film which I hope you will enjoy. Read more here: And there:
Henry Daubrez 🌸💀51,050 görüntüleme • 1 yıl önce
VEO 2 by Google DeepMind : MY CHEAT SHEET Alright, so after 500h-ish spent on VEO and giving birth to both "Kitsune" and "Banished", tons of people asked for a making-of. Instead, I decided to give you what I actually know of VEO 2 to this day. Please share! it's made to be spread around! 1/ If you're not using a LLM (Gemini, ChatGPT, whatever), you're doing it wrong. VEO 2 currently has a sweet spot when it comes to prompt length: too short is poor, too long drops information, action, description etc. I did a lot of back and forth to find my sweet spot, but once I got in a place I thought felt right, I used a LLM to help me keep my structure, length, and help me draft actions. I would then spent an extensive amount of time tweaking, iterating, removing words, changing order, adding others, but the draft would come from a LLM and a conversation I built and trained to understand what my structure looked like, what was a success, or a failure. I would also share the prompts working well for further reference, and sharing the failures also for further reference. This would ensure my LLM conversation became a true companion. 2/ Structure, structure, structure Structure is important. Each recipe is different but same as any GenAI text-to something, it looks like the "higher on the prompt has more weight" rule applies. So, in my case I would start by describing the aesthetics I am looking for, time of day, colors, mood, then move to camera, subject, action, and all the rest. Once again, you might have a different experience but what is important is to stick to whatever structure you have as you move forward. Keeping it organized also makes it easier to edit later. 3/ Only describe what you see in the frame If you have a character you want to keep consistent, but you want a close-up on the face for example, your reflex will be to describe the character from head to toe and then mention you want a close-up...It's not that simple. If I tell VEO I want a face close-up but then proceed to describe the character's feet, the close-up mention will be dropped by VEO... Once again, the LLM can help you in this by giving it the instruction to only describe what is in the frame. 4/ Patience Well, it can get costly to be patient, but even if you repeat the same structure, sometimes changing one word can still throw the entire thing out and totally change the aesthetics of your scene. It is by nature extremely consistent if you conserve most words, but sometimes it happens. In those situations, trace your steps back and try to figure out which words are triggering a larger change. 5/ Documenting When I started "Kitsune" (and did the same for all others), the first thing I did was start a Figjam file so I could save the successful prompts and come back to them for future reference. Why Figjam? So I could also upload 1 to 4 generations from this prompt, and browse through them in the future. 6/ VEO is the Midjourney of video Currently, no text-to-video tool (Minimax being the closest behind) gave me a feeling I could provide strong art directions and actually get them. I have been a designer for nearly 20 years, and art direction to me has been one of the strongest foundations of most of my work. Dark, light, happy, sad, colorful or not, it doesn't matter as long as you have a point of view and please...have a point of view. Recently watched a great video about the slow death of art direction in film (link in comments) and oh boy, did VEO 2 deliver on giving me the feeling I was listened. Try starting your prompts with different kinds of medium (watercolor for example), the mood you are trying to achieve, the kind of lighting you want, the dust in the rays of light, etc... which gets me to the next one 7/ You can direct your colors in VEO It's as simple as mentioning the hues you want to have in the final result, in which quantity, and where. When I direct shots, I am constantly describing colors for two reasons: 1. Well, having a point of view and 2. reaching better consistency through text-to-video. If I have a strong and consistent mood but my character is slightly different because of text-to-video, the impact won't be dramatic because a strong art direction helps a lot with consistency. 8/ Describe your life away Some people asked me how I achieved a good consistency between shots knowing it's only text-to-video and the answer is simple: I describe my characters, their unique traits, their clothing, their haircut, etc..anything which could help someone visually impaired have a very precise mental representation of the subject. 9/ But don't describe too much either... It would be magical if you could stuff 3000 words in the window and have exactly what you asked for, right? Well, it turns out VEO is amazing with its prompt adherence, but there is always a moment where it starts dropping animations or visual elements when your prompt stretches for a tad too long. This actually happens way before the character limit allowed by VEO is reached, so don't overdo it, it's no use and will play against the results. For info, 200-250 words seems like a sweet spot! 10/ Natural movements but... VEO is great with natural movements and this is also one of the reasons why I used it so extensively: people walking don't walk in slow-motion. That being said, don't try to be too ambitious on some of the expected movements: multiple camera movements won't work, full 360 revolutions around a subject won't work, anime-style crazy camera movements won't work, etc... what it can do is already great, but there are still some limitations...
