Advanced Rich Transcription System for Estonian Speech

TL;DR

This paper presents an advanced Estonian speech transcription system capable of handling diverse acoustic conditions and semi-spontaneous speech, achieving low error rates and incorporating features like punctuation and speaker ID.

Contribution

The paper introduces a robust Estonian speech transcription system utilizing multi-condition training, phoneme-based decoding, and weakly supervised speaker identification, advancing language-specific speech recognition technology.

Findings

Word error rate of 8.1% on broadcast conversations

Effective noise robustness through background noise profile adaptation

Successful integration of punctuation and speaker identification

Abstract

This paper describes the current TT\"U speech transcription system for Estonian speech. The system is designed to handle semi-spontaneous speech, such as broadcast conversations, lecture recordings and interviews recorded in diverse acoustic conditions. The system is based on the Kaldi toolkit. Multi-condition training using background noise profiles extracted automatically from untranscribed data is used to improve the robustness of the system. Out-of-vocabulary words are recovered using a phoneme n-gram based decoding subgraph and a FST-based phoneme-to-grapheme model. The system achieves a word error rate of 8.1% on a test set of broadcast conversations. The system also performs punctuation recovery and speaker identification. Speaker identification models are trained using a recently proposed weakly supervised training method.