Dereverberation using joint estimation of dry speech signal and acoustic system
The purpose of speech dereverberation is to remove quality-degrading effects of a time-invariant impulse response filter from the signal. In this report, we describe an approach to speech dereverberation that involves joint estimation of the dry speech signal and of the room impulse response. We explore deep learning models that apply to each task separately, and how these can be combined in a joint model with shared parameters.
Code (0)
등록된 구현이 없습니다.
Tasks
Room Impulse Response (RIR)Speech DereverberationSimilar Papers 제목 키워드 기반
Dereverberation of Autoregressive Envelopes for Far-field Speech Recognition
The task of speech recognition in far-field environments is adversely affected by the reverberant artifacts that elicit as the temporal smearing of the sub-band envelopes. In this paper, we develop a neural model for spe…
Speech Dereverberationspeech-recognitionSpeech RecognitionA Composite T60 Regression and Classification Approach for Speech Dereverberation
Dereverberation is often performed directly on the reverberant audio signal, without knowledge of the acoustic environment. Reverberation time, T60, however, is an essential acoustic factor that reflects how reverberatio…
regressionSpeech DereverberationUnsupervised Blind Joint Dereverberation and Room Acoustics Estimation with Diffusion Models
This paper presents an unsupervised method for single-channel blind dereverberation and room impulse response (RIR) estimation, called BUDDy. The algorithm is rooted in Bayesian posterior sampling: it combines a likeliho…
Room Impulse Response (RIR)Speech DereverberationBlind dereverberation of single channel speech signal based on harmonic structure
The paper presents a new method for dereverberation of speech signals with a single microphone. For applications such as speech recognition, reverberant speech causes serious problems when a distant microphone is used in…
speech-recognitionSpeech RecognitionBUDDy: Single-Channel Blind Unsupervised Dereverberation with Diffusion Models
In this paper, we present an unsupervised single-channel method for joint blind dereverberation and room impulse response estimation, based on posterior sampling with diffusion models. We parameterize the reverberation o…