# Two caption layers in iCouldPlay: distinguish recorded text from selectable subtitles

Language: en
Canonical: [Two caption layers in iCouldPlay: distinguish recorded text from selectable subtitles](<https://allluckly.cn/en/guides/icouldplay-burned-selectable-caption-check/>)
中文：[中文 HTML](<https://allluckly.cn/guides/icouldplay-burned-selectable-caption-check/>)
English: [English HTML](<https://allluckly.cn/en/guides/icouldplay-burned-selectable-caption-check/>)
Published: 2026-10-27T00:00:00+08:00
Modified: 2026-10-27T13:09:52+08:00
Author: Bison

Use reversible comparisons to identify baked-in text, embedded tracks and external captions, and establish which layer controls affect.

English translation of the [Chinese original](<https://allluckly.cn/guides/icouldplay-burned-selectable-caption-check/>), translated on 2026-10-27. Original publication dates, attribution, code and version notes are retained.

Two visible caption layers can combine recorded text with player captions, or involve selectable tracks and external files. Use observable off/on comparisons to identify them and retain needed information. This does not promise removal of words already stored as picture pixels or require converting the original first.

## Choose one repeatable text moment

Record both layers' wording, colours and positions at one timestamp, then watch briefly. Use authorised samples and retain originals. Appearance alone cannot identify origin because recorded and selectable text can share similar styles.

Record file, source and selected captions, ruling out another cut. Different words may serve labels and dialogue rather than accidental duplication. Establish what to retain before removing useful information merely for a cleaner picture.

## Disable current selectable captions

Use available off controls and revisit the exact moment, observing actual disappearance rather than menu state alone. Remaining text may belong to the picture or another layer outside that control, without automatically establishing a broken switch.

If everything disappears, test the desired selection independently. Otherwise preserve the off condition for further comparison. Avoid loading many files or renaming extensions to guess identity; those actions do not establish semantic source.

## Compare the identical original elsewhere

Open the same file in another suitable tool with its selectable captions disabled. Persistent text suggests possible baked-in pixels, but verify settings and source production information. Two tools showing text under default settings alone cannot prove its structure.

For your own recording, inspect the authoring project for burned-in export. Without it, report observed evidence rather than certainty from one playback. Any clean re-export or image editing belongs in a separate copy with fresh quality checks.

## Select one explicit external file

Load a known simple counterpart in iCouldPlay, verify the fixed sentence, then disable it again. This establishes that layer's control relationship. Compare embedded options only where available, without assuming identical simultaneous-track behaviour across paths.

An embedded container track can remain selectable text or image subtitles, distinct from burned-in words. Embedded does not automatically mean permanently visible. Record whether text comes from container, external import or picture to choose the appropriate remedy.

## Check readability of retained information

Inspect chosen captions against ordinary and bright backgrounds. Where important recorded labels remain, try available placement and style controls to avoid them. Change one factor at the same scene, keeping size separate from timing offsets.

External caption controls cannot recover picture details covered by recorded words. For a clean copy, use your own project or another appropriate legitimate source. Do not describe image-text removal as lossless or promise automatic reconstruction of obscured subjects.

## Recheck identity at other locations

Check beginning, middle and end because recorded titles can be intermittent and selectable cues vanish during silence. One blank moment does not prove a layer absent throughout. Course headings, corner labels and dialogue can serve different purposes despite similar colours.

Reopen and verify selections again after audio or video-version changes. Audio choice does not necessarily alter recorded text, and matching caption names do not establish timing for a new cut. Identify sources before clearing caption folders or removing video entries.

## Finish with a layer inventory

List recorded words, selectable container captions and external captions with source, off behaviour, purpose and edition. Mark absent or uncertain categories honestly. For help, use a shareable small sample and exact times rather than private footage or credentials.

Completion means predicting and verifying which layer a control changes while retaining readable information. Investigate uncertain structure from preserved evidence instead of reinstalling. Source identity is more useful than attractive styling for resolving duplicate text.

## Related tutorials and sources

Continue with [External caption basics](<https://allluckly.cn/en/guides/icouldplay-external-subtitles/>), [Caption appearance](<https://allluckly.cn/en/guides/icouldplay-subtitle-style/>).

Checked on October 27, 2026 through [Apple's public lookup](<https://itunes.apple.com/lookup?id=6755653178&country=cn>): iCouldPlay 1.1.0 requires iOS / iPadOS 13.0 or later. Feature scope follows the [official App Store listing](<https://apps.apple.com/cn/app/id6755653178>). This article describes a workflow and checks, without treating development-only controls as verified public features or claiming step-by-step physical-device or private-server testing. Follow the options in your installed version.