ManYue scan orientation checks: distinguish stored pixels from displayed rotation

A scan that appears upright in one viewer but sideways in a comic can involve source pixels, orientation metadata or viewer handling. Compare clear corner marks and text without equating device rotation with source rotation or assuming a universal auto-correct button. Preserve originals before deciding whether an external reading derivative is necessary.

Create a sample that reveals orientation and mirroring

Make a picture labelled top, bottom, left and right, plus an asymmetric arrow and number. Symmetric scenery or a round icon cannot reliably reveal mirroring. Preserve an ordinary source verified in a familiar tool. Any simulated rotation belongs in a separately documented working copy. Use a previously verified common-image flow rather than a special format; the labels provide objective evidence independent of whether the phone looks portrait or landscape.

Record dimensions and available orientation information

Inspect stored dimensions and any exposed orientation information. Distinguish pixel arrangement from a viewer-adjusted appearance. Not every format/tool exposes equivalent fields; absence of a visible field means unconfirmed, not necessarily no metadata. Avoid undocumented numeric edits. Record two trusted viewing results and preserve any disagreement before the app comparison so differing interpretation of orientation remains a testable possibility.

Open the actual page with device posture fixed

Import through the established local-image/archive route and hold orientation fixed. Inspect labels and arrow on the body page, not only the cover. Record the resulting top edge, mirroring and completeness. Keep device auto-rotation, reader orientation and source direction separate. A temporary single-page view can simplify checking; record its baseline. Document the sideways result before rotating and repackaging away the original evidence.

Check what device rotation actually changes

Rotate under the actual permitted system/reader conditions and observe whether scale, visible area or image direction changes. Landscape holding does not necessarily repair upside-down content. Check orientation lock before diagnosing image damage. Return to the baseline posture each time with the same file/page. No unrelated accessibility changes or universal claims about decoding paths are needed; record the conditions observed.

Distinguish mirroring from a scene viewpoint

Use asymmetric lettering and arrows rather than a character’s handedness or a changing viewpoint. Reversed text together with swapped corner labels is stronger mirroring evidence; sideways rotation is another case. If the original is mirrored, the reader cannot be assumed to recover an earlier scan. For actual artwork compare a reliable same-edition source before correcting deliberate reversed signage or a mirror scene.

Make an explicitly oriented derivative

When modification is authorised and justified, use a familiar external tool to export a separately oriented copy and reopen it. A rotate click alone may affect only view state or metadata. Test the marked sample first for upright text, no mirroring, preserved ratio and intact edges. Correct only confirmed problem pages and retain the archive original. Label the derivative, without promising every viewer will behave identically or treating extension changes as processing.

Verify identity and position after import

Import the derivative as a separate sample and verify that the result comes from that output rather than another file or cache. For a repackaged book inspect ordering, structure, beginning/middle/end and corrected pages. A new entry may have separate records; map by chapter and scene. For a remote-only discrepancy inspect server derivatives before changing every original to compensate for an unconfirmed delivery path.

Acceptance and reporting material

Finish with verifiable corner orientation, legible text, intact edges and a clear distinction between posture and file processing, while preserving the original. Success in one external tool does not certify every reader. Report the self-made sample, source/derivative operations, app version and observations without private books. Retain a usable copy, label unknowns and avoid repeated destructive rotations that erase the initial state or assume identical metadata across formats.

Related guides and sources

ManYue test comic: verify an import with six numbered pages; ManYue large-image diagnostics: separate file size, dimensions and loading.

Checked on November 3, 2026: Apple public version information lists ManYue 2.0.4, requiring iOS / iPadOS 15.0 or later. Feature scope follows the official App Store description and the installed version. This article describes actions and content checks you can perform; it does not claim step-by-step physical-device or private-server testing. Manual records, external tools and application controls are identified separately, and code observations or personal inferences are not presented as verified automatic capabilities.

For orientation information, see Apple ImageIO image orientation. This does not imply identical handling across all formats and viewing paths.