A global voice change may not control every scene. MetaStudio’s live-video guide gives scene-level voice settings priority over global settings and treats uploaded audio as its own source. Inspect the active source for the affected scene before assuming the voice control failed or regenerating the whole project.
Trace each scene’s sound authority
For every scene, identify whether speech comes from text synthesis or an uploaded recording. For text-driven scenes, check whether a local voice choice exists. For audio-driven scenes, listen to the selected file. A global voice label elsewhere in the workspace does not prove that all audio passes through that voice.
Suppose a hypothetical three-scene presentation uses text synthesis in scenes one and three, but a previously approved recording in scene two. Changing the global voice may leave scene two unchanged by design. A separate local setting in scene three can create a third voice. This mixture can be an intentional production choice or an accidental inconsistency; the source inventory reveals which.
Build a scene-level voice ledger
Record a decision for each scene rather than repeatedly toggling the project setting. Keep the words, voice identity and audio file version distinct. An uploaded recording with the correct words can still contain an outdated speaker.
| Scene | Speech source | Local override | Required action |
|---|---|---|---|
| Introduction | Text synthesis | Check actual value | Use approved global or local voice |
| Recorded example | Uploaded audio | Not a substitute for file review | Replace recording only if required |
| Closing | Text synthesis | Check for older choice | Remove or approve intentional difference |
This is an original diagnostic example, not a report of a generated presentation. It does not claim that another platform uses the same priority rules.
Change only the source that governs the defect
If an unwanted local override controls the scene, use the supported setting to correct it. If the old voice is baked into uploaded audio, a different global voice cannot rewrite that recording. Decide whether to retain it, rerecord it or create a reviewed substitute through an appropriate tool. Preserve content approval when making that change.
Preview from before the affected scene through the following transition. Listen for abrupt speaker identity changes, duplicate audio and a recording that starts late. Check the opening and ending of each source, not only a middle sentence. A scene can have the intended voice while still containing the wrong script version.
Deliver an intentional voice pattern
If multiple voices are deliberate, explain their roles in the project record. If one voice is required throughout, verify the actual complete output rather than a global-setting screenshot. Record which scenes were synthesized and which contain supplied audio so the next editor knows where to make changes.
For an Imagild film, the ledger can accompany externally prepared audio and presenter footage. This article does not establish equivalent voice-priority controls in Imagild. Its pass condition is a deliberate, reviewed sound source for every scene and a final playback that matches the approved voice pattern.