On the use of X-vectors for Robust Speaker Recognition

Abstract

Text-independent speaker verification (SV) is currently in the process of embracing DNN modeling in every stage of SV system. Slowly, the DNN-based approaches such as end-to-end modelling and systems based on DNN embeddings start to be competitive even in challenging and diverse channel conditions of recent NIST SREs. Domain adaptation and the need for a large amount of training data are still a challenge for current discriminative systems and (unlike with generative models), we see significant gains from data augmentation, simulation and other techniques designed to overcome lack of training data. We present an analysis of a SV system based on DNN embeddings (x-vectors) and focus on robustness across diverse data domains such as standard telephone and microphone conversations, both in clean, noisy and reverberant environments. We also evaluate the system on challenging far-field data created by re-transmitting a subset of NIST SRE 2008 and 2010 microphone interviews. We compare our results with the stateof-the-art i-vector system. In general, we were able to achieve better performance with the DNN-based systems, but most importantly, we have confirmed the robustness of such systems across multiple data domains.

Factorization of Discriminatively Trained i-Vector Extractor for Speaker Recognition

“…We show that, with such an approach, we can achieve a reasonable performance. Our results are perhaps not as competitive as those achieved with current state-of-the-art x-vector systems [18], nevertheless, we are now closer to our goal which is to further use this model in the fully end-to-end discriminative system [19] that can be initialized from a robust generative baseline. Figure 1: Scheme of an end-to-end speaker verification system based on a feed forward NN designed to mimic a generic speaker verification system ( [19]).…”

Discriminatively Re-trained I-vector Extractor for Speaker Recognition

“…We are fully aware that we do not reach the performance of x-vectors. Results presented here can be directly compared to our previous work [14] focused on analyzing the performance of the state-of-the-art i-vector and x-vector systems on the very same datasets. Here we present the i-vector system that is based purely on MFCCs, while in [14] we were using concatenation of MFCCs and DNN bottleneck features.…”

“…To mix the reverberation, noise and signal at given SNR, we followed the procedure outlined in [14]. When jointly augmenting the data by noise and reverberation, the speech and noise are reverberated separately and different RIRs from the same room are used for speech signal and noise to simulate different positions of their sources.…”

Analysis of DNN Speech Signal Enhancement for Robust Speaker Recognition

“…After we explore the benefits of DNN-based audio pre-processing with standard generative SV systems based on i-vectors and PLDA, we attempt to improve an already better baseline system where DNN replaces the crucial i-vector extraction step. We use the architecture proposed by David Snyder Snyder (2017), Snyder et al (2017) which already presents the x-vector (the embedding) as a robust feature for PLDA modeling, and provides state-of-the-art results across various acoustic conditions (Novotný et al, 2018b). We experiment with using the denoising autoencoder as a pre-processing step while training the x-vector extractor or just during the test stage.…”