正文 · AI 翻译
IT Home reported on October 26 that ByteDance’s Jimeng AI announced the official launch of Jimeng Short Drama Studio on October 26, offering its self-developed script model DreamScript, image model Anycook, and storyboard model DreamShot.
Self-developed script model DreamScript: Supports script generation across multiple genres, covering mainstream types such as original works, adaptations, content for overseas markets, and narrated content; analyzes the potential for scripts to become hits by identifying platform trends and the factors behind popular works, helping optimize plot structure; analyzes long texts of up to 100,000 Chinese characters, automatically understanding and parsing them to greatly improve creative efficiency.
Self-developed image model Anycook: Supports distinctive character designs; tuned specifically for short-drama art production, it generates more realistic and recognizable short-drama characters; builds scenes in multiple styles, covering diverse visual styles to match story settings precisely.
Self-developed storyboard model DreamShot: Offers two creation modes, allowing users to switch freely between canvas and storyboard to suit their workflow; enables coordinated editing across the entire storyboard, supporting revisions through dialogue, global changes, and flexible plot adjustments while keeping shots organized; provides dynamic previews through a 3D director’s console, with freely configurable camera positions and movements to preview motion effects in advance; intelligently previews key storyboard shots, linking key shots to assess visual continuity and narrative pacing in advance.
Dedicated post-production tools: Reshoot a segment when a single shot is unsatisfactory, targeting that shot without starting the entire production over; reconstruct audio by separating voices from background sound and generating new dubbing; automatically split shots with intelligent cuts to make complex footage easier to edit; and share assets across Jimeng, so creative assets connect seamlessly and a change in one place takes effect throughout the workflow.
Remaking and visual conversion: Intelligent breakdown of the original video automatically identifies characters, scenes, props, and shot structure after a short drama is uploaded; visual remaking regenerates footage according to the selected visual style, aspect ratio, and target region; localization preserves the original drama’s plot pacing and shot language while adapting it for different markets.