Simultaneous Speech Translation for Live Subtitling: from Delay to Display

TL;DR

This paper explores adapting simultaneous speech translation systems for live subtitling by predicting subtitle breaks and using scrolling display modes, improving readability and delay management in multilingual live events.

Contribution

It introduces a novel display mode that predicts subtitle breaks and presents subtitles in scrolling lines, enhancing readability and delay control in live translation systems.

Findings

Scrolling line display achieves acceptable reading speed

Delay remains close to 4 seconds with scrolling mode

Translation quality remains a key challenge

Abstract

With the increased audiovisualisation of communication, the need for live subtitles in multilingual events is more relevant than ever. In an attempt to automatise the process, we aim at exploring the feasibility of simultaneous speech translation (SimulST) for live subtitling. However, the word-for-word rate of generation of SimulST systems is not optimal for displaying the subtitles in a comprehensible and readable way. In this work, we adapt SimulST systems to predict subtitle breaks along with the translation. We then propose a display mode that exploits the predicted break structure by presenting the subtitles in scrolling lines. We compare our proposed mode with a display 1) word-for-word and 2) in blocks, in terms of reading speed and delay. Experiments on three language pairs (en$\rightarrow$it, de, fr) show that scrolling lines is the only mode achieving an acceptable reading speed while keeping delay close to a 4-second threshold. We argue that simultaneous translation for readable live subtitles still faces challenges, the main one being poor translation quality, and propose directions for steering future research.