Crafting AI-Generated "Journey to the West" Videos with Internet Jargon

Creating AI-generated videos that blend classic narratives with contemporary internet slang involves a multi-step process, beginning with script development and extending through shot-by-shot production. The "Journey to the West" theme has proven particularly effective for this approach, according to information reviewed by toolmesh.ai.

Desk with laptop showing script and tablet displaying a storyboard for video production.

Script and Shot-by-Shot Planning

The initial script can be co-authored with an AI by providing a concept, such as integrating internet jargon into "Journey to the West" scenes. Once the core script is established, the next phase involves breaking it down into individual shots. Given that current AI video generation tools typically produce clips lasting 5 to 15 seconds, multiple segments must be created and then edited together. For example, a single video might ultimately comprise four distinct shots after extensive editing.

Maintaining character consistency across these short clips is crucial. This requires reference images for each character, ideally encompassing both a starting and an ending frame, or at least a starting frame, as supported by tools like Sora. Since drawing these images can be challenging for non-professionals, AI can assist in generating a detailed shot-by-shot script. This process involves providing the main script and character images to an AI, such as Gemini, with a prompt to outline each shot.

Computer screen showing an AI prompt for generating a shot-by-shot script, with character reference images nearby.

A typical prompt for a shot-by-shot script might request:

1. A description of the shot.

2. A Chinese prompt for drawing the start frame, including detailed scene and character descriptions.

3. A Chinese video prompt based on the start frame image and text, incorporating recommended camera angles and character dialogue.

4. Any other necessary information.

For scenarios requiring both start and end frames, the prompt can be adjusted to include drawing prompts for both frames and a video prompt based on both images, along with character dialogue. Each shot is typically designed to be around 8-10 seconds long.

Split image showing a 'start frame' and 'end frame' for a character, illustrating video continuity.

Ensuring Character Consistency and Video Generation

To ensure visual continuity, it is advisable to first generate a comprehensive character reference image that includes all characters appearing in the video. This image serves as a baseline for subsequent generations. When using AI tools like Gemini, maintaining the same conversation can help ensure that previously generated images are automatically referenced, aiding in consistency. If a new conversation is initiated, the character reference image and relevant previous shot images must be provided to the AI to maintain character and artistic style.

After generating the start frame image for each shot, the video creation process can begin. In platforms like Sora, users select a "Storyboard" option, input the start frame reference image, and manually segment the video prompt for each shot before initiating generation. Other video creation tools follow a similar methodology, relying on reference images and text prompts. If a generated shot does not meet expectations, it may need to be re-generated multiple times before the final clips are manually stitched together to form the complete video.