2.06 Drafting narrative scripts for audio and video production

From story to script

Once you have a collection of stories and descriptive texts, you may want to create narrated content for outputs like an exhibition audio guide, a short documentary, or a podcast. Transforming a raw transcript into a compelling, spoken script requires a different style of writing. A script needs to be concise, clear, and written specifically to be heard, often using shorter sentences and more direct language. An AI assistant can be an excellent partner in this drafting process, helping to quickly convert your source material into a well-structured script.

AI demonstration: Drafting a narration script

Imagine you want to create a short, 30-second audio introduction for a section of your digital exhibit about traditional life on the Danube. You can provide an AI with the summary you created in Module 4 and ask it to write a script in a specific tone and length.

Prompt:

Based on the story summary about Ana’s spring festival by the river, write a 30-second spoken introduction for an audio guide. It should welcome listeners and set the scene in an engaging way.

Example AI response (script):

Narrator: “Welcome. For generations, life in this village moved to the rhythm of the Danube. It was more than just a river. It was a source of food, a place of work, and the heart of the community’s culture. In the stories that follow, you will hear about the great fishing seasons… the art of weaving baskets from riverbank willows… and the songs that were sung on these very shores. Listen closely, and you can almost hear the echoes of that life today.”

AI demonstration: Translating and localising a script

For a multilingual project like HER[AI]TAGE, AI is a crucial tool for making content accessible to partners and communities in Croatia, Slovenia, and Serbia. An AI can quickly translate your script for different audiences, though it is always best practice to have a native speaker review the translation for cultural nuance and accuracy.

Prompt:

Translate the above introduction script into Croatian in a friendly, spoken tone.

Example AI response (translated script):

Pripovjedač: “Dobrodošli. Generacijama je život u ovom selu tekao u ritmu Dunava. Bio je više od rijeke. Bio je izvor hrane, mjesto rada i srce kulture zajednice. U pričama koje slijede, čut ćete o velikim ribolovnim sezonama… o umijeću pletenja košara od vrbovog pruća s obala rijeke… i o pjesmama koje su se pjevale na ovim istim obalama. Poslušajte pažljivo i gotovo da možete čuti odjeke tog života i danas.”

Conceptual note on recording and audio production

While this course does not require you to record audio, it is important to understand the conceptual process. To record the script, you would follow the best practices from Module 2: find a quiet space, use a good microphone positioned correctly, and speak clearly at a steady, natural pace.

An alternative to human narration is to use AI Text-to-Speech (TTS) tools. Platforms like ElevenLabs or Narration Box can convert a written script into highly realistic, natural-sounding speech in multiple languages.

This technology offers great consistency and accessibility. However, an ethical consideration arises: a synthetic voice, no matter how realistic, may lack the authentic emotional nuance of a human narrator, which is particularly important for personal heritage stories. This decision should be made with care, considering the project’s goals and, if possible, with community consultation.

Tip: Use Google Gemini in Google AI Studio to create these audio recordings with different voices that sound natural and all for free. To do this, you would use Gemini’s multimodal interface, providing the script text and a prompt such as: “Read the following script aloud in a warm, clear female voice for an audio guide.”

In a real-world project, you would likely create a richer audio experience by combining your narration with clips from the original interviews and background sounds (like river ambience or music). This layering is typically done using free audio editing software like Audacity, where you can arrange and mix different audio tracks together. In a tool like Audacity, this would involve importing your narration and interview clips onto separate tracks on the timeline and using the “Time Shift Tool” to position them. As with all created assets, it is good practice to log your scripts and the prompts used to generate them, creating a clear record of your creative process.