We describe the large vocabulary automatic speech recognition system developed for Modern Standard Arabic by the SRI/Nightingale team, and used for the 2007 GALE evaluation as part of the speech translation system. We show how system performance is affected by different development choices, ranging from text processing and lexicon to decoding system architecture design. Word error rate results are reported on broadcast news and conversational data from the GALE development and evaluation test sets.
Bibliographic reference. Vergyri, D. / Mandal, A. / Wang, Wen / Stolcke, Andreas / Zheng, Jing / Graciarena, Martin / Rybach, D. / Gollan, Christian / Schlüter, Ralf / Kirchhoff, Katrin / Faria, A. / Morgan, Nelson (2008): "Development of the SRI/nightingale Arabic ASR system", In INTERSPEECH-2008, 1437-1440.