Audio Description

Part 1: Create an Audio Description project

1. Select the project type

When creating a new project, select Audio Description as the project type.

2. Upload the source video

Upload the source video and give the file a clear, recognizable name.

Supported formats: mp3, mov. Maximum file size/duration: 5 GB.

3. Set the source language

Select the language spoken in the source video.

Dubformer analyzes the original dialogue to identify where speech occurs and capture its timing. These timings define the available spaces where AD narration can be placed without overlapping the original dialogue.

Note: automatic gap detection works best with clear dialogue separation.

4. Select the AD language

Choose the language for the AD narration.

By default, the AD language matches the source language, as Audio Description is typically created in the same language as the original dialogue.

5. Upload supporting materials (optional)

Add any supporting files that may be useful during AD production:

• M&E (Music & Effects) — improves mix quality of the final AD track

• Vocals audio — improves dialogue gap detection accuracy

• As recorded (original) script — if an existing script should be used instead of ASR. Accepted formats: xlsx, csv, .srt with timecodes.

6. Generate Audio Description automatically

Select Auto-generate audio description to have Dubformer generate an AD script automatically.

The AD engine analyzes source dialogue timings and generates description that fits within the available spaces between spoken lines, following standard AD writing conventions (objective, present-tense description; no interpretation of intent or emotion unless visually evident).

7. Create the project

When all settings are ready, click Create project.

Part 2: Working inside the Audio Description task

Once the project is created, open the task to review, refine, and finalize the AD script and narration.

1. Open the task or assign an editor

Open the task yourself, or assign it to an editor.

If you selected auto-generation, opening the task will display the automatically generated description text, aligned to the identified gaps in dialogue.

2. Select narration voices

Go to the Speakers tab to choose voices for AD narration. Three options are available:

• Library voices — pre-built voices available in Dubformer.

• Cloned voices — create a voice clone from an audio sample.

• Voice design — generate a voice from a text prompt.

3. Edit text and performance

Refine the script and narration using two tabs:

• Script editor — click a chunk to edit its text. [See: Task editor: tools and controls]

• Visual editor

Control pronunciation and delivery using performance prompts and other options. [See: https://help.dubformer.ai/en/help/collections/7773399-ai-dubbing-enhancement-toolkit

4. Preview narration

Play back the generated narration against the video to confirm timing, pacing, and mix balance before finalizing.

5. Finalize

Save your progress and Process the changes to generate the final result. 

Part 3: Deliverables

On completion, export results.