Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

"speech": models, code, and papers

AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms

Nov 09, 2018
Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, Nobukatsu Hojo

This paper describes a method based on a sequence-to-sequence learning (Seq2Seq) with attention and context preservation mechanism for voice conversion (VC) tasks. Seq2Seq has been outstanding at numerous tasks involving sequence modeling such as speech synthesis and recognition, machine translation, and image captioning. In contrast to current VC techniques, our method 1) stabilizes and accelerates the training procedure by considering guided attention and proposed context preservation losses, 2) allows not only spectral envelopes but also fundamental frequency contours and durations of speech to be converted, 3) requires no context information such as phoneme labels, and 4) requires no time-aligned source and target speech data in advance. In our experiment, the proposed VC framework can be trained in only one day, using only one GPU of an NVIDIA Tesla K80, while the quality of the synthesized speech is higher than that of speech converted by Gaussian mixture model-based VC and is comparable to that of speech generated by recurrent neural network-based text-to-speech synthesis, which can be regarded as an upper limit on VC performance.

* Submitted to ICASSP2019 

  Access Paper or Ask Questions

Exploiting Single-Channel Speech for Multi-Channel End-to-End Speech Recognition: A Comparative Study

Mar 31, 2022
Keyu An, Zhijian Ou

Recently, the end-to-end training approach for multi-channel ASR has shown its effectiveness, which usually consists of a beamforming front-end and a recognition back-end. However, the end-to-end training becomes more difficult due to the integration of multiple modules, particularly considering that multi-channel speech data recorded in real environments are limited in size. This raises the demand to exploit the single-channel data for multi-channel end-to-end ASR. In this paper, we systematically compare the performance of three schemes to exploit external single-channel data for multi-channel end-to-end ASR, namely back-end pre-training, data scheduling, and data simulation, under different settings such as the sizes of the single-channel data and the choices of the front-end. Extensive experiments on CHiME-4 and AISHELL-4 datasets demonstrate that while all three methods improve the multi-channel end-to-end speech recognition performance, data simulation outperforms the other two, at the cost of longer training time. Data scheduling outperforms back-end pre-training marginally but nearly consistently, presumably because that in the pre-training stage, the back-end tends to overfit on the single-channel data, especially when the single-channel data size is small.

* submitted to INTERSPEECH 2022. arXiv admin note: substantial text overlap with arXiv:2107.02670 

  Access Paper or Ask Questions

Read the Room: Adapting a Robot's Voice to Ambient and Social Contexts

May 10, 2022
Emma Hughson, Paige Tuttosi, Akihiro Matsufuji, Angelica Lim

Adapting one's voice to different ambient environments and social interactions is required for human social interaction. In robotics, the ability to recognize speech in noisy and quiet environments has received significant attention, but considering ambient cues in the production of social speech features has been little explored. Our research aims to modify a robot's speech to maximize acceptability in various social and acoustic contexts, starting with a use case for service robots in varying restaurants. We created an original dataset collected over Zoom with participants conversing in scripted and unscripted tasks given 7 different ambient sounds and background images. Voice conversion methods, in addition to altered Text-to-Speech that matched ambient specific data, were used for speech synthesis tasks. We conducted a subjective perception study that showed humans prefer synthetic speech that matches ambience and social context, ultimately preferring more human-like voices. This work provides three solutions to ambient and socially appropriate synthetic voices: (1) a novel protocol to collect real contextual audio voice data, (2) tools and directions to manipulate robot speech for appropriate social and ambient specific interactions, and (3) insight into voice conversion's role in flexibly altering robot speech to match different ambient environments.

* 8 pages 

  Access Paper or Ask Questions

Royalflush Speaker Diarization System for ICASSP 2022 Multi-channel Multi-party Meeting Transcription Challenge

Feb 10, 2022
Jingguang Tian, Xinhui Hu, Xinkang Xu

This paper describes the Royalflush speaker diarization system submitted to the Multi-channel Multi-party Meeting Transcription Challenge. Our system comprises speech enhancement, overlapped speech detection, speaker embedding extraction, speaker clustering, speech separation and system fusion. In this system, we made three contributions. First, we propose an architecture of combining the multi-channel and U-Net-based models, aiming at utilizing the benefits of these two individual architectures, for far-field overlapped speech detection. Second, in order to use overlapped speech detection model to help speaker diarization, a speech separation based overlapped speech handling approach, in which the speaker verification technique is further applied, is proposed. Third, we explore three speaker embedding methods, and obtained the state-of-the-art performance on the CNCeleb-E test set. With these proposals, our best individual system significantly reduces DER from 15.25% to 6.40%, and the fusion of four systems finally achieves a DER of 6.30% on the far-field Alimeeting evaluation set.

  Access Paper or Ask Questions

Investigations on Speech Recognition Systems for Low-Resource Dialectal Arabic-English Code-Switching Speech

Aug 29, 2021
Injy Hamed, Pavel Denisov, Chia-Yu Li, Mohamed Elmahdy, Slim Abdennadher, Ngoc Thang Vu

Code-switching (CS), defined as the mixing of languages in conversations, has become a worldwide phenomenon. The prevalence of CS has been recently met with a growing demand and interest to build CS ASR systems. In this paper, we present our work on code-switched Egyptian Arabic-English automatic speech recognition (ASR). We first contribute in filling the huge gap in resources by collecting, analyzing and publishing our spontaneous CS Egyptian Arabic-English speech corpus. We build our ASR systems using DNN-based hybrid and Transformer-based end-to-end models. In this paper, we present a thorough comparison between both approaches under the setting of a low-resource, orthographically unstandardized, and morphologically rich language pair. We show that while both systems give comparable overall recognition results, each system provides complementary sets of strength points. We show that recognition can be improved by combining the outputs of both systems. We propose several effective system combination approaches, where hypotheses of both systems are merged on sentence- and word-levels. Our approaches result in overall WER relative improvement of 4.7%, over a baseline performance of 32.1% WER. In the case of intra-sentential CS sentences, we achieve WER relative improvement of 4.8%. Our best performing system achieves 30.6% WER on ArzEn test set.

