Embedded cover art
Embedded cover art carries image resources inside media metadata, with picture type, encoding and selection determining which image reaches the library.
A media file can carry cover images in its metadata structures. Their bytes form image resources associated with the item; they are not additional sound samples.
The enclosing tag describes the picture
| Representation | Relevant boundary |
|---|---|
| ID3v2 APIC | Picture type, MIME information, description and image bytes |
| FLAC PICTURE metadata | Structured picture information and encoded image payload |
| MP4 cover metadata | Cover entries interpreted through that metadata convention |
The actual image still needs a matching image decoder. A tag reader locating an APIC frame does not itself decompress its JPEG or PNG payload.
Several pictures need a selection rule
ID3's attached-picture types distinguish, among other roles, front and back covers. A file can consequently contain more than one legitimate image.
Selecting the first encountered picture can pick a back cover or another role. An integration should retain the type and relevant description where its model uses them.
In Emby music, compare the actual embedded candidates with the selected library image. A provider's album artwork and the file's embedded front cover can both be available sources under another selection policy.
Image bytes have their own cost
An original file might carry encoded pictures of 60,000 and 80,000 bytes. Their combined image payload is 140,000 bytes, before tag/block overhead.
Those bytes contribute to file size and retrieval but do not increase the audio's sample count. A whole-file bitrate calculation includes them; a codec-only audio rate has another scope.
Decoded image memory also depends on its image grid and output representation, not simply on this encoded byte total.
Metadata preservation is independent of audio copying
A rewrite can retain coded audio while replacing or omitting cover metadata. The sound payload can remain the same while the whole-file checksum changes.
An extraction or conversion therefore needs a picture-preservation decision as well as an audio decision. Keep role, image type and selected resource together when assessing the resulting cover.