This paper presents BUT ReverbDB -a dataset of real room impulse responses (RIR), background noises and re-transmitted speech data. The retransmitted data includes LibriSpeech test-clean, 2000 HUB5 English evaluation and part of 2010 NIST Speaker Recognition Evaluation datasets. We provide a detailed description of RIR collection (hardware, software, post-processing) that can serve as a "cook-book" for similar efforts. We also validate BUT ReverbDB in two sets of automatic speech recognition (ASR) experiments and draw conclusions for augmenting ASR training data with real and artificially generated RIRs. We show that a limited number of real RIRs, carefully selected to match the target environment, provide results comparable to a large number of artificially generated RIRs, and that both sets can be combined to achieve the best ASR results. The dataset is distributed for free under a non-restrictive license and it currently contains data from 8 rooms, which is growing. The distribution package also contains a Kaldi-based recipe for augmenting publicly available AMI closetalk meeting data and test the results on an AMI single distant microphone set, allowing it to reproduce our experiments.

No abstract

This paper presents the approach developed by the BUT team for the first DIHARD speech diarization challenge, which is based on our Bayesian Hidden Markov Model with eigenvoice priors system. Besides the description of the approach, we provide a brief analysis of different techniques and data processing methods tested on the development set. We also introduce a simple attempt for overlapped speech detection that we used for attaining cleaner speaker models and reassigning overlapped speech to multiple speakers. Finally, we present results obtained on the evaluation set and discuss findings we made during the development phase and with the help of the DIHARD leaderboard feedback.

For real-world speech recognition applications, noise robustness is still a challenge. In this work, we adopt the teacherstudent (T/S) learning technique using a parallel clean and noisy corpus for improving automatic speech recognition (ASR) performance under multimedia noise. On top of that, we apply a logits selection method which only preserves the k highest values to prevent wrong emphasis of knowledge from the teacher and to reduce bandwidth needed for transferring data. We incorporate up to 8000 hours of untranscribed data for training and present our results on sequence trained models apart from cross entropy trained ones. The best sequence trained student model yields relative word error rate (WER) reductions of approximately 10.1%, 28.7% and 19.6% on our clean, simulated noisy and real test sets respectively comparing to a sequence trained teacher.Index Termsautomatic speech recognition, noise robustness, teacher-student training, domain adaptation * Ladislav Mosner performed the work while he was a research intern at Amazon.

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.