360° film subtitles. Check beyond the frame.

Prepare timed text on Mac, verify the immersive experience in VR

For an independent filmmaker releasing an authorized narrated 360° film, subtitle preparation has two separate review stages. The words and times must match the final soundtrack. The receiving VR player must also display those words comfortably while the viewer looks around. A successful SRT export establishes neither subtitle depth nor visibility inside a headset.

Polyglot can transcribe supported audio or video, translate reviewed subtitle text locally on an Apple Silicon Mac and export SRT or WebVTT. Use it for that timed-text work. It does not author immersive scenes, encode spherical video, place captions in 3D space or simulate a headset’s view. This workflow covers prerecorded film delivery, not live VR conversations or interactive game dialogue.

Choose the receiving player before building the caption package

Record the intended headset, player and version, video projection mode, storage or streaming method, and accepted subtitle format. Treat each combination as a separate delivery target. A player’s support for subtitles in a flat virtual cinema does not by itself prove the same behavior in a 360° viewing mode.

HereSphere’s developer documented external SRT support in its February 2023 Quest update: the subtitle file goes beside the video with the same base filename. The same document describes VR subtitle depth following autofocus, options for a fixed distance and vertical placement. This establishes a documented receiving workflow, not a test of your current headset or a Polyglot integration. Verify the installed version and actual delivery path before promising compatibility.

Obtain permission for the recording, participants, soundtrack and intended translations. Review the whole sphere, including what is behind the initial view: a person or private detail outside the opening viewport can still become visible. Resolve any restriction in the released media itself; removing words from a subtitle file does not remove them from the audio or picture.

Keep the audio reference on the film’s clock

Keep the final immersive video unchanged. If its media format is unsuitable for your transcription workflow, use your editing tools to export a supported, full-duration audio reference from that exact final cut. Preserve the beginning, pauses and end without trimming, speed changes or extra lead-in. Check speech near the beginning, middle and end against the film before using the reference for subtitle timing.

Generate the source captions from that reference, or import an already reviewed SRT. Correct names, uncertain phrases and meaningful sound information through human review. If spatial audio or simultaneous voices make the words ambiguous, consult the approved script or recording owner; do not infer what someone said from their apparent position. Polyglot does not promise speaker identification or automatic spatial-audio interpretation.

Translate only after the source wording is ready. Check each language for reading load and preserved meaning. An exported subtitle’s timestamps follow media playback; they do not describe a direction, viewing angle or depth. Keep any spatial placement instructions in a separate delivery note for the receiving player.

Mark where looking and reading compete

Make a short review sheet for moments when narration asks the viewer to turn, a relevant sound comes from elsewhere, or a scene cut changes the apparent distance of the subject. Review whether the viewer can find the subject and still read the caption. W3C’s XR accessibility note identifies caption support and customization as user needs; it is not a certification checklist or a promise that one placement works for everyone.

In this fictional scene, a narrator asks the viewer to find a blue doorway behind them. The caption faithfully follows the spoken instruction. The reviewer tests whether it remains readable during the turn, whether it hides the doorway, and whether a recenter action changes the result. These are proposed acceptance checks, not a report of a headset test.

Fictional review sheet
Film: courtyard-cut03.mp4
Audio: courtyard-cut03.wav
Track: courtyard-cut03.srt

00:28–00:31
Heard: “Behind you, the blue
doorway leads to
the courtyard.”

Check in the target headset:
- Read while turning around.
- Find the doorway without
  the caption hiding it.
- Recenter and replay.
- Check after the next cut.

Player/version: __________
Projection/settings: ____
Reviewer/result: ________

Separate text fixes from player placement fixes

If the caption has the wrong words or arrives late, correct the source text or timing and review affected translations. If the text is correctly timed but appears doubled, too close, outside the useful view or in front of an important subject, investigate the receiving player’s rendering and subtitle settings. Changing SRT timestamps cannot repair stereo placement.

HereSphere’s documented depth and pitch controls illustrate why placement belongs to the receiver. Check readability with both eyes during ordinary head movement and across cuts; a readable single-eye view does not establish comfortable stereo presentation. Stop a review if it is uncomfortable, adjust the supported settings and repeat with the intended viewers. Do not compensate by permanently changing faithful caption wording to describe a direction the narrator never said.

W3C’s caption guidance includes speech and meaningful non-speech sound. Review those elements and any essential visual information separately. Captions alone do not make the full immersive experience accessible, and this workflow does not certify accessibility or comfort.

Release one verified video, track and playback configuration

Export the format the destination accepts; for the documented HereSphere example, use SRT and confirm filename pairing in the same folder. Check the exact storage or streaming path you will distribute. Preserve the original exports and keep language-specific packages identifiable so one language never overwrites another file with the same required name.

In the target headset, select the intended captions, start from the beginning, seek across a scene change, pause and resume, recenter, and replay after closing the video. Review both normal forward viewing and the attention changes in your sheet. Confirm the correct language remains selected, the first and last cues appear, and captions clear during silence where appropriate. Repeat after replacing the video, subtitle file or relevant player settings.

Keep the player version, projection, subtitle settings and review result with the delivery package. Leave untested combinations pending. A conventional captioned companion version or descriptive transcript may be useful for other viewers, but requires its own editorial and timing review. Polyglot prepares local text files; headset playback, transfer and spatial acceptance remain separate work.

Keep exploring.