Subtitling is becoming increasingly important for disseminating information,\ngiven the enormous amounts of audiovisual content becoming available daily.\nAlthough Neural Machine Translation (NMT) can speed up the process of\ntranslating audiovisual content, large manual effort is still required for\ntranscribing the source language, and for spotting and segmenting the text into\nproper subtitles. Creating proper subtitles in terms of timing and segmentation\nhighly depends on information present in the audio (utterance duration, natural\npauses). In this work, we explore two methods for applying Speech Translation\n(ST) to subtitling: a) a direct end-to-end and b) a classical cascade approach.\nWe discuss the benefit of having access to the source language speech for\nimproving the conformity of the generated subtitles to the spatial and temporal\nsubtitling constraints and show that length is not the answer to everything in\nthe case of subtitling-oriented ST.\n