TS-RIR: Translated synthetic room impulse responses for speech augmentation

TL;DR

This paper introduces TS-RIRGAN, a novel method to enhance synthetic room impulse responses by translating them into more realistic versions, significantly improving far-field speech recognition accuracy.

Contribution

The paper presents TS-RIRGAN, a new architecture that translates synthetic RIRs into more realistic ones, bridging the fidelity gap for speech augmentation.

Findings

Improved synthetic RIRs lead to up to 19.9% WER reduction.

Translation with TS-RIRGAN enhances RIR realism.

Method benefits far-field speech recognition performance.

Abstract

We present a method for improving the quality of synthetic room impulse responses for far-field speech recognition. We bridge the gap between the fidelity of synthetic room impulse responses (RIRs) and the real room impulse responses using our novel, TS-RIRGAN architecture. Given a synthetic RIR in the form of raw audio, we use TS-RIRGAN to translate it into a real RIR. We also perform real-world sub-band room equalization on the translated synthetic RIR. Our overall approach improves the quality of synthetic RIRs by compensating low-frequency wave effects, similar to those in real RIRs. We evaluate the performance of improved synthetic RIRs on a far-field speech dataset augmented by convolving the LibriSpeech clean speech dataset [1] with RIRs and adding background noise. We show that far-field speech augmented using our improved synthetic RIRs reduces the word error rate by up to 19.9% in Kaldi far-field automatic speech recognition benchmark [2].