Duplicate Cleaner searches selected folders and groups files that match chosen rules. It can compare actual file content, names and dates, image appearance, or audio information. The scan itself does not remove anything. Duplicate Cleaner first builds result groups, then the user marks particular copies and chooses a removal action. This separates it from a storage cleaner that deletes caches according to preset rules without asking which personal copy should remain.
Content beats names
Exact-content mode compares the data inside each file. Two files can match even when their names and dates differ, and exact matches always have the same size. This is the safest starting point for ordinary documents copied into several folders.
Ignore-content mode changes the question. It can group files by chosen combinations of name, size, and date without proving that their bytes match. Two different photographs with a repeated camera filename can then appear related. Use this mode only when the naming rule itself is the subject of the search.
Similarity costs time
Similar-content comparison uses a threshold rather than an exact identity test. A tighter threshold produces closer matches, while a looser threshold can reveal edited variants. The extra analysis takes longer than an exact scan and produces groups that require visual judgment.
Do not interpret a similarity percentage as permission to remove one copy. Cropping, recompression, or a small text change can leave two files visually close while each still has a separate purpose. Preview both items and inspect their folders before marking either one.
Images use appearance
Image mode compares visual content across different formats and resolutions. It can look for rotated or flipped versions and can restrict results by dimensions or aspect ratio. Duplicate Cleaner caches image metrics, so later scans can avoid repeating some of the analysis.
A thumbnail can hide useful differences in sharpness or editing. Open the full images before choosing a keeper. Resolution rules also deserve attention: an exported web copy may look identical in the result pane but contain far fewer pixels than the original photograph.
Audio trusts metadata
Audio mode can compare title and artist tags alongside other track information. This finds the same song when filenames differ, but it inherits every error in the tags. A mistyped artist or a generic title can split a real duplicate group or join unrelated recordings.
An exact audio-data comparison asks a narrower question and does not depend on the tags. Re-encoded files may not match byte for byte even when they contain the same performance. Listen to the beginning and end, then compare bitrate or duration before removing a track from an archive.
Marking is separate
The Selection Assistant marks members of each group according to rules for location, age, size, name, or media properties. Mark All But One keeps one row according to the current ordering. It does not understand which folder is the master archive or which copy carries the preferred metadata.
Set the preferred location rule before a bulk mark and display groups where every member has a mark. That view catches the most damaging selection error: sending every copy in a group to removal. Review unusual groups individually instead of adding another broad rule to force a result.
Protection needs rescanning
Duplicate Cleaner protected folders prevent their files from receiving removal marks. System locations can also stay outside normal deletion work. This protects installed components, but it can confuse a user who tries to mark a visible result and sees no change.
Changing the protected-folder setting after the scan does not rewrite the existing result state. Run the scan again. Check the scan paths at the same time, because a protected master folder and an unprotected backup folder can create a result where only the backup copy remains eligible.
Removal needs a plan
Duplicate Cleaner can move, rename, or delete marked files. The Recycle Bin is useful when the filesystem and path support it, but very long paths can limit that safety net. For a first large cleanup, move marked items into a quarantine folder on the same volume and use the computer normally before permanent deletion. A scan can prove duplication under its rules; it cannot prove that no application still expects one particular path.
