Skip to content

Glossary

Re-encoding

Re-encoding is decoding an image and compressing it again, which discards detail on every pass through a lossy format.

Updated

Also written: transcoding, recompression.

A lossy codec throws away information judged imperceptible. Do that once and the result usually looks fine. Do it repeatedly — each time on the already-degraded output of the last pass — and losses compound, visibly.

This is the hidden cost in most online metadata tools. Decoding to a bitmap and encoding a fresh file drops the metadata as a side effect and takes a few lines of code, so it is the common implementation. The metadata does go, and the picture is quietly worse.

It is usually detectable without any tooling: a re-encoded file comes back a very different size from the original, for reasons that have nothing to do with the metadata that was removed.

For video the cost is steeper still — transcoding is slow, and motion and fine texture degrade first.

Related terms

  • Lossless — Lossless means the image data comes out bit-for-bit identical to what went in.
  • Container format — A container format is the file structure that wraps compressed image or video data together with the metadata describing it.
  • Provenance — Provenance is the record of where an image came from and what has been done to it since.

Questions that turn on this

Back to the glossary →