Speech-to-Speech Translation

Select a WAV file corresponding to an ENGLISH text. Please ensure it was recoreded as clearly as possible.

Processing chain:

For this implementation we used the following:

  • Romanian ASR from the ROBIN Project: Andrei-Marius Avram, Vasile Păiș, Dan Tufiș. 2020. Towards a Romanian end-to-end automatic speech recognition based on DeepSpeech2. Proc. Ro. Acad., Series A, Volume 21, No. 4, pp. 395-402.

  • Romanian TTS from http://romaniantts.com : Adriana Stan, Junichi Yamagishi, Simon King, Matthew Aylett, "The Romanian speech synthesis (RSS) corpus: Building a high quality HMM-based speech synthesis system using a high sampling rate", In Speech Communication, vol. 53, no. 3, pp. 442-450, 2011.

  • Romanian SSLA TTS: Tiberiu Boroș, Ștefan D. Dumitrescu, Vasile Păiș, "Tools and resources for Romanian text-to-speech and speech-to-text applications", CoRR, vol. abs/1802.05583, 2018. https://arxiv.org/pdf/1802.05583.pdf

  • Translation using the EU Council Presidency Translator developed by Tilde with support from RACAI during the Romanian presidency.

  • English DeepSpeech2 ASR from: Amodei et al. 2016. Deep Speech 2 : End-to-End Speech Recognition in English and Mandarin. In Proceedings of The 33rd International Conference on Machine Learning (PMLR), 48:173-182.

  • English Mozilla DeepSpeech ASR: Hannun et al. 2016. Deep Speech: Scaling up end-to-end speech recognition. arXiv:1412.5567 [cs.CL]

  • English Mozilla TTS: https://github.com/mozilla/TTS