Create YouTube Shorts from a script with voiceover and captions

YouTube

Requested by

Request

English translation

- A specific scenario script is given as input - Generate TTS according to the given scenario - Generate the video and subtitles to match the TTS - The brand must come across well, and the brand intro must not be cut off - The result must be completed properly for the 9:16 Shorts format - Follow the reference video as closely as possible - The subtitle style and editing concept must remain the same

Show original text

- 입력으로 특정 시나리오 스크립트가 주어짐 - 주어진 시나리오에 맞게 tts를 생성 - tts에 맞게 영상과 자막을 생성 - 브랜드가 잘 들어나고 브랜드의 인트로가 잘리면 안됨 - 쇼츠 포맷인 9:16에 알맞게 폼이 완성 되어야함 - 레퍼런스 하는 영상을 최대한 따라하도록 해야함 - 자막 스타일이나 편집 컨셉은 동일하게 유지되어야 함

Provided assets

0

No input materials are provided. Complete the task from the original brief.

Fixed: what stays the same

  • Reference video editing style, subtitle style, character consistency

Variable: up to the maker

  • The video to generate, background music, sound effects

Evaluation criteria

  • Follow the reference video as closely as possible. When generating the narration that appears in the reference video, match it as closely as possible to how a person would pronounce it. Match subtitle sync as closely as possible. Make scen…

Brief

English translation
brief.md
# Auto-Producing YouTube Shorts from a Scenario (TTS + Subtitles)

## What to make
- Reproduce the editing style of the reference video
- Reproduce the subtitle style of the reference video
- Maintain consistent character appearance
- Generate natural-sounding narration pronunciation
- Sync subtitles and audio
- Scene transitions that do not feel awkward
- Appropriate pacing and immersion in the video
- Editing that does not disrupt the overall mood
- Characters presented realistically
- Remove excessive AI characteristics (glossy sheen, unnaturalness)
- No microphone audio quality problems
- Consistent camera angles

## Fixed: what does not change
- Reference video editing style, subtitle style, character consistency

## Variable: what the creator decides
- The video to generate, background music, sound effects

## Evaluation criteria
- Follow the reference video as closely as possible. When generating the narration that appears in the reference video, match it as closely as possible to how a person would pronounce it. Match subtitle sync as closely as possible. Make scene transitions free of awkwardness. For immersive editing, edit at a reasonably brisk pace; however, keep it within bounds that do not disrupt the overall mood. Editing points must be appropriate and must not break the user's immersion. Generate consistently without breaking the character's consistency. Produce the video of the people without a cheap, tacky look. Depict realistic-looking people; they must not look too AI-like or glossy. To produce people as realistically as possible, avoid referencing influencer images or the like as much as possible. I will give you good and bad cases; create the video by referencing these. In the good cases, editing elements such as appropriate editing points and microphone seem well evident. In the bad cases, there are problems such as being edited at fast-forward speed in the middle or camera angles not matching. In particular, there is a problem in the middle where the microphone is hard to hear; you need to avoid such problems as much as possible.

## Client's original text
> If I give you a YouTube reference video link, generate a vlog video with AI in that video's style. It will be made using a video generation API, and you need to use it freely to produce a consistent video. In particular, the person's consistency must not break. You need to make a good video in which the same person appears in the vlog as if explaining things.

No one has submitted an output for this task yet.

More tasks in this category