Environment Classification via Blind Roomprints Estimation

Abstract

In this paper we present a novel approach for environment classification for speech recordings, which does not require the selection of decaying reverberation tails. It is based on a multi-band RT60 analysis of blind channel estimates and achieves an accuracy of up to 93.6% on test recordings derived from the ACE corpus.

References in corpus (2)

• Speaker-Independent Microphone Identification in Noisy Conditions
• Spectral Denoising for Microphone Classification

Cited by in corpus (1)

• Blind identification of Ambisonic reduced room impulse response