Henry Daubrez 🌸💀30,602 görüntüleme • 1 yıl önce
OVERGROWN 🌱 | Official Teaser • 1,000+ hours of craft. • 200 shots in Photoshop & Google Flow. • 13-minute contemplative drama. A contemplative drama exploring chronic resilience and the heavy beauty of survival. Currently in consideration for Tribeca 2026. Best experienced with sound on 🔊
Henry Daubrez 🌸💀11,221 görüntüleme • 4 ay önce
THE RITUAL OF T̴́̀.A̴̔͗.Ṕ̵̊. Tomorrow My humble contribution to 6529's Meme Card drop. 🔊 ON A short thread 🧵
Henry Daubrez 🌸💀24,394 görüntüleme • 1 yıl önce
KIDD-0 🌸 🏴☠️ - AN ANIME OPENING (and the future of IP development) For the last year or so, one of the most debated topics in the GenAI space has been its potential to disrupt the animation industry. I briefly touched on this with KITSUNE and the waves surrounding the film. Another area of the entertainment industry that seems ripe for change is intellectual property (IP) development. The massive opportunity with Generative AI lies in fast-tracking world-building for new IPs. In the past, how costly was it to give producers, studios, and investors a proper glimpse of what they were buying into? Buying a script is one thing, but effectively communicating the visual universe supporting that script was always a challenge. If you've been following me for the past few years, you know that KIDD-0 (the pink-haired boy) is a character and universe I started developing through my early GenAI work (think Jax Diffusion, late 2021). I've long wanted to expand this world: somewhere between Peter Pan, fantasy, steampunk, and coming-of-age stories. Given the potential I saw for AI-assisted animation with KITSUNE, I decided to create an anime opening for what a KIDD-0 TV series could look like (because, yes... it's on my bucket list). Quick backstory: as a kid (here we go again), I had a CD-ROM with a trailer to Record of Lodoss War but couldn't watch the series (no internet, remember?), so I watched this trailer again and again and again...took a few years until I could watch it along with a few other Animes from the 90's, and obviously Lodoss' openings stuck with me for a long time after that. Anyway, there's still plenty of room for world-building, but after pushing this project as far as I could, including writing Japanese titles and credits (thanks, ChatGPT!), I see even greater potential for this IP. Of course, I expect the final product to be a hybrid between traditional film making, VFX, and generative AI, but this reinforces my belief in leveraging Generative AI to prototype IPs, reduce costs, and attract funding for new projects. The video is a mix of VEO2, Udio, Screenflow (yeah, I know, but...), Topaz Video Enhance, Photoshop, and a few magic tricks here and there. (PSSSST... please poke me if you want to discuss developing KIDD-0 into a series or feature...it's high on my bucket list!) 4K video link in the comments :)
Henry Daubrez 🌸💀17,274 görüntüleme • 1 yıl önce
RUN RABBIT RUN 🐇 – VEO3 CHEAT SHEET – Part I ( 🔈Sound on) So, I’ve been more extensively testing Veo3 over the past few weeks, and I wanted to see how much dynamism I could get from the new model and how I could use motion and placement to consistently tell a story while changing universes each time. Now, my findings: 1/ First, it does feel more like a Veo2.5 than a full leap. But it’s a meaningful step forward, especially in motion and complexity. Maybe I'm a little jaded, but I expect new versions of models to be strong departures from the previous one. Veo2 was already great, this one feels like an evolution, not a revolution. 2/ My Veo2 Cheat Sheet still mostly apply for Veo3: For example, I still prompt the most important elements first (often style). Just as a reminder: 3/ It's pretty good at filling the gaps when you're writing simple prompts. For cool, out-of-the-box results, just describe short scenes, but be careful with consistency. Weird tradeoff, I guess. 4/ The token window is a LOT longer, so you can get more complex, multi-step actions from your prompt. Veo2 would often lose the thread, Veo3 can keep up with sequential and layered instructions. That’s a win. 5/ Camera motion is much better. Handheld, tracking, wide-angle moves, it holds together well. Still lacks control, but it’s a clear jump from Veo2. Definitely worth noting, especially given how many people have been experimenting with Higgsfield lately. 6/ Like Veo2, POV and fantasy prompts seem to lean into a game engine vibe. Not quite realism — more like stylized cutscenes or in-game shots. Sometimes it works, sometimes it’s distracting. 7/ Aspect ratio tip (thanks Kyle Salazar): want to reduce the black bars? Add “aspect ratio 1.85:1” at the start of your prompt. Haven’t tested it yet, but apparently it helps. -8/ You can generate 360° videos. Try “360 video,” “monoscopic,” or “equirectangular.” Not flawless (some seams), but it works surprisingly well. See my previous post for this. 9/ Prompting for aesthetics is still hit-or-miss. Consistency is tough. You can over-describe your characters and key elements to keep them stable throughout, but for style...especially animation, it’s still hard to maintain. Hopefully "ingredients" or style-locking features will help in the near future. 10/ Bodies and turning movements are still tricky. Fast actions still distort limbs, creates artifacts, etc... Most shot are fine, but if you push too much for extremely dynamic scenes, you'll definitely notice some distortions (some in this video too) - What’s still missing (but probably coming): – Image-to-video – Start/end frame control – Locking a voice to a character – Prompt ingredients or persistent visual logic This is just Part I :) I’ll keep posting notes as I go.
Henry Daubrez 🌸💀14,683 görüntüleme • 1 yıl önce
Daha fazla içerik yok.