Specify the film's language independently of the website interface, then decide whether the film actually contains speech. An English menu does not establish the language of a generated soundtrack. Conversely, an English screenplay can describe a wordless film. Treat interface, story, audio and captions as four separate decisions.
Make a language and sound card
Before planning shots, write down the audience's reading language, spoken language if any, caption language and whether characters should move their lips. Include pronunciation notes only when speech will be recorded. A silent visual story needs no fictional dialogue for a model to guess.
| Layer | Example decision | Review method |
|---|---|---|
| Interface | Use the site in Spanish | Check navigation separately |
| Story prose | Write the treatment in English | Read the draft |
| Final speech | No speech; environmental sound only | Listen to the full output |
| Captions | English title added during editing | Proofread the actual graphic |
This avoids a common brief conflict: requesting an ambience-only scene, adding a voice-over paragraph and then asking the generator to maintain silent lips. Remove the conflicting paragraph rather than repeating the language instruction more forcefully.
Choose ambience when the action carries the meaning
An ambience-led wedding film can use footsteps, fabric movement, a door opening and a ring box closing. The script describes what viewers see; it is not a transcript to be spoken. If a power or a relationship detail only makes sense after a narrator explains it, revise the visible action before generating.
Film language: English. This is a visual story with location ambience and gentle physical sound effects, no dialogue and no narration. The couple communicate through gaze, a three-finger countdown and taking hands. Any title will be added and proofread separately after the footage is selected.
Review the returned audio even with this instruction. Listen for unintended speech, stray words, repetitive effects and sounds that contradict the setting. Muting a defective track may be preferable to keeping an unexplained voice. Sound replacement requires an appropriate editing tool and cleared audio assets.
Narration requires a different approval point
A narration-led project needs the complete approved spoken text and an actual recording before precise timing can be trusted. Word count gives a rough planning estimate, not a measured duration. Pauses, names and delivery style affect the schedule.
Imagild's wedding workflow does not promise a finished voice-audition service. If narration progression is unavailable or paused in your interface, do not bypass it by treating prose as recorded audio. Prepare the recording externally and establish how it will be edited before paying to render a sequence that must match it.
Test meaning without sound as well
Watch a rough selection muted. The audience should still understand who is acting and what changes. Then listen without looking: are there words you did not authorize? Those two checks catch different problems and are more reliable than trusting the interface's selected language.
Will changing the interface translate an existing video?
No. Navigation localization is not dubbing. Existing media must be reviewed or edited separately.
Where should I start in Imagild?
Use Film Studio for the story and shot workflow. Approve language and sound intent before generation, and verify the actual media rather than assuming the prompt guarantees compliance.