@JSFILMZ0412 Wow. Crazy difference. Contrast, saturation, lighting etc.
English
Retropunk
9.9K posts
Retropunk
@RetropunkAI
Creative Engineer Lead, AI Explorer, Developer, Musician ✨CPPs: Runway, Dreamina✨
Cedar Park, TX Katılım Ekim 2009
1.3K Takip Edilen1.7K Takipçiler
Still think there’s no difference in Seedance 2.5? 👀
Look at the skin tones.
I used to own an ARRI ALEXA, and I’m telling you—the skin tones in 2.5 are noticeably more organic. The way it handles color, texture, and subtle tonal shifts in the skin just feels more natural and filmic.
I’ve been seeing a lot of discussion about whether 2.5 is even worth the upgrade since, on the surface, not much seems to have changed.
For me, this alone makes the upgrade absolutely worth it.
English
This indie studio is making a game where you deliver packages in a miniature tilt-shift zombie apocalypse
- Dodge zombie hordes hauling TNT and fragile boxes
- Solve puzzles to clear a path
- Don't drop packages
It's called Tiny Hero Courier. Would you play this?
English
@seiiiiiiiiiiru @Hailuo_AI excellent test with simplae animation reference guide. Have you done anything with depth? Like a greyscale low poly 3D animation test as a reference? I'm definitely gonna try this. Thanks for sharing.
English
【Hailuo AI(@Hailuo_AI)】H3📢
AfterEffectsで作成した雑な図形アニメーションをH3で参照して画像を動かす実験。
これが出来るモデルは優秀です🧐✨
#MiniMaxH3
SEIIIRU😈動画生成AIを使う映像クリエイター@seiiiiiiiiiiru
【Hailuo AI(@Hailuo_AI)】H3📢 久々にHailuoから新しいモデルが登場しました！ クレジット消費少な目で性能良いので現状コスパ最強と言えるかもしれません！ 昔からアニメスタイルやリアルスタイルなどバランスのとれたモデルだったので、今後様々なシーンで使用されると思います。 #MiniMaxH3 [ 仕様 - MiniMaxH3 ] ✅ネイティブマルチモーダルテキスト ・画像 / 動画 / 音声 を組み合わせて生成・編集できる。 ✅Omni Referenceキャラクターの一貫性 ・カメラワーク / モーション / スタイル を複数の素材から参照できる。 ✅高い指示追従性 / 一部分だけ修正 ・キャラだけ変更 / 背景だけ変更 のような編集が得意。 ✅商用レベルの品質CM ・MV / UIモック / ゲームPV / ロゴアニメーション などを想定。 あ、それと日本語性能かなり良いっすよ！ 「すごい"ん"らしいんですよー」の甘噛みはコチラの入力ミスでした🙏 (※たぶん早期アクセスです)
日本語
@lena_z01 @BytePlusGlobal Why not just share it? This whole comment and I’ll give you something is not cool. Especially when the guide might be useless.
English
Stop wasting credits testing Seedance 2.5.
@BytePlusGlobal just released the official prompt guide, and it reveals how to get much better results from the model.
Highlights:
→ Extend videos with continuous creation
→ Generate long-form videos up to 180s
→ Control timing with text prompts
→ Remove unwanted subtitles & BGM
→ Improve reference consistency with multimodal inputs
Comment below and I’ll send you the full guide 👇
English
@Gertywood5 @Cannon_Studio_ you did a great job still. Its very hard to see any deliberate cuts unless you really slow it down.
I have used RIFE again recently to interpolate frames that need that extra smoothing.
English
If I were to do this again I'd probably plot the whole thing out very meticulously by storyboarding all first and last frames per 15 seconds. Upscale those and make sure they all look the same before starting. I kind of hamfisted this tbh and worked a combination of first and last frame plus scene extension. The key is to use motion to hide things.
English
My attempt at making this one-shot sequence. It was more difficult than I expected.
More thoughts below.
English
Minimax H3 v2v test
This was the first try, I am confident if I worked on the prompt a bit I can get this thing to do 1:1 mocap.
English
@dreamingtulpa fantastic series. I love it.
Could you do Unc doing a Blair Witch selfie "I'm so scared" parody?
Maybe, "boo hoo, I ain't scared mother******!" LOL
English
@Uncanny_Harry Derp. I mistyped. I meant to ask if they appreciate using AI as old-school analog artists?
English
@Uncanny_Harry Wow how exciting and fun. A family of artists. So cool. You must’ve destroyed arts and crafts day at elementary school! :)
As old school analog artists do they appreciate it?
English
Uncanny Harry is fast becoming a family affair. My dad is a royal college of Art scholarship graduate, an Illustrator with over 50 years of experience. Here he is drawing the concept drawings for the main characters in my IP "Here be monsters!". AI doesn't kill original art, it can instead be used to amplify it. (The colorful stuff in the background is my mum's work, she's a potter and a painter).
English
@SheriBanks There's a whole lot of fun things happening here. Keep going! It’s amazing how the right music piece elevates. 🤘
English
little expensive mistakes... edit..edit edit....UGH!!!!! this "weekend" project is coming a month....lol
English
@rileybrown Oh damn. Can’t wait to see video. Stoked you set it up already!
English
Asked Codex in Buzz to configure setting up Devin with Kimi and Cursor with Grok, and it did it.
Now the whole gang is here. 5.6 Sol is default, Fable 5, Kimi K3, and Grok 4.5.
English
ElevenLabs has got real competition with Wondercraft and now possibly Fish Audio.
Fish Audio@FishAudio
Today we’ve raised $52M Seed and we are announcing the public launch of S2.1 Pro. >It can clone a voice from 5 seconds of audio >2x faster than Cartesia & 1/6th the cost of Eleven Labs >most expressive model with word level control over emotion, intonation, pacing etc We support frontier AI companies including HeyGen, LiveKit, Retell, Sanas, and OpenArt all run our model in production. If you're a business and we can't cut your voice AI costs by 50%, we'll give you 1 year of Fish Audio for free. Book a demo: s.fish.audio/tmapke To celebrate our first birthday, we'll give you 1 month of S2.1 Pro for free. Like, retweet, and comment “Fish” to get it.
English
@seiiiiiiiiiiru Interesting concept. I’ll have to find time to try it.
English
リリックモーションアプリを作る修行中です🤖
今回は、音楽に合わせて右から左へ流れる歌詞のモーションを試してみました。
AIが多くのことをやってくれると言っても、やればやるほどAIに条件を提示するのも簡単じゃないなと感じています。
このアプリではイントロや間奏やアウトロがあるとその部分も歌詞が流れてしまってズレてくるのを、どういう条件提示すればいいか悩み中…
誰かリリックモーションアプリ作ってる人おる？
コード系のAI使い始めると、今までプラグインが必要だった人力で難しい表現もできるようになりそうなので、今後もコツコツ練習を続けます😋🙏
音楽：Suno @suno
動画：Higgsfield Apps @higgsfield
日本語
Retropunk retweetledi
