Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

🏃‍♂️ Great tips on a run cycle loop from Sony Senior Anim Chalermphol Wattanawongtrakool! Feel that rhythm! Source: #agorastudio #framebyframe #animation #animator #reference #resource

205,200 Aufrufe • vor 1 Jahr •via X (Twitter)

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

🇨🇳 Another great Chinese Model, OmniHuman-1.5 from ByteDance Turns 1 image plus a voice track into expressive avatar video by pairing a System 1 and System 2 inspired planner with a Diffusion Transformer, Produces coherent motion for over 1 minute with moving camera and multi character scenes. Most avatar models move to the beat of the audio but miss meaning, so gestures feel generic and emotions feel shallow. The fix here is a Multimodal LLM planner that listens to the speech and drafts a structured plan describing intent, emotions, beats, and high level actions, which gives the motion engine clear semantic targets instead of only rhythm. The motion engine is a Multimodal Diffusion Transformer that fuses the plan with audio, the single reference image, and optional text prompts, then synthesizes continuous body, face, and head motion that matches both words and tone. A key trick is a Pseudo Last Frame, a synthetic target that summarizes the next expected state, which stabilizes fusion across modalities and keeps motion consistent over long spans. From just 1 image and speech, the system outputs speaking avatars with synchronized lips, context aware gestures, and continuous camera movement, and it also supports multi character interactions without manual choreography. Reported results show strong lip sync accuracy, high video quality, natural motion, and close match to text prompts, and the same setup works on nonhuman characters too.

Rohan Paul

63,859 Aufrufe • vor 10 Monaten

[Throwback] Kobayashi Arata,a professional dancer, give detailed abt taehyung’s dance abilities on Fire and Dynamite 👤Whenever I see Taetae dance, I got the impression that he is a groove-type dancer. He has good detail and dances like he is riding the music. I like dances that convey the dancer's own vibe that only they have, and Taetae's dance is like that. His neck's trajectory is continuously moving back and forth. It's a professional's technique. His dance is very professional. It's very difficult to grab the rhythm in a simple way, but Taetae has a very good sense for that. Even when he isn't moving, the texture of his dance movements are very exquisite.The way he dances to his heart's content is wonderfully cool. The way he feels the sound (music) is expressed through the movement of his neck. The way he burns his feelings into the music is at a very professional level. One more thing to mention, the way he approaches a certain melody is always very outstanding and beautiful. You can see his great sense for dancing in his ad-libs, besides the set choreography. The neck is the part of the body that can express rhythm, so it especially requires skill. Taetae uses his neck extremely well. People at a high level of dancing can draw a hot response from just moving their neck a little. A great dancer knows that just a simple neck movement can look really cool. 1 speculate that Taetae is a dancer who knows this. The way he paces himself is no joke. The reason he can comfortably maintain the rhythm with his whole body is because he properly feels the music inside his body His dance has a mature charm. It's something different than the brilliance and dynamism that can be seen from the outside. From an angle that surpasses all that, he expresses how he perceives the beauty of the dance. He is the type that enthralls others by actually enjoying the music himself. He is like a miso soup made by a first-class chef. It doesn't make sense how smooth he is. (All of his movements are very smooth.) People, who enjoy music itself in the first place, look cool with the way they ride the rhythm. He dances as if he has surrendered himself to the sound (music), and as if he is submerged in the music. He is able to feel the rhythm with his whole body. Taetae's hidden details, in which he accomplishes things like they're nothing, make my heart beat. It's a difficult thing to do. You can't do this just by learning it. He has a good sense for dancing. The way he dances is very smart. In conclusion, Taetae is at an extremely high level of dance. From the little details that are hard to catch at first glance, to his dance skills that read the music, making it feel like a "professional dance" style and feel, can all be seen in the ad-libs as well. These parts all depend on the individual's sense for dance. What can be seen from here is that Taetae has a good sense of understanding and feeling sound(music), and i really want to become someone like that too. 28min video: trans cr: KTHbase/KTH_FanBase

