TSUBAKI

453 posts

TSUBAKI banner
TSUBAKI

TSUBAKI

@AI__TSUBAKI

AI Video Creator|SNS総フォロワー30万人 CPP: Kling, Dreamina, Hailuo, PixVerse, Pollo, Domo, ImagineArt, JoyPix, CapCut, SJinn, Anthum, Video Rebirth, Renoise等|作品はハイライト

Japan Katılım Temmuz 2015
30 Takip Edilen196 Takipçiler
TSUBAKI
TSUBAKI@AI__TSUBAKI·
Prompt: Style: Ultra-realistic mass celebrity arrival scene. Single continuous shot. Handheld camera from crowd perspective. Natural micro-shake. No cuts. Documentary-level realism. Audio: Only natural environment sound loud crowd cheering, overlapping voices shouting, rapid camera shutter clicks, phones recording audio, distant city noises, footsteps, fabric movement. Lighting: Nighttime, illuminated from street lights. Mixed reflections on polished surfaces. Soft but realistic shadows. Slight atmospheric haze for depth. Main Character: (Subject Reference Image) Calm, controlled presence. Subtle confident smile. Face identity must remain perfectly consistent across all frames. Outfit: Exactly the same from reference image. Location: An event in Roppongi, Tokyo. Scene Flow 0–3s: Camera starts from inside a dense crowd behind barricades. Handheld, slightly unstable. View partially blocked by people in front. Multiple phones raised, some screens visible recording. Crowd energy is loud, chaotic, restless. 3–6s: Camera lifts slightly above shoulder level, still handheld. Focus shifts naturally between heads, waving hands, and we see our subject exit a fancy venue in Roppongi, Tokyo. Occasional bright flashes from media cameras. Anticipation rises as crowd leans forward. 6–10s: Security personnel step in, pushing the crowd back slightly. Camera reacts with natural shake. Through shifting gaps, the main character appears in the distance—initially soft and partially obscured, gradually becoming clearer while walking forward with a small escort team. She walks up to a fan and signs a shikishi (autograph board) the fan is holding. 10–13s: Subtle handheld push-in (natural movement, not digital zoom). The main character is now clearly visible, walking confidently at center frame. Path is being cleared. She raises one hand and gives a calm, controlled wave with a slight smile. Camera struggles slightly to keep framing due to crowd movement. 13–15s: Camera shifts and tilts trying to follow. A luxury convoy becomes partially visible featuring three large premium SUVs. A security member opens the back door of the middle SUV. The main character enters quickly and rolls the window down. Vehicles begin moving forward as she waves to the crowd. Camera lifts slightly as people jump and try to capture the moment.
TSUBAKI tweet media
English
0
0
3
81
TSUBAKI
TSUBAKI@AI__TSUBAKI·
SJinnのSeedance 2.0いいですね 実写画像もOKです プロンプトはリプ @SJinn_Agent
日本語
1
1
12
445
TSUBAKI
TSUBAKI@AI__TSUBAKI·
BACH-1.0 Preview ―まだ知られてない生成モデル― 最大30秒マルチショット 解像度1080p、リファレンス画像14枚 正しい日本語生成 初登場でKlingと肩を並べる実力 5月7日 正式リリース @video_rebirth #BachCPP
日本語
5
2
33
4.1K
TSUBAKI
TSUBAKI@AI__TSUBAKI·
BACHを使用して驚いたのは 「焼きそば」「金魚すくい」などの日本語が正しく生成されたことです あとりんご飴もちゃんとりんご飴の見た目してます🍎 ベビーカステラは難しかったね?でも十分すごい リファレンス画像はキャラクターシートのみです BACH↓ bach.art/?invitecode=AL…
TSUBAKI tweet mediaTSUBAKI tweet media
日本語
0
0
3
229
Wei Liu
Wei Liu@DrLiuWei·
Our BACH 1.0 Preview model just debuted at #6 on @ArtificialAnlys. Same leaderboard as Alibaba, ByteDance, and xAI. Proud of this team. We're just warming up.
Artificial Analysis@ArtificialAnlys

