Fidelity: Why Does Faithfulness Not Mean Being Identical?

Fidelity describes which important properties of something remain preserved after recording, translation, compression or another transformation. It does not require the result to be identical to the original; if no difference were permitted, transformation itself would lose its purpose.

When music is compressed into a smaller audio file, some data are discarded, yet the melody, rhythm and main timbral qualities may remain sufficient for listening. That can count as high fidelity for enjoying music. The same file may be inadequate if the task is to detect faint noise introduced by recording equipment. Fidelity therefore requires a stated object, purpose and acceptable loss.

It is related to accuracy but not identical to it. Accuracy asks whether a result matches facts or a target; fidelity focuses on the correspondence across a transformation. A translation may contain no factual error yet lose the original tone. Nor does fidelity mean preserving the greatest possible amount of detail: irrelevant detail can obscure the structure that matters. The decisive question is which changes are acceptable and which lost properties would make the result no longer faithful.