❀ ✿

46,199 Aufrufe • vor 1 Jahr

Made with GPT-Image-2 + Seedance 2.0 by Yapper Full prompt below: "Create a 15-second stylized 3D animated cooking comedy with playful Eastern martial-arts cooking energy. Use the uploaded character sheet as the fixed visual reference for BAOBAO and the senior chef. Keep both characters consistent throughout the video, including their faces, proportions, outfits, colors, signature props, and warm cute design style. Scene: A cozy Chinese restaurant kitchen with one clear wooden prep table in the center. The background includes bamboo steamers, simple red lanterns, warm wooden shelves, ceramic bowls, chopsticks, soft steam, flour dust, and golden restaurant lighting. The background should feel rich and warm, but not crowded. Keep the two characters as the main focus. Characters: BAOBAO is a cute small chef, hardworking and expressive. The senior chef is round, cheerful, sleepy, lazy, and lovable. Story: BAOBAO works very hard to make perfect xiaolongbao. The senior chef pretends to supervise, but slowly gets sleepy. When BAOBAO finishes a beautiful basket of xiaolongbao, the senior chef cannot resist the smell and secretly eats one. BAOBAO catches him. Instead of getting angry, BAOBAO uses playful martial-arts-style cooking footwork to protect the remaining dumplings. The moment becomes a funny cooking performance. In the end, the senior chef admits the dumplings are delicious and proudly approves BAOBAO. Timeline: 0-3s: Medium full-body shot inside the warm Chinese restaurant kitchen. BAOBAO stands at the central wooden prep table, focused and hardworking, rolling dough and folding xiaolongbao with small precise hand movements. Flour floats in the golden light. The senior chef sits nearby, pretending to supervise, but slowly becomes sleepy while holding chopsticks. 3-5s: BAOBAO speeds up with graceful training-style cooking movements, folding dumplings neatly one after another with rhythmic fingertip taps. The senior chef opens one eye, impressed, then lazily leans back again as if pretending not to care. 5-7s: BAOBAO proudly opens the bamboo steamer to reveal a basket of perfect xiaolongbao. Warm steam rises in soft ribbons. BAOBAO turns away briefly to wipe flour from his hands. The senior chef suddenly wakes up, eyes bright, attracted by the smell. 7-9s: The senior chef quietly reaches in with chopsticks, takes one xiaolongbao, and eats it quickly. BAOBAO turns back and freezes in surprise. The senior chef freezes too, cheeks full, trying to look innocent. Hold this funny pause for clear comedy timing. 9-12s: The senior chef slowly reaches toward the steamer again for a second dumpling. BAOBAO quickly slides in front of the basket with playful training-style footwork. He spins the bamboo steamer lid in a smooth circular cooking flourish and gently covers the basket. The senior chef reaches left and right comically, but BAOBAO blocks him each time with the lid and a proud little stance. 12-15s: The senior chef stops, laughs warmly, and gives BAOBAO a big thumbs-up. BAOBAO blushes proudly. The senior chef pats the table beside the steamer in approval. BAOBAO presents the bamboo steamer toward the camera. Both characters smile together in a warm heroic final pose as soft steam curls around them. Visual style: Premium stylized 3D animated family-film quality, expressive cartoon face acting, warm cinematic Chinese restaurant lighting, polished bamboo textures, soft flour particles, gentle steam ribbon effects, cozy restaurant atmosphere. Camera: Smooth cinematic camera, mostly medium full-body shots. Keep BAOBAO, the senior chef, the wooden table, bamboo steamer, dumplings, chopsticks, hands, feet, and facial expressions clearly visible. Use a gentle push-in during the secret dumpling eating moment. Use a small controlled camera move during BAOBAO's steamer-lid cooking flourish. End with a warm final hero shot. Motion style: Playful martial-arts cooking choreography, rhythmic hand movements, cute comedy timing, smooth readable animation, expressive body acting, energetic but gentle. Audio: Cheerful Eastern cooking rhythm with light drums, wooden percussion, bamboo flute melody, rolling dough sounds, fingertip folding taps, steamer lid sound, steam whoosh, chopstick tap, funny pause beat when the senior chef is caught, warm triumphant ending music. No lyrics."

Zar⭕on

10,472 Aufrufe • vor 1 Monat

One of the most common mistakes manufacturers make is relying on engineering logic to shape user behavior, while ignoring how users actually feel. Engineers often believe that as long as a feature exists, moves, or looks different, users will appreciate it. Users, however, care about something much simpler: whether the experience feels natural, refined, and intuitive. Great user experience is never created by stacking options. It is created by coherence and emotional consistency. Samsung’s recent update to Lock Star is a clear example. The new version introduces unlock animations that appear to add playfulness, yet the real experience feels disappointing. Many of these animations come across as childish, low level, poorly connected, and mechanically rigid. They feel closer to children’s UI effects than something designed for Samsung’s broad premium user base. The issue is not the presence of animations themselves. The problem lies in the mindset behind them. These effects are clearly driven by implementation logic rather than user perception. The assumption seems to be that adding visual flair automatically creates delight, while ignoring rhythm, emotional tone, and aesthetic maturity. This reflects a long standing pattern in Samsung’s software design. Random stickers, cartoon like emojis, and decorative elements often appear without a cohesive design language. There is a belief that younger users are attracted to these visuals, yet for Gen Z, such elements often feel outdated, immature, and disconnected from a premium identity. Strong product design allows user experience to constrain engineering decisions. When engineering begins to entertain itself, users quietly begin to disengage. Let's see how far apart Samsung and Xiaomi are.

