# ManYue: Distinguish identical files, similar images and intentional repeated pages

Language: en
Canonical: [ManYue: Distinguish identical files, similar images and intentional repeated pages](<https://allluckly.cn/en/guides/manyue-suspected-duplicate-page-identity/>)
中文：[中文 HTML](<https://allluckly.cn/guides/manyue-suspected-duplicate-page-identity/>)
English: [English HTML](<https://allluckly.cn/en/guides/manyue-suspected-duplicate-page-identity/>)
Published: 2026-11-08T00:00:00+08:00
Modified: 2026-11-08T09:34:15+08:00
Author: Bison

Use position records, optional file fingerprints and detail checks without batch deletion based on names.

English translation of the [Chinese original](<https://allluckly.cn/guides/manyue-suspected-duplicate-page-identity/>), translated on 2026-11-08. Original publication dates, attribution, code and version notes are retained.

Two similar consecutive images may be duplicate exports, changed dialogue, revisions or intentional repetition. Matching names do not establish matching content, and different bytes do not guarantee a visible difference. Record positions in ManYue and compare original page files externally. The result is an evidence-preserving judgement, not automatic deduplication, similarity recognition or one-click duplicate-page deletion.

## Fix each page’s separate position

In a local comic you may read, record both suspected pages’ file positions, chapters, neighbouring images and edition, naming them earlier A and later B. Retain the archive instead of deleting one to see whether the story improves. If a paired layout creates the impression, use an available single-page view to establish two actual positions rather than two halves of one spread. Note the first occurrence and surrounding continuity, preserving a distinct locator for each. Without original page files, visual comparison remains possible, but mark file identity unavailable.

## Record full paths alongside filenames

Locate corresponding originals with a trusted archive or file tool, recording full relative path, name, size and pixel dimensions. Page one in different subfolders need not be one entry, and copy in a filename does not prove duplication. Inspect read-only or extract the two known images from a working copy without overwriting namesakes. Preserve both sources when a name conflict appears instead of comparing after overwrite. If reader positions cannot be mapped to files, defer file-level conclusions rather than substituting two similar covers.

## Optionally compare bytes with an existing fingerprint tool

If a trusted SHA-256 tool already exists on a computer, calculate both image fingerprints and record algorithm, full result and matching path. Otherwise skip this part without installing unfamiliar deduplication software. Matching fingerprints strongly support byte identity only when the same algorithm and complete files are compared. Do not compare an archive fingerprint with one image or inspect only a few initial characters. Different fingerprints establish different bytes, potentially from metadata, encoding or image content, not necessarily newly drawn content. Fingerprints certify neither rights, authentic provenance nor image safety.

## Compare image and lettering at matched display scale

In familiar viewers, compare at similar display scale, checking borders, poses, dialogue, corner page numbers and props. Record at least three specific regions. If one is a thumbnail, revisit dimensions rather than interpreting blurred missing letters as authorial deletion. Recompression can create small noise differences that are not plot changes. Preserve genuinely changed dialogue with a necessary short excerpt and location, retaining both originals instead of overwriting one for consistency. Visually matching but byte-different is a valid result; the two judgement types need not agree.

## Use context to assess the purpose of repetition

Read a page before and after in ManYue. Repeated composition might create a pause, recall, viewpoint comparison or dialogue variation, but each interpretation needs textual support. Even identical files can be intentionally repeated, so byte identity does not establish redundancy. For repeated covers at a chapter boundary, determine whether each opens a different chapter or both occur consecutively within one. Without author explanation or a reliable edition comparison, retain suspected repetition pending review rather than declaring a page deletable from personal preference. Keep narrative purpose separate from file identity.

## Change only a confirmed item in a comparison copy

Only for your own organized material with a confirmed misplaced duplicate export should you consider a separately saved comparison copy. Move the confirmed item outside the archive and retain it rather than deleting permanently. Record omitted path, original position and reason, leaving other names and order unchanged. After import, check neighbouring continuity and the final page, not merely a count reduced by one. Revert to the original for any unexplained change instead of expanding the operation. Do not remove intentional repetitions or changed-dialogue pages this way, call a comparison copy the complete edition or discard old reading records.

## Check mapping and provenance when comparison fails

If matching fingerprints seem to accompany clearly different images, verify result-to-path mapping, stale viewer windows and whether only names were compared. For unclear visual differences, match scale and dimensions and wait for stable loading. Preserve mappings for tool-renamed extracted files instead of treating new names as original paths. If only the reader repeats a page while the archive does not, record positions, page-turn actions and version for a minimal reproduction. That differs from duplicate source entries; do not delete source pages to accommodate a display anomaly.

## Preserve three layers of findings after returning

Leave the reader and return to A and B and their surrounding anchors, ensuring position changes did not substitute different objects. Summarize three layers: confirmed file identity, visible similarities/differences and evidence of narrative purpose. Mark fingerprint or copy steps not performed. Preserve original, path mapping, necessary local records and comparison copy separately, sharing only authorized material. Completion explains retention, uncertainty or isolation in a copy of your own material. It does not reduce every similar image to one or authorize automatic cleanup of the whole library.

## Related guides and sources

[ManYue archive companion-file audit: distinguish pages, notes and system helpers](<https://allluckly.cn/en/guides/manyue-archive-auxiliary-file-audit/>); [ManYue tall-image seam evidence: distinguish repeated content, gaps and loading blanks](<https://allluckly.cn/en/guides/manyue-strip-seam-evidence-pair/>).

Checked on November 8, 2026: [Apple public version information](<https://itunes.apple.com/lookup?id=6753895889&country=cn>) lists ManYue 2.0.4, requiring iOS / iPadOS 15.0 or later. Feature scope follows the [official App Store description](<https://apps.apple.com/cn/app/id6753895889>) and the installed version. This article provides reading and checking methods you can perform; it does not claim step-by-step physical-device or private-server testing. Manual records, external tools and actually visible controls are identified separately, and personal analysis or code observations are not presented as verified automatic capabilities.

For file fingerprints and existing tool interfaces, see the [official Perl Digest::SHA documentation](<https://perldoc.perl.org/Digest::SHA>). Fingerprints assist file-identity checks here, not visual similarity or provenance certification.