Bach-1.0 Preview from Video Rebirth debuts at #6 on the Artificial Analysis Text to Video Leaderboard (No Audio)! Bach-1.0 Preview is the latest Text to Video model from @video_rebirth, with similar performance to Vidu Q3 Pro, Kling 3.0 Omni 1080p (Pro), and grok-imagine-video. Bach-1.0 Preview is intended for broad release later in May. See example generations from Bach-1.0 Preview in the Artificial Analysis Video Arena below 🧵

English
5
0
9
1K
TSUBAKI
TSUBAKI@AI__TSUBAKI·
Video Rebirthの「Bach-1.0 Preview」が Artificial AnalysisのT2Vで6位になりました これはKling 3.0 Omniと 同等の性能を備えていることを意味します Bach-1.0 Previewは 5月下旬に一般公開される予定です
TSUBAKI tweet media
TSUBAKI@AI__TSUBAKI

VIDEO REBIRTH ―まだ知られてない生成モデル― なんと30秒マルチショット 解像度1080p、リファレンス画像14枚まで ↑これって世界初?どうなんだろう コンテストの作品作らないといけないのに 色々試したくて生成してしまう… @video_rebirth #BachCPP

日本語
1
0
7
310
TSUBAKI
TSUBAKI@AI__TSUBAKI·
If you want to become a CPP, Anthum AI(@anthum_ai)is an option. Follower count and views don’t matter. All that matters is winning the contest. Only winners become CPP. And once you become a CPP, you’ll get prize money,plus free access to Seedance 2.0 and Kling through Anthum Studio.
Anthum AI@anthum_ai

We're looking for the best AI ad creators on the internet. Not just the best generators - the ones who understand advertising. Hook. Story. Brand alignment. Conversion. $3,000 in prizes 💰💸. 20,000 credits. 💳 The Anthum Ad Challenge is live. Link in bio.

English
0
0
1
152
TSUBAKI
TSUBAKI@AI__TSUBAKI·
AEW版で作ってみた GPT Images 2.0→Seedance 2.0 Video Prompt: Hyper-realistic cinematic 15-second continuous video of a women's championship finale in a sold-out arena. Featuring Riho (Japanese, idol-like cuteness, black hair, very petite, light blue and white gear) vs Mercedes Moné (American, powerful and charismatic physique, gold-themed gear). Follow the storyboard sequence precisely: exhausted mid-match staredown → hard clothesline → submission struggle → DDT reversal → near fall → German suplex → top-rope moonsault finisher → three-count pin → championship victory moment. Maintain consistent facial identity and body proportions throughout. Realistic sweat, impact physics, and intense crowd reactions. Use slow motion for key moments. End with confetti and corner pyrotechnics as a cinematic crane shot rises on the winner. Photorealistic, AEW broadcast quality, 60fps, high-contrast lighting. Made with @dreamina_ai #DreaminaCPP
TSUBAKI tweet media
WasifAI@doctorwasif

Chatgpt GPT-2 & Seedance 2 on @yapper_so Video Prompt: Hyper-realistic cinematic 15-second continuous video of a WWE championship finale between two beautiful athletic Japanese women inside a sold-out arena. Follow the storyboard sequence exactly — from exhausted mid-match staredown, clothesline, submission struggle, DDT reversal, near fall, German suplex, top rope moonsault finisher, three-count pin, to championship belt victory celebration with confetti and corner pyrotechnics firing. Both women in matching ring gear throughout with consistent identity. Sweat, impact physics, crowd reactions, and slow-motion on key moments. Cinematic crane shot rising on final victory pose. Photorealistic, WWE broadcast quality, 60fps.

