Studies on stop unrelease in second language acquisition have hitherto focused on the productions of Slavic learners of English (Šimáčková & Podlipský, 2015) and experiments on Polish learners of English; the latter show the tendency to release stops on a more regular basis depending on the type of stop combinations (Rojczyk et al. 2013). In the present study, we aim to test the efficiency of audio-visual explanations as opposed to distracted imitation in pronunciation teaching amongst French learners of English. While unreleased stops are rather frequent in French and English - especially in plosives clusters (Byrd, 1993; Davidson, 2010), unreleased plosives in final positions are less common in French (Van Dommelen, 1983). During phase 1 of the experiment, three groups of 12 native French learners of English (level A1/A2, B1/B2 and C1/C2) were asked to read idiomatic expressions containing both homogeneous and heterogeneous sequences of voiceless stops straddled between words, namely, in sequences like “that cat” [ðæt˺ kæt˺], and stops at the end of sentences like “I told him to speak” [tə spiːk˺]. In the second phase of the experiment, one half in each group was given a different task. The first group heard recorded versions of phase 1 sentences and before reading them out loud, counted up to five in their L1. Stimuli for imitation contained no release in the contexts under scrutiny. The other half had to watch a video explaining the phenomenon of unreleased stops with a production of phase-two expressions propped up by hand gestures. They were then asked to re-read the sentences given in phase 1. Based on these results the current study makes recommendations about what working environment should be prioritized in pronunciation teaching both in class and online (Kröger et al. 2010), and suggests ways to assess students and visually keep track of their progress.