This article outlines a number of serious problems arising from the handling within the legal process of covert recordings used as evidence in criminal trials. These problems relate specifically to four key areas, namely: translation of material in languages other than English, transcription of indistinct English, attribution of utterances to speakers and “enhancing” of poor quality audio. The paper traces the problems back to the landmark High Court judgment of Butera 1987, and attributes them to insufficient understanding within the judiciary of well-established but counterintuitive findings of linguistic science regarding factors that affect the reliable interpretation of recorded speech. Several possible solutions to the problems are canvassed, and it is recommended that the most promising way forward is via enhanced communication and collaboration between law, law enforcement and linguistic science.