* To be published in Computer Speech and Language Journal 

  Access Paper or Ask Questions

Radically Old Way of Computing Spectra: Applications in End-to-End ASR

Apr 02, 2021
Samik Sadhu, Hynek Hermansky

We propose a technique to compute spectrograms using Frequency Domain Linear Prediction (FDLP) that uses all-pole models to fit the squared Hilbert envelope of speech in different frequency sub-bands. The spectrogram of a complete speech utterance is computed by overlap-add of contiguous all-pole model responses. A long context window of 1.5 seconds allows us to capture the low frequency temporal modulations of speech in the spectrogram. For an end-to-end automatic speech recognition task, the FDLP spectrogram performs on par with the standard mel spectrogram features for clean read speech training and test data. For more realistic speech data with train-test domain mismatches or reverberations, FDLP spectrogram shows up to 25% and 22% relative WER improvements over mel spectrogram respectively.

* submitted to INTERSPEECH 2021 

  Access Paper or Ask Questions

Towards Effective Rebuttal: Listening Comprehension using Corpus-Wide Claim Mining

Jul 27, 2019
Tamar Lavee, Matan Orbach, Lili Kotlerman, Yoav Kantor, Shai Gretz, Lena Dankin, Shachar Mirkin, Michal Jacovi, Yonatan Bilu, Ranit Aharonov, Noam Slonim

Engaging in a live debate requires, among other things, the ability to effectively rebut arguments claimed by your opponent. In particular, this requires identifying these arguments. Here, we suggest doing so by automatically mining claims from a corpus of news articles containing billions of sentences, and searching for them in a given speech. This raises the question of whether such claims indeed correspond to those made in spoken speeches. To this end, we collected a large dataset of $400$ speeches in English discussing $200$ controversial topics, mined claims for each topic, and asked annotators to identify the mined claims mentioned in each speech. Results show that in the vast majority of speeches debaters indeed make use of such claims. In addition, we present several baselines for the automatic detection of mined claims in speeches, forming the basis for future work. All collected data is freely available for research.

* 6th Argument Mining Workshop @ ACL 2019 

  Access Paper or Ask Questions

SpEx: Multi-Scale Time Domain Speaker Extraction Network

Apr 17, 2020
Chenglin Xu, Wei Rao, Eng Siong Chng, Haizhou Li

Speaker extraction aims to mimic humans' selective auditory attention by extracting a target speaker's voice from a multi-talker environment. It is common to perform the extraction in frequency-domain, and reconstruct the time-domain signal from the extracted magnitude and estimated phase spectra. However, such an approach is adversely affected by the inherent difficulty of phase estimation. Inspired by Conv-TasNet, we propose a time-domain speaker extraction network (SpEx) that converts the mixture speech into multi-scale embedding coefficients instead of decomposing the speech signal into magnitude and phase spectra. In this way, we avoid phase estimation. The SpEx network consists of four network components, namely speaker encoder, speech encoder, speaker extractor, and speech decoder. Specifically, the speech encoder converts the mixture speech into multi-scale embedding coefficients, the speaker encoder learns to represent the target speaker with a speaker embedding. The speaker extractor takes the multi-scale embedding coefficients and target speaker embedding as input and estimates a receptive mask. Finally, the speech decoder reconstructs the target speaker's speech from the masked embedding coefficients. We also propose a multi-task learning framework and a multi-scale embedding implementation. Experimental results show that the proposed SpEx achieves 37.3%, 37.7% and 15.0% relative improvements over the best baseline in terms of signal-to-distortion ratio (SDR), scale-invariant SDR (SI-SDR), and perceptual evaluation of speech quality (PESQ) under an open evaluation condition.

* IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2020 
* ACCEPTED in IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP) 

  Access Paper or Ask Questions

Building Bilingual and Code-Switched Voice Conversion with Limited Training Data Using Embedding Consistency Loss

Apr 22, 2021
Yaogen Yang, Haozhe Zhang, Xiaoyi Qin, Shanshan Liang, Huahua Cui, Mingyang Xu, Ming Li

Building cross-lingual voice conversion (VC) systems for multiple speakers and multiple languages has been a challenging task for a long time. This paper describes a parallel non-autoregressive network to achieve bilingual and code-switched voice conversion for multiple speakers when there are only mono-lingual corpora for each language. We achieve cross-lingual VC between Mandarin speech with multiple speakers and English speech with multiple speakers by applying bilingual bottleneck features. To boost voice cloning performance, we use an adversarial speaker classifier with a gradient reversal layer to reduce the source speaker's information from the output of encoder. Furthermore, in order to improve speaker similarity between reference speech and converted speech, we adopt an embedding consistency loss between the synthesized speech and its natural reference speech in our network. Experimental results show that our proposed method can achieve high quality converted speech with mean opinion score (MOS) around 4. The conversion system performs well in terms of speaker similarity for both in-set speaker conversion and out-set-of one-shot conversion.

* Submitted to Interspeech 2021 

  Access Paper or Ask Questions