Ice Universe

59,669 Aufrufe • vor 6 Monaten

Introducing ml-intern, the agent that just automated the post-training team Hugging Face It's an open-source implementation of the real research loop that our ML researchers do every day. You give it a prompt, it researches papers, goes through citations, implements ideas in GPU sandboxes, iterates and builds deeply research-backed models for any use case. All built on the Hugging Face ecosystem. It can pull off crazy things: We made it train the best model for scientific reasoning. It went through citations from the official benchmark paper. Found OpenScience and NemoTron-CrossThink, added 7 difficulty-filtered dataset variants from ARC/SciQ/MMLU, and ran 12 SFT runs on Qwen3-1.7B. This pushed the score 10% → 32% on GPQA in under 10h. Claude Code's best: 22.99%. In healthcare settings it inspected available datasets, concluded they were too low quality, and wrote a script to generate 1100 synthetic data points from scratch for emergencies, hedging, multilingual etc. Then upsampled 50x for training. Beat Codex on HealthBench by 60%. For competitive mathematics, it wrote a full GRPO script, launched training with A100 GPUs on watched rewards claim and then collapse, and ran ablations until it succeeded. All fully backed by papers, autonomously. How it works? ml-intern makes full use of the HF ecosystem: - finds papers on arxiv and reads them fully, walks citation graphs, pulls datasets referenced in methodology sections and on - browses the Hub, reads recent docs, inspects datasets and reformats them before training so it doesn't waste GPU hours on bad data - launches training jobs on HF Jobs if no local GPUs are available, monitors runs, reads its own eval outputs, diagnoses failures, retrains ml-intern deeply embodies how researchers work and think. It knows how data should look like and what good models feel like. Releasing it today as a CLI and a web app you can use from your phone/desktop. CLI: Web + mobile: And the best part? We also provisioned 1k$ GPU resources and Anthropic credits for the quickest among you to use.

Aksel

1,264,490 Aufrufe • vor 2 Monaten

I learned this the hard way: do NOT use SwiftUI if you want your app to look and feel amazing. At least when coding with AI. (sorry, Apple colleagues reading this 😅) I'm sharing my process vibe coding this calorie tracker. I get a lot of questions about the fluid transition in the video. Here's the whole story. Initially, Claude built the grid with SwiftUI. It was quick and easy, and looked good! But the transition to the day view was a boring navigation push/pop. No fun. I wanted something custom. I asked Claude to make it a fluid transition that remaps the food tiles from their source to destination positions. All hell broke loose. Claude tried a bunch of horrible things. Initially it used matched geometry effects, which worked OK but didn't lend themselves well to gesture-driven animations. So it resorted to SwiftUI preference keys + geometry readers to figure out the source and destination positions and calculate the interpolated position based on gesture progress, coordinating across grid and day views. But this meant it had to write a custom layout because it couldn't reposition tiles inside the native SwiftUI grid. And it had to do an awkward handoff between views, which always created ugly pops or jumps. And don't get me started on trying to put it on a bouncy spring, that only made the math 10x buggier. Fortunately, Claude Fable was smart enough to see that this was becoming a disaster (and discover most of the issues itself, in the simulator), so it pivoted away from SwiftUI. Opus might not be so wise, so you'll have to pay attention and intervene. Ultimately, it rewrote it in plain UIKit and everything turned out great. After that, we moved from 2D images to 3D assets, which introduced a new set of performance challenges and yet another rewrite to a single Metal layer, which is what you see below. I can write more about the 2D-to-3D saga if anyone's interested. If I were to do it again, I'd just say "Don't use SwiftUI" from the very first prompt, and save a few hours of headaches. SwiftUI can be amazing for a human iterating directly in code. But agents don't benefit from any of its advantages. Plus, agents have seen decades of UIKit training data, so they're great at writing it, and it's far more flexible. Here's hoping we see more agent-friendly iterations of SwiftUI in the future. Till then, I'm probably going to avoid it.

Anshu

107,685 Aufrufe • vor 25 Tagen