日本語
2
1
7
1.1K
TSUBAKI
TSUBAKI@AI__TSUBAKI·
Storyboard Prompt: Hyper-realistic 4×4 AEW-style storyboard (16 panels), set in a dramatic arena with spotlights, haze, and a roaring crowd. Featuring Riho, a Japanese female wrestler (idol-like cuteness, black hair, 23 years old, 156 cm, very petite, wearing light blue and white gear), versus Mercedes Moné, an American female wrestler (28 years old, 165 cm). Maintain consistent facial features across all panels. Realistic sweat, dynamic motion, and cinematic compositions capturing strikes and high-flying moves. High-contrast lighting enhances the intensity of the match and the energy of the crowd.
English
0
0
0
204
TSUBAKI
TSUBAKI@AI__TSUBAKI·
BL盗撮ドラマにしてみた 教師と生徒の恋愛はダメですよ @dreamina_ai #DreaminaCPP @capcutapp_jp #CapCut生成AI
神奇小喷菇AIGC@xpg0970

seedance2.0 prompt:偷拍风格,手持iPhone从街对面较远距离偷偷拍摄,使用手机相机**数字放大(digital zoom in)**效果,夜景,强烈抖动镜头,典型iPhone数字放大时的颗粒感、轻微模糊和噪点,手机画质,轻微玻璃反光和街灯折射,画面不稳、偶尔失焦,像路人站在街对面用iPhone远距离偷录并慢慢放大。 场景:高端餐厅内部,桌子上摆着红酒杯、餐具。桌子前坐着一位亚洲长短发男生( @图片1)和一位短发的外国男( @图片2),他们在低声交谈、微笑、看iPhone 17 Pro Max、碰杯。背景有其他食客和柔和灯光。 镜头运动:从街对面较远位置先拍摄餐厅外立面,然后用手机相机**缓慢数字zoom in**推近玻璃窗内部,镜头持续抖动、寻找焦点,zoom过程中画面出现iPhone典型的数字放大颗粒和轻微失真,偶尔被路过的行人或车辆短暂挡住,极强真实偷拍感。最后3-4秒,两人突然发现被拍,同时回头看向镜头方向,露出惊讶、错愕的表情,动作自然。 时长:15秒左右。 旁白:用@音频1 的男声中文旁白同步播放,内容是惊讶的吐槽语气。 风格:真实手机偷拍视频质感,手持摇晃强烈,夜景暖光,冷暖色混合,轻微噪点,像真实iPhone远距离zoom拍摄。特别强调街对面数字放大时的手机画质和抖动感。(无背景音乐)

日本語
0
0
0
171
TSUBAKI
TSUBAKI@AI__TSUBAKI·
Higgsfieldが23分の映像を4日間で制作🧩 相変わらず彼らの作る映像はすごい てか進撃の巨人みたいなの出てきた 日本語喋ってるし 制作工程の一部↓ ・基本的にSeedance 2.0の15秒で生成 ・「文脈優先」のプロンプト方式を採用 ・声の一貫性を保つために各キャラクターの声の特徴をすべてのプロンプトに記載(例:「低い日本語の声」) プロって凄いな かっこいい
Higgsfield AI 🧩@higgsfield

Check the full breakdown of the first 23 minutes here x.com/PJaccetturo/st…

日本語
0
0
0
227
TSUBAKI
TSUBAKI@AI__TSUBAKI·
『余花 ―Afterbloom―』 米国アカデミー賞公認 日本発・アジア最大級の国際短編映画祭 「ホラー&サスペンス部門」応募作品 SSFF & ASIA専用シナリオ 3分 / マルチエンディング(Bパターン) #最恐ホラーショート #友情に潜む恐怖
国際短編映画祭 SSFF & ASIA@s_s_f_f

バズより、震え。 米国アカデミー賞公認SSFF & ASIA × CRG 縦型ホラー&サスペンス部門 作品募集スタート。 👻 ▶ 応募方法・詳細はこちら app.lifelogbox.com/festival/short… ―――――― フォロワー数も、再生数も問わない。 必要なのは、 30秒で記憶に刻む“最恐の1カット”。 「最震賞 Supported by CRG」受賞者は、 CRG所属SNSホラー映像作家として 活動できる可能性も。 🎬 賞金最大70万円

日本語
0